File size: 10,218 Bytes
d1ceb73 |
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 |
import os
import re
from pathlib import Path
from importlib.machinery import all_suffixes
from jedi.inference.cache import inference_state_method_cache
from jedi.inference.base_value import ContextualizedNode
from jedi.inference.helpers import is_string, get_str_or_none
from jedi.parser_utils import get_cached_code_lines
from jedi.file_io import FileIO
from jedi import settings
from jedi import debug
_BUILDOUT_PATH_INSERTION_LIMIT = 10
def _abs_path(module_context, str_path: str):
path = Path(str_path)
if path.is_absolute():
return path
module_path = module_context.py__file__()
if module_path is None:
# In this case we have no idea where we actually are in the file
# system.
return None
base_dir = module_path.parent
return base_dir.joinpath(path).absolute()
def _paths_from_assignment(module_context, expr_stmt):
"""
Extracts the assigned strings from an assignment that looks as follows::
sys.path[0:0] = ['module/path', 'another/module/path']
This function is in general pretty tolerant (and therefore 'buggy').
However, it's not a big issue usually to add more paths to Jedi's sys_path,
because it will only affect Jedi in very random situations and by adding
more paths than necessary, it usually benefits the general user.
"""
for assignee, operator in zip(expr_stmt.children[::2], expr_stmt.children[1::2]):
try:
assert operator in ['=', '+=']
assert assignee.type in ('power', 'atom_expr') and \
len(assignee.children) > 1
c = assignee.children
assert c[0].type == 'name' and c[0].value == 'sys'
trailer = c[1]
assert trailer.children[0] == '.' and trailer.children[1].value == 'path'
# TODO Essentially we're not checking details on sys.path
# manipulation. Both assigment of the sys.path and changing/adding
# parts of the sys.path are the same: They get added to the end of
# the current sys.path.
"""
execution = c[2]
assert execution.children[0] == '['
subscript = execution.children[1]
assert subscript.type == 'subscript'
assert ':' in subscript.children
"""
except AssertionError:
continue
cn = ContextualizedNode(module_context.create_context(expr_stmt), expr_stmt)
for lazy_value in cn.infer().iterate(cn):
for value in lazy_value.infer():
if is_string(value):
abs_path = _abs_path(module_context, value.get_safe_value())
if abs_path is not None:
yield abs_path
def _paths_from_list_modifications(module_context, trailer1, trailer2):
""" extract the path from either "sys.path.append" or "sys.path.insert" """
# Guarantee that both are trailers, the first one a name and the second one
# a function execution with at least one param.
if not (trailer1.type == 'trailer' and trailer1.children[0] == '.'
and trailer2.type == 'trailer' and trailer2.children[0] == '('
and len(trailer2.children) == 3):
return
name = trailer1.children[1].value
if name not in ['insert', 'append']:
return
arg = trailer2.children[1]
if name == 'insert' and len(arg.children) in (3, 4): # Possible trailing comma.
arg = arg.children[2]
for value in module_context.create_context(arg).infer_node(arg):
p = get_str_or_none(value)
if p is None:
continue
abs_path = _abs_path(module_context, p)
if abs_path is not None:
yield abs_path
@inference_state_method_cache(default=[])
def check_sys_path_modifications(module_context):
"""
Detect sys.path modifications within module.
