def test_url_to_path(url, win_expected, non_win_expected):
if sys.platform == 'win32':
expected_path = win_expected
else:
expected_path = non_win_expected
if expected_path is None:
with pytest.raises(ValueError):
url_to_path(url)
else:
assert url_to_path(url) == expected_path
def test_url_to_path(url: str, win_expected: str,
non_win_expected: str) -> None:
if sys.platform == "win32":
expected_path = win_expected
else:
expected_path = non_win_expected
if expected_path is None:
with pytest.raises(ValueError):
url_to_path(url)
else:
assert url_to_path(url) == expected_path
def get_file_content(url, session, comes_from=None):
# type: (str, PipSession, Optional[str]) -> Tuple[str, Text]
"""Gets the content of a file; it may be a filename, file: URL, or
http: URL. Returns (location, content). Content is unicode.
Respects # -*- coding: declarations on the retrieved files.
:param url: File path or url.
:param session: PipSession instance.
:param comes_from: Origin description of requirements.
"""
scheme = get_url_scheme(url)
if scheme in ['http', 'https']:
# FIXME: catch some errors
resp = session.get(url)
raise_for_status(resp)
return resp.url, resp.text
elif scheme == 'file':
if comes_from and comes_from.startswith('http'):
raise InstallationError(
'Requirements file {} references URL {}, '
'which is local'.format(comes_from, url)
)
url = url_to_path(url)
try:
with open(url, 'rb') as f:
content = auto_decode(f.read())
except IOError as exc:
raise InstallationError(
'Could not open requirements file: {}'.format(exc)
)
return url, content
def send(self,
request,
stream=None,
timeout=None,
verify=None,
cert=None,
proxies=None):
pathname = url_to_path(request.url)
resp = Response()
resp.status_code = 200
resp.url = request.url
try:
stats = os.stat(pathname)
except OSError as exc:
resp.status_code = 404
resp.raw = exc
else:
modified = email.utils.formatdate(stats.st_mtime, usegmt=True)
content_type = mimetypes.guess_type(pathname)[0] or "text/plain"
resp.headers = CaseInsensitiveDict({
"Content-Type": content_type,
"Content-Length": stats.st_size,
"Last-Modified": modified,
})
resp.raw = open(pathname, "rb")
resp.close = resp.raw.close
return resp
def install_req_from_editable(
editable_req, # type: str
comes_from=None, # type: Optional[str]
use_pep517=None, # type: Optional[bool]
isolated=False, # type: bool
options=None, # type: Optional[Dict[str, Any]]
wheel_cache=None, # type: Optional[WheelCache]
constraint=False # type: bool
):
# type: (...) -> InstallRequirement
name, url, extras_override = parse_editable(editable_req)
if url.startswith('file:'):
source_dir = url_to_path(url)
else:
source_dir = None
if name is not None:
try:
req = Requirement(name)
except InvalidRequirement:
raise InstallationError("Invalid requirement: '%s'" % name)
else:
req = None
return InstallRequirement(
req, comes_from, source_dir=source_dir,
editable=True,
link=Link(url),
constraint=constraint,
use_pep517=use_pep517,
isolated=isolated,
options=options if options else {},
wheel_cache=wheel_cache,
extras=extras_override or (),
)
def open_local_or_remote_file(link, session):
"""
Open local or remote file for reading.
:type link: pip.index.Link
:type session: requests.Session
:raises ValueError: If link points to a local directory.
