File size: 15,083 Bytes
7885a28 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 |
"""unit tests for sparse utility functions"""
import numpy as np
from numpy.testing import assert_equal
import pytest
from pytest import raises as assert_raises
from scipy.sparse import _sputils as sputils, csr_array, bsr_array, dia_array, coo_array
from scipy.sparse._sputils import matrix
class TestSparseUtils:
def test_upcast(self):
assert_equal(sputils.upcast('intc'), np.intc)
assert_equal(sputils.upcast('int32', 'float32'), np.float64)
assert_equal(sputils.upcast('bool', complex, float), np.complex128)
assert_equal(sputils.upcast('i', 'd'), np.float64)
def test_getdtype(self):
A = np.array([1], dtype='int8')
assert_equal(sputils.getdtype(None, default=float), float)
assert_equal(sputils.getdtype(None, a=A), np.int8)
with assert_raises(
ValueError,
match="scipy.sparse does not support dtype object. .*",
):
sputils.getdtype("O")
with assert_raises(
ValueError,
match="scipy.sparse does not support dtype float16. .*",
):
sputils.getdtype(None, default=np.float16)
def test_isscalarlike(self):
assert_equal(sputils.isscalarlike(3.0), True)
assert_equal(sputils.isscalarlike(-4), True)
assert_equal(sputils.isscalarlike(2.5), True)
assert_equal(sputils.isscalarlike(1 + 3j), True)
assert_equal(sputils.isscalarlike(np.array(3)), True)
assert_equal(sputils.isscalarlike("16"), True)
assert_equal(sputils.isscalarlike(np.array([3])), False)
assert_equal(sputils.isscalarlike([[3]]), False)
assert_equal(sputils.isscalarlike((1,)), False)
assert_equal(sputils.isscalarlike((1, 2)), False)
def test_isintlike(self):
assert_equal(sputils.isintlike(-4), True)
assert_equal(sputils.isintlike(np.array(3)), True)
assert_equal(sputils.isintlike(np.array([3])), False)
with assert_raises(
ValueError,
match="Inexact indices into sparse matrices are not allowed"
):
sputils.isintlike(3.0)
assert_equal(sputils.isintlike(2.5), False)
assert_equal(sputils.isintlike(1 + 3j), False)
assert_equal(sputils.isintlike((1,)), False)
assert_equal(sputils.isintlike((1, 2)), False)
def test_isshape(self):
assert_equal(sputils.isshape((1, 2)), True)
assert_equal(sputils.isshape((5, 2)), True)
assert_equal(sputils.isshape((1.5, 2)), False)
assert_equal(sputils.isshape((2, 2, 2)), False)
assert_equal(sputils.isshape(([2], 2)), False)
assert_equal(sputils.isshape((-1, 2), nonneg=False),True)
assert_equal(sputils.isshape((2, -1), nonneg=False),True)
assert_equal(sputils.isshape((-1, 2), nonneg=True),False)
assert_equal(sputils.isshape((2, -1), nonneg=True),False)
assert_equal(sputils.isshape((1.5, 2), allow_nd=(1, 2)), False)
assert_equal(sputils.isshape(([2], 2), allow_nd=(1, 2)), False)
assert_equal(sputils.isshape((2, 2, -2), nonneg=True, allow_nd=(1, 2)),
False)
assert_equal(sputils.isshape((2,), allow_nd=(1, 2)), True)
assert_equal(sputils.isshape((2, 2,), allow_nd=(1, 2)), True)
assert_equal(sputils.isshape((2, 2, 2), allow_nd=(1, 2)), False)
def test_issequence(self):
assert_equal(sputils.issequence((1,)), True)
assert_equal(sputils.issequence((1, 2, 3)), True)
assert_equal(sputils.issequence([1]), True)