"""
def get_sys_path_powers(names):
for name in names:
power = name.parent.parent
if power is not None and power.type in ('power', 'atom_expr'):
c = power.children
if c[0].type == 'name' and c[0].value == 'sys' \
and c[1].type == 'trailer':
n = c[1].children[1]
if n.type == 'name' and n.value == 'path':
yield name, power
if module_context.tree_node is None:
return []
added = []
try:
possible_names = module_context.tree_node.get_used_names()['path']
except KeyError:
pass
else:
for name, power in get_sys_path_powers(possible_names):
expr_stmt = power.parent
if len(power.children) >= 4:
added.extend(
_paths_from_list_modifications(
module_context, *power.children[2:4]
)
)
elif expr_stmt is not None and expr_stmt.type == 'expr_stmt':
added.extend(_paths_from_assignment(module_context, expr_stmt))
return added
def discover_buildout_paths(inference_state, script_path):
buildout_script_paths = set()
for buildout_script_path in _get_buildout_script_paths(script_path):
for path in _get_paths_from_buildout_script(inference_state, buildout_script_path):
buildout_script_paths.add(path)
if len(buildout_script_paths) >= _BUILDOUT_PATH_INSERTION_LIMIT:
break
return buildout_script_paths
def _get_paths_from_buildout_script(inference_state, buildout_script_path):
file_io = FileIO(str(buildout_script_path))
try:
module_node = inference_state.parse(
file_io=file_io,
cache=True,
cache_path=settings.cache_directory
)
except IOError:
debug.warning('Error trying to read buildout_script: %s', buildout_script_path)
return
from jedi.inference.value import ModuleValue
module_context = ModuleValue(
inference_state, module_node,
file_io=file_io,
string_names=None,
code_lines=get_cached_code_lines(inference_state.grammar, buildout_script_path),
).as_context()
yield from check_sys_path_modifications(module_context)
def _get_parent_dir_with_file(path: Path, filename):
for parent in path.parents:
try:
if parent.joinpath(filename).is_file():
return parent
except OSError:
continue
return None
def _get_buildout_script_paths(search_path: Path):
"""
if there is a 'buildout.cfg' file in one of the parent directories of the
given module it will return a list of all files in the buildout bin
directory that look like python files.
:param search_path: absolute path to the module.
"""
project_root = _get_parent_dir_with_file(search_path, 'buildout.cfg')
if not project_root:
return
bin_path = project_root.joinpath('bin')
if not bin_path.exists():
return
for filename in os.listdir(bin_path):
try:
filepath = bin_path.joinpath(filename)
with open(filepath, 'r') as f:
firstline = f.readline()
if firstline.startswith('#!') and 'python' in firstline:
yield filepath
except (UnicodeDecodeError, IOError) as e:
# Probably a binary file; permission error or race cond. because
# file got deleted. Ignore it.
debug.warning(str(e))
continue
def remove_python_path_suffix(path):
for suffix in all_suffixes() + ['.pyi']:
if path.suffix == suffix:
path = path.with_name(path.stem)
break
return path
def transform_path_to_dotted(sys_path, module_path):
"""
Returns the dotted path inside a sys.path as a list of names. e.g.
>>> transform_path_to_dotted([str(Path("/foo").absolute())], Path('/foo/bar/baz.py').absolute())
(('bar', 'baz'), False)
Returns (None, False) if the path doesn't really resolve to anything.
The second return part is if it is a package.
"""
# First remove the suffix.
module_path = remove_python_path_suffix(module_path)
if module_path.name.startswith('.'):
return None, False
# Once the suffix was removed we are using the files as we know them. This
# means that if someone uses an ending like .vim for a Python file, .vim
# will be part of the returned dotted part.
is_package = module_path.name == '__init__'
if is_package:
module_path = module_path.parent
def iter_potential_solutions():
for p in sys_path:
if str(module_path).startswith(p):
# Strip the trailing slash/backslash
rest = str(module_path)[len(p):]
# On Windows a path can also use a slash.
if rest.startswith(os.path.sep) or rest.startswith('/'):
# Remove a slash in cases it's still there.
rest = rest[1:]
if rest:
split = rest.split(os.path.sep)
if not all(split):
# This means that part of the file path was empty, this
# is very strange and is probably a file that is called
# `.py`.
return
# Stub folders for foo can end with foo-stubs. Just remove
# it.
yield tuple(re.sub(r'-stubs$', '', s) for s in split)
potential_solutions = tuple(iter_potential_solutions())
if not potential_solutions:
return None, False
# Try to find the shortest path, this makes more sense usually, because the
# user usually has venvs somewhere. This means that a path like
# .tox/py37/lib/python3.7/os.py can be normal for a file. However in that
# case we definitely want to return ['os'] as a path and not a crazy
# ['.tox', 'py37', 'lib', 'python3.7', 'os']. Keep in mind that this is a
# heuristic and there's now ay to "always" do it right.
return sorted(potential_solutions, key=lambda p: len(p))[0], is_package
|