def test_as_path_registering():
@generic_path.register_pathlike_cls('my_path://')
class MyPath(gpath.PosixGPath):
pass
my_path = generic_path.as_path('my_path://abc')
assert isinstance(my_path, MyPath)
assert generic_path.as_path(my_path) is my_path
def test_as_path_new():
p0 = type_utils.ReadWritePath('some/path/to/xyz')
p1 = type_utils.ReadOnlyPath('some/path/to/xyz')
p2 = generic_path.as_path('some/path/to/xyz')
assert p0 == p1
assert p0 == p2
assert type(p0) is type(p1)
assert type(p0) is type(p2)
def gcs_path(*relative_path: type_utils.PathLike) -> type_utils.ReadWritePath:
"""Returns the GCS URI path.
Args:
*relative_path: Eventual relative path in the bucket.
Returns:
path: The GCS uri.
"""
return generic_path.as_path(GCS_ROOT_DIR).joinpath(*relative_path)
def test_windows_encoding():
with mock.patch('os.name', 'nt'):
assert os.name == 'nt'
# On windows, paths should be `WindowsGPath`
path = generic_path.as_path('c:/Program Files/text.txt')
assert isinstance(path, gpath.WindowsGPath)
path = generic_path.as_path(pathlib.PosixPath('some_dir/abc'))
assert isinstance(path, gpath.WindowsGPath)
# Other `GPath` and `gs://` should be `PosixPurePath`
path = generic_path.as_path('gs://some_dir/abc')
assert not isinstance(path, gpath.WindowsGPath)
assert isinstance(path, gpath.PosixGPath)
path = generic_path.as_path(gpath.PosixGPath('some_dir/abc'))
assert not isinstance(path, gpath.WindowsGPath)
assert isinstance(path, gpath.PosixGPath)
def test_resource_path():
path = resource_utils.ResourcePath(make_zip_file())
assert isinstance(path, os.PathLike)
assert path.joinpath('b/c.txt').read_text() == 'content of c'
sub_dirs = list(path.joinpath('b').iterdir())
assert len(sub_dirs) == 3
for p in sub_dirs: # Childs should be `ResourcePath` instances
assert isinstance(p, resource_utils.ResourcePath)
# Forwarded to `as_path` keep the resource.
path = generic_path.as_path(path)
assert isinstance(path, resource_utils.ResourcePath)
assert path.joinpath() == path
assert path.joinpath('abc', 'def.txt').name == 'def.txt'
def resource_path(package: Union[str, types.ModuleType]) -> ReadOnlyPath:
"""Returns `importlib.resources.files`."""
path = importlib_resources.files(package) # pytype: disable=module-attr
if isinstance(path, pathlib.Path):
# TODO(tfds): To ensure compatibility with zipfile.Path, we should ensure
# that the returned `pathlib.Path` isn't missused. More specifically:
# * `os.fspath` should only be called on files (not directories)
# * `str(path)` should be forbidden (only `__format__` allowed).
# In practice, it is trickier to do as `__fspath__` and `__str__` are
# called internally.
# Convert to `GPath` for consistency and compatibility with `MockFs`.
return generic_path.as_path(path)
elif isinstance(path, zipfile.Path):
path = ResourcePath(path.root, path.at)
return typing.cast(ReadOnlyPath, path)
else:
raise TypeError(f'Unknown resource path: {type(path)}: {path}')
def __new__(cls: Type[T], *args: PathLike) -> T:
if cls in (ReadOnlyPath, ReadWritePath):
from tensorflow_datasets.core.utils import generic_path # pytype: disable=import-error # pylint: disable=g-import-not-at-top
return generic_path.as_path(*args)
else:
return super().__new__(cls, *args)