assert_equal(sputils.issequence([1, 2, 3]), True)
assert_equal(sputils.issequence(np.array([1, 2, 3])), True)
assert_equal(sputils.issequence(np.array([[1], [2], [3]])), False)
assert_equal(sputils.issequence(3), False)
def test_ismatrix(self):
assert_equal(sputils.ismatrix(((),)), True)
assert_equal(sputils.ismatrix([[1], [2]]), True)
assert_equal(sputils.ismatrix(np.arange(3)[None]), True)
assert_equal(sputils.ismatrix([1, 2]), False)
assert_equal(sputils.ismatrix(np.arange(3)), False)
assert_equal(sputils.ismatrix([[[1]]]), False)
assert_equal(sputils.ismatrix(3), False)
def test_isdense(self):
assert_equal(sputils.isdense(np.array([1])), True)
assert_equal(sputils.isdense(matrix([1])), True)
def test_validateaxis(self):
assert_raises(TypeError, sputils.validateaxis, (0, 1))
assert_raises(TypeError, sputils.validateaxis, 1.5)
assert_raises(ValueError, sputils.validateaxis, 3)
# These function calls should not raise errors
for axis in (-2, -1, 0, 1, None):
sputils.validateaxis(axis)
@pytest.mark.parametrize("container", [csr_array, bsr_array])
def test_safely_cast_index_compressed(self, container):
# This is slow to test completely as nnz > imax is big
# and indptr is big for some shapes
# So we don't test large nnz, nor csc_array (same code as csr_array)
imax = np.int64(np.iinfo(np.int32).max)
# Shape 32bit
A32 = container((1, imax))
# indices big type, small values
B32 = A32.copy()
B32.indices = B32.indices.astype(np.int64)
B32.indptr = B32.indptr.astype(np.int64)
# Shape 64bit
# indices big type, small values
A64 = csr_array((1, imax + 1))
# indices small type, small values
B64 = A64.copy()
B64.indices = B64.indices.astype(np.int32)
B64.indptr = B64.indptr.astype(np.int32)
# indices big type, big values
C64 = A64.copy()
C64.indices = np.array([imax + 1], dtype=np.int64)
C64.indptr = np.array([0, 1], dtype=np.int64)
C64.data = np.array([2.2])
assert (A32.indices.dtype, A32.indptr.dtype) == (np.int32, np.int32)
assert (B32.indices.dtype, B32.indptr.dtype) == (np.int64, np.int64)
assert (A64.indices.dtype, A64.indptr.dtype) == (np.int64, np.int64)
assert (B64.indices.dtype, B64.indptr.dtype) == (np.int32, np.int32)
assert (C64.indices.dtype, C64.indptr.dtype) == (np.int64, np.int64)
for A in [A32, B32, A64, B64]:
indices, indptr = sputils.safely_cast_index_arrays(A, np.int32)
assert (indices.dtype, indptr.dtype) == (np.int32, np.int32)
indices, indptr = sputils.safely_cast_index_arrays(A, np.int64)
assert (indices.dtype, indptr.dtype) == (np.int64, np.int64)
indices, indptr = sputils.safely_cast_index_arrays(A, A.indices.dtype)
assert indices is A.indices
assert indptr is A.indptr
with assert_raises(ValueError):
sputils.safely_cast_index_arrays(C64, np.int32)
indices, indptr = sputils.safely_cast_index_arrays(C64, np.int64)
assert indices is C64.indices
assert indptr is C64.indptr
def test_safely_cast_index_coo(self):
# This is slow to test completely as nnz > imax is big
# So we don't test large nnz
imax = np.int64(np.iinfo(np.int32).max)
# Shape 32bit
A32 = coo_array((1, imax))
# coords big type, small values
B32 = A32.copy()
