def parse_path(path):
"""Parse a dataset's identifier or path into its parts
Parameters
----------
path : str or path-like object
The path to be parsed.
Returns
-------
ParsedPath or UnparsedPath
Notes
-----
When legacy GDAL filenames are encountered, they will be returned
in a UnparsedPath.
"""
# Windows drive letters (e.g. "C:\") confuse `urlparse` as they look like
# URL schemes
if sys.platform == "win32" and re.match("^[a-zA-Z]\\:", path):
return UnparsedPath(path)
elif path.startswith('/vsi'):
return UnparsedPath(path)
else:
parts = urlparse(path)
# if the scheme is not one of Rasterio's supported schemes, we
# return an UnparsedPath.
if parts.scheme and not all(p in SCHEMES for p in parts.scheme.split('+')):
return UnparsedPath(path)
else:
return ParsedPath.from_uri(path)
def parse_path(path):
"""Parse a dataset's identifier or path into its parts
Parameters
----------
path : str or path-like object
The path to be parsed.
Returns
-------
ParsedPath or UnparsedPath
Notes
-----
When legacy GDAL filenames are encountered, they will be returned
in a UnparsedPath.
"""
# Windows drive letters (e.g. "C:\") confuse `urlparse` as they look like
# URL schemes
if sys.platform == "win32" and re.match("^[a-zA-Z]\\:", path):
return UnparsedPath(path)
elif path.startswith('/vsi'):
return UnparsedPath(path)
else:
parts = urlparse(path)
# if the scheme is not one of Rasterio's supported schemes, we
# return an UnparsedPath.
if parts.scheme and not all(p in SCHEMES for p in parts.scheme.split('+')):
return UnparsedPath(path)
else:
return ParsedPath.from_uri(path)
def parse_path(uri, vfs=None):
"""Parse a URI or Apache VFS URL into its parts
Returns: tuple
(path, archive, scheme)
"""
archive = scheme = None
path = uri
if vfs:
parts = urlparse(vfs)
scheme = parts.scheme
archive = parts.path
if parts.netloc and parts.netloc != 'localhost': # pragma: no cover
archive = parts.netloc + archive
else:
parts = urlparse(path)
scheme = parts.scheme
path = parts.path
if parts.query:
path += "?" + parts.query
if parts.netloc and parts.netloc != 'localhost':
path = parts.netloc + path
# There are certain URI schemes we favor over GDAL's names.
if scheme in SCHEMES:
parts = path.split('!')
path = parts.pop() if parts else None
archive = parts.pop() if parts else None
# For filesystem paths.
elif scheme.lower() in FILE_SCHEMES:
pass
# We permit GDAL's idiosyncratic URI-like dataset paths such as
# 'netcdf':... to fall right through with no parsed archive
# or scheme.
else:
archive = scheme = None
path = uri
return path, archive, scheme
def parse_path(uri, vfs=None):
"""Parse a URI or Apache VFS URL into its parts
Returns: tuple
(path, archive, scheme)
"""
archive = scheme = None
path = uri
if vfs:
parts = urlparse(vfs)
scheme = parts.scheme
archive = parts.path
if parts.netloc and parts.netloc != 'localhost': # pragma: no cover
archive = parts.netloc + archive
else:
parts = urlparse(path)
scheme = parts.scheme
path = parts.path
if parts.query:
path += "?" + parts.query
if parts.netloc and parts.netloc != 'localhost':
path = parts.netloc + path
# There are certain URI schemes we favor over GDAL's names.
if scheme in SCHEMES:
parts = path.split('!')
path = parts.pop() if parts else None
archive = parts.pop() if parts else None
# For filesystem paths.
elif scheme.lower() in FILE_SCHEMES:
pass
# We permit GDAL's idiosyncratic URI-like dataset paths such as
# 'netcdf':... to fall right through with no parsed archive
# or scheme.
else:
archive = scheme = None
path = uri
return path, archive, scheme
def parse_path(path):
"""Parse a dataset's identifier or path into its parts
Parameters
----------
path : str or path-like object
The path to be parsed.
Returns
-------
ParsedPath or UnparsedPath
Notes
-----
When legacy GDAL filenames are encountered, they will be returned
in a UnparsedPath.
"""
if isinstance(path, Path):
return path
elif pathlib and isinstance(path, pathlib.PurePath):
return ParsedPath(path.as_posix(), None, None)
elif isinstance(path, string_types):
if sys.platform == "win32" and re.match(r"^[a-zA-Z]\:", path):
if pathlib:
return ParsedPath(pathlib.Path(path).as_posix(), None, None)
else:
return UnparsedPath(path)
elif path.startswith('/vsi'):
return UnparsedPath(path)
else:
parts = urlparse(path)
else:
raise PathError("invalid path '{!r}'".format(path))
# if the scheme is not one of Rasterio's supported schemes, we
# return an UnparsedPath.
if parts.scheme:
if all(p in SCHEMES for p in parts.scheme.split('+')):
return ParsedPath.from_uri(path)
return UnparsedPath(path)
def from_uri(cls, uri):
parts = urlparse(uri)
path = parts.path
scheme = parts.scheme or None
if parts.query:
path += "?" + parts.query
if parts.scheme and parts.netloc:
path = parts.netloc + path
parts = path.split('!')
path = parts.pop() if parts else None
archive = parts.pop() if parts else None
return ParsedPath(path, archive, scheme)
def from_uri(cls, uri):
parts = urlparse(uri)
path = parts.path
scheme = parts.scheme or None
if parts.query:
path += "?" + parts.query
if parts.scheme and parts.netloc:
path = parts.netloc + path
parts = path.split('!')
path = parts.pop() if parts else None
archive = parts.pop() if parts else None
return ParsedPath(path, archive, scheme)
def parse_path(path, vfs=None):
"""Parse a dataset's path into its parts
**DEPRECATED**
Parameters
----------
path : str
The path or filename to be parsed.
vfs : str, optional **DEPRECATED**
A virtual file system path.
Returns
-------
path, archive, scheme : str
Parts of the parsed path.
"""
warnings.warn(
"This function will be removed in version 1.0",
RasterioDeprecationWarning
)
if vfs:
parts = urlparse(vfs)
scheme = parts.scheme
archive = parts.path
if parts.netloc and parts.netloc != 'localhost': # pragma: no cover
archive = parts.netloc + archive
parsed = ParsedPath(path, archive, scheme)
return parsed.path, parsed.archive, parsed.scheme
else:
parsed = future_parse_path(path)
if isinstance(parsed, ParsedPath):
return parsed.path, parsed.archive, parsed.scheme
else:
return parsed.path, None, None