:return: a context manager to a FileStream with the opened file-like object
"""
url = link.url_without_fragment
if link.is_file:
# Local URL
local_path = url_to_path(url)
if os.path.isdir(local_path):
raise ValueError(f"Cannot open directory for read: {url}")
else:
st = os.stat(local_path)
with open(local_path, "rb") as local_file:
yield FileStream(stream=local_file, size=st.st_size)
else:
# Remote URL
headers = {"Accept-Encoding": "identity"}
response = session.get(url, headers=headers, stream=True)
# Content length must be int or None
try:
content_length = int(response.headers["content-length"])
except (ValueError, KeyError, TypeError):
content_length = None
try:
yield FileStream(stream=response.raw, size=content_length)
finally:
response.close()
def send(
self,
request: PreparedRequest,
stream: bool = False,
timeout: Optional[Union[float, Tuple[float, float]]] = None,
verify: Union[bool, str] = True,
cert: Optional[Union[str, Tuple[str, str]]] = None,
proxies: Optional[Mapping[str, str]] = None,
) -> Response:
pathname = url_to_path(request.url)
resp = Response()
resp.status_code = 200
resp.url = request.url
try:
stats = os.stat(pathname)
except OSError as exc:
resp.status_code = 404
resp.raw = exc
else:
modified = email.utils.formatdate(stats.st_mtime, usegmt=True)
content_type = mimetypes.guess_type(pathname)[0] or "text/plain"
resp.headers = CaseInsensitiveDict(
{
"Content-Type": content_type,
"Content-Length": stats.st_size,
"Last-Modified": modified,
}
)
resp.raw = open(pathname, "rb")
resp.close = resp.raw.close
return resp
def group_locations(locations, expand_dir=False):
# type: (Sequence[str], bool) -> Tuple[List[str], List[str]]
"""
Divide a list of locations into two groups: "files" (archives) and "urls."
:return: A pair of lists (files, urls).
"""
files = []
urls = []
# puts the url for the given file path into the appropriate list
def sort_path(path):
# type: (str) -> None
url = path_to_url(path)
if mimetypes.guess_type(url, strict=False)[0] == 'text/html':
urls.append(url)
else:
files.append(url)
for url in locations:
is_local_path = os.path.exists(url)
is_file_url = url.startswith('file:')
if is_local_path or is_file_url:
if is_local_path:
path = url
else:
path = url_to_path(url)
if os.path.isdir(path):
if expand_dir:
path = os.path.realpath(path)
for item in os.listdir(path):
sort_path(os.path.join(path, item))
elif is_file_url:
urls.append(url)
else:
logger.warning(
"Path '%s' is ignored: it is a directory.",
path,
)
elif os.path.isfile(path):
sort_path(path)
else:
logger.warning(
"Url '%s' is ignored: it is neither a file "
"nor a directory.",
url,
)
elif is_url(url):
# Only add url with clear scheme
urls.append(url)
else:
logger.warning(
"Url '%s' is ignored. It is either a non-existing "
"path or lacks a specific scheme.",
url,
)
return files, urls
def send(self, request, stream=None, timeout=None, verify=None, cert=None,
proxies=None):
pathname = url_to_path(request.url)
resp = Response()
resp.status_code = 200
resp."""PipSession and supporting code, containing all pip-specific
def get_file_content(url, session):
# type: (str, PipSession) -> Tuple[str, str]
"""Gets the content of a file; it may be a filename, file: URL, or
http: URL. Returns (location, content). Content is unicode.
Respects # -*- coding: declarations on the retrieved files.
:param url: File path or url.
:param session: PipSession instance.
"""
scheme = get_url_scheme(url)
if scheme in ['http', 'https']:
# FIXME: catch some errors
resp = session.get(url)
raise_for_status(resp)
return resp.url, resp.text
elif scheme == 'file':
url = url_to_path(url)
try:
with open(url, 'rb') as f:
content = auto_decode(f.read())
except OSError as exc:
raise InstallationError(f'Could not open requirements file: {exc}')
return url, content
def build_source(
location: str,
*,
candidates_from_page: CandidatesFromPage,
page_validator: PageValidator,
expand_dir: bool,
cache_link_parsing: bool,
) -> Tuple[Optional[str], Optional[LinkSource]]:
path: Optional[str] = None
url: Optional[str] = None
if os.path.exists(location): # Is a local path.
url = path_to_url(location)
path = location
elif location.startswith("file:"): # A file: URL.
url = location
path = url_to_path(location)
elif is_url(location):
url = location
if url is None:
msg = (
"Location '%s' is ignored: "
"it is either a non-existing path or lacks a specific scheme."
)
logger.warning(msg, location)
return (None, None)
if path is None:
source: LinkSource = _RemoteFileSource(
candidates_from_page=candidates_from_page,
page_validator=page_validator,
link=Link(url, cache_link_parsing=cache_link_parsing),
)
return (url, source)
if os.path.isdir(path):
if expand_dir:
source = _FlatDirectorySource(
candidates_from_page=candidates_from_page,
path=path,
)
else:
source = _IndexDirectorySource(
candidates_from_page=candidates_from_page,
link=Link(url, cache_link_parsing=cache_link_parsing),
)
return (url, source)
elif os.path.isfile(path):
source = _LocalFileSource(
candidates_from_page=candidates_from_page,
link=Link(url, cache_link_parsing=cache_link_parsing),
)
return (url, source)
logger.warning(
"Location '%s' is ignored: it is neither a file nor a directory.",
location,
)
return (url, None)
def is_dir_url(link):
# type: (Link) -> bool
"""Return whether a file:// Link points to a directory.
``link`` must not have any other scheme but file://. Call is_file_url()
first.
"""
link_path = url_to_path(link.url_without_fragment)
return os.path.isdir(link_path)
def unpack_file_url(
link, # type: Link
location, # type: str
download_dir=None, # type: Optional[str]
hashes=None # type: Optional[Hashes]
):
# type: (...) -> None
"""Unpack link into location.
If download_dir is provided and link points to a file, make a copy
of the link file inside download_dir.
"""
link_path = url_to_path(link.url_without_fragment)
# If it's a url to a local directory
if is_dir_url(link):
if os.path.isdir(location):
rmtree(location)
_copy_source_tree(link_path, location)
if download_dir:
logger.info('Link is a directory, ignoring download_dir')
return
# If --require-hashes is off, `hashes` is either empty, the
# link's embedded hash, or MissingHashes; it is required to
# match. If --require-hashes is on, we are satisfied by any
# hash in `hashes` matching: a URL-based or an option-based
# one; no internet-sourced hash will be in `hashes`.
if hashes:
hashes.check_against_path(link_path)
# If a download dir is specified, is the file already there and valid?
already_downloaded_path = None
if download_dir:
already_downloaded_path = _check_download_dir(link, download_dir,
hashes)
if already_downloaded_path:
from_path = already_downloaded_path
else:
from_path = link_path
content_type = mimetypes.guess_type(from_path)[0]
# unpack the archive to the build dir location. even when only downloading
# archives, they have to be unpacked to parse dependencies
unpack_file(from_path, location, content_type, link)
# a download dir is specified and not already downloaded
if download_dir and not already_downloaded_path:
_copy_file(from_path, download_dir, link)
def find_all_candidates(self, project_name):
# type: (str) -> List[InstallationCandidate]
"""Find all available InstallationCandidate for project_name
This checks index_urls and find_links.
All versions found are returned as an InstallationCandidate list.
See LinkEvaluator.evaluate_link() for details on which files
are accepted.
"""
collected_links = self._link_collector.collect_links(project_name)
link_evaluator = self.make_link_evaluator(project_name)
find_links_versions = self.evaluate_links(
link_evaluator,
links=collected_links.find_links,
)
page_versions = []
for page_url, page_links in collected_links.pages.items():
logger.debug('Analyzing links from page %s', page_url)
with indent_log():
new_versions = self.evaluate_links(
link_evaluator,
links=page_links,
)
page_versions.extend(new_versions)
file_versions = self.evaluate_links(
link_evaluator,
links=collected_links.files,
)
if file_versions:
file_versions.sort(reverse=True)
logger.debug(
'Local files found: %s',
', '.join([
url_to_path(candidate.link.url)
for candidate in file_versions
])
)
# This is an intentional priority ordering
return file_versions + find_links_versions + page_versions
def find_all_candidates(self, project_name: str) -> List[InstallationCandidate]:
"""Find all available InstallationCandidate for project_name
This checks index_urls and find_links.
All versions found are returned as an InstallationCandidate list.
See LinkEvaluator.evaluate_link() for details on which files
are accepted.
"""
link_evaluator = self.make_link_evaluator(project_name)
collected_sources = self._link_collector.collect_sources(
project_name=project_name,
candidates_from_page=functools.partial(
self.process_project_url,
link_evaluator=link_evaluator,
),
)
page_candidates_it = itertools.chain.from_iterable(
source.page_candidates()
for sources in collected_sources
for source in sources
if source is not None
)
page_candidates = list(page_candidates_it)
file_links_it = itertools.chain.from_iterable(
source.file_links()
for sources in collected_sources
for source in sources
if source is not None
)
file_candidates = self.evaluate_links(
link_evaluator,
sorted(file_links_it, reverse=True),
)
if logger.isEnabledFor(logging.DEBUG) and file_candidates:
paths = [url_to_path(c.link.url) for c in file_candidates]
logger.debug("Local files found: %s", ", ".join