B32.coords = tuple(co.astype(np.int64) for co in B32.coords)
# Shape 64bit
# coords big type, small values
A64 = coo_array((1, imax + 1))
# coords small type, small values
B64 = A64.copy()
B64.coords = tuple(co.astype(np.int32) for co in B64.coords)
# coords big type, big values
C64 = A64.copy()
C64.coords = (np.array([imax + 1]), np.array([0]))
C64.data = np.array([2.2])
assert A32.coords[0].dtype == np.int32
assert B32.coords[0].dtype == np.int64
assert A64.coords[0].dtype == np.int64
assert B64.coords[0].dtype == np.int32
assert C64.coords[0].dtype == np.int64
for A in [A32, B32, A64, B64]:
coords = sputils.safely_cast_index_arrays(A, np.int32)
assert coords[0].dtype == np.int32
coords = sputils.safely_cast_index_arrays(A, np.int64)
assert coords[0].dtype == np.int64
coords = sputils.safely_cast_index_arrays(A, A.coords[0].dtype)
assert coords[0] is A.coords[0]
with assert_raises(ValueError):
sputils.safely_cast_index_arrays(C64, np.int32)
coords = sputils.safely_cast_index_arrays(C64, np.int64)
assert coords[0] is C64.coords[0]
def test_safely_cast_index_dia(self):
# This is slow to test completely as nnz > imax is big
# So we don't test large nnz
imax = np.int64(np.iinfo(np.int32).max)
# Shape 32bit
A32 = dia_array((1, imax))
# offsets big type, small values
B32 = A32.copy()
B32.offsets = B32.offsets.astype(np.int64)
# Shape 64bit
# offsets big type, small values
A64 = dia_array((1, imax + 2))
# offsets small type, small values
B64 = A64.copy()
B64.offsets = B64.offsets.astype(np.int32)
# offsets big type, big values
C64 = A64.copy()
C64.offsets = np.array([imax + 1])
C64.data = np.array([2.2])
assert A32.offsets.dtype == np.int32
assert B32.offsets.dtype == np.int64
assert A64.offsets.dtype == np.int64
assert B64.offsets.dtype == np.int32
assert C64.offsets.dtype == np.int64
for A in [A32, B32, A64, B64]:
offsets = sputils.safely_cast_index_arrays(A, np.int32)
assert offsets.dtype == np.int32
offsets = sputils.safely_cast_index_arrays(A, np.int64)
assert offsets.dtype == np.int64
offsets = sputils.safely_cast_index_arrays(A, A.offsets.dtype)
assert offsets is A.offsets
with assert_raises(ValueError):
sputils.safely_cast_index_arrays(C64, np.int32)
offsets = sputils.safely_cast_index_arrays(C64, np.int64)
assert offsets is C64.offsets
def test_get_index_dtype(self):
imax = np.int64(np.iinfo(np.int32).max)
too_big = imax + 1
# Check that uint32's with no values too large doesn't return
# int64
a1 = np.ones(90, dtype='uint32')
a2 = np.ones(90, dtype='uint32')
assert_equal(
np.dtype(sputils.get_index_dtype((a1, a2), check_contents=True)),
np.dtype('int32')
)
# Check that if we can not convert but all values are less than or
# equal to max that we can just convert to int32
a1[-1] = imax
assert_equal(
np.dtype(sputils.get_index_dtype((a1, a2), check_contents=True)),
np.dtype('int32')
)
# Check that if it can not convert directly and the contents are
# too large that we return int64
a1[-1] = too_big
assert_equal(
np.dtype(sputils.get_index_dtype((a1, a2), check_contents=True)),
np.dtype('int64')
)
# test that if can not convert and didn't specify to check_contents
# we return int64
a1 = np.ones(89, dtype='uint32')
a2 = np.ones(89, dtype='uint32')
assert_equal(
np.dtype(sputils.get_index_dtype((a1, a2))),
np.dtype('int64')
)
# Check that even if we have arrays that can be converted directly
# that if we specify a maxval directly it takes precedence
a1 = np.ones(12, dtype='uint32')
a2 = np.ones(12, dtype='uint32')
assert_equal(
np.dtype(sputils.get_index_dtype(
(a1, a2), maxval=too_big, check_contents=True
)),
np.dtype('int64')
)
# Check that an array with a too max size and maxval set
# still returns int64
a1[-1] = too_big
assert_equal(
np.dtype(sputils.get_index_dtype((a1, a2), maxval=too_big)),
np.dtype('int64')
)
# tests public broadcast_shapes largely from
# numpy/numpy/lib/tests/test_stride_tricks.py
# first 3 cause np.broadcast to raise index too large, but not sputils
@pytest.mark.parametrize("input_shapes,target_shape", [
[((6, 5, 1, 4, 1, 1), (1, 2**32), (2**32, 1)), (6, 5, 1, 4, 2**32, 2**32)],
[((6, 5, 1, 4, 1, 1), (1, 2**32)), (6, 5, 1, 4, 1, 2**32)],
[((1, 2**32), (2**32, 1)), (2**32, 2**32)],
[[2, 2, 2], (2,)],
[[], ()],
[[()], ()],
[[(7,)], (7,)],
[[(1, 2), (2,)], (1, 2)],
[[(2,), (1, 2)], (1, 2)],
[[(1, 1)], (1, 1)],
[[(1, 1), (3, 4)], (3, 4)],
[[(6, 7), (5, 6, 1), (7,), (5, 1, 7)], (5, 6, 7)],
[[(5, 6, 1)], (5, 6, 1)],
[[(1, 3), (3, 1)], (3, 3)],
[[(1, 0), (0, 0)], (0, 0)],
[[(0, 1), (0, 0)], (0, 0)],
[[(1, 0), (0, 1)], (0, 0)],
[[(1, 1), (0, 0)], (0, 0)],
[[(1, 1), (1, 0)], (1, 0)],
[[(1, 1), (0, 1)], (0, 1)],
[[(), (0,)], (0,)],
[[(0,), (0, 0)], (0, 0)],
[[(0,), (0, 1)], (0, 0)],
[[(1,), (0, 0)], (0, 0)],
[[(), (0, 0)], (0, 0)],
[[(1, 1), (0,)], (1, 0)],
[[(1,), (0, 1)], (0, 1)],
[[(1,), (1, 0)], (1, 0)],
[[(), (1, 0)], (1, 0)],
[[(), (0, 1)], (0, 1)],
[[(1,), (3,)], (3,)],
[[2, (3, 2)], (3, 2)],
[[(1, 2)] * 32, (1, 2)],
[[(1, 2)] * 100, (1, 2)],
[[(2,)] * 32, (2,)],
])
def test_broadcast_shapes_successes(self, input_shapes, target_shape):
assert_equal(sputils.broadcast_shapes(*input_shapes), target_shape)
# tests public broadcast_shapes failures
@pytest.mark.parametrize("input_shapes", [
[(3,), (4,)],
[(2, 3), (2,)],
[2, (2, 3)],
[(3,), (3,), (4,)],
[(2, 5), (3, 5)],
[(2, 4), (2, 5)],
[(1, 3, 4), (2, 3, 3)],
[(1, 2), (3, 1), (3, 2), (10, 5)],
[(2,)] * 32 + [(3,)] * 32,
])
def test_broadcast_shapes_failures(self, input_shapes):
with assert_raises(ValueError, match="cannot be broadcast"):
sputils.broadcast_shapes(*input_shapes)
def test_check_shape_overflow(self):
new_shape = sputils.check_shape([(10, -1)], (65535, 131070))
assert_equal(new_shape, (10, 858967245))
def test_matrix(self):
a = [[1, 2, 3]]
b = np.array(a)
assert isinstance(sputils.matrix(a), np.matrix)
assert isinstance(sputils.matrix(b), np.matrix)
c = sputils.matrix(b)
c[:, :] = 123
assert_equal(b, a)
c = sputils.matrix(b, copy=False)
c[:, :] = 123
assert_equal(b, [[123, 123, 123]])
def test_asmatrix(self):
a = [[1, 2, 3]]
b = np.array(a)
assert isinstance(sputils.asmatrix(a), np.matrix)
assert isinstance(sputils.asmatrix(b), np.matrix)
c = sputils.asmatrix(b)
c[:, :] = 123
assert_equal(b, [[123, 123, 123]])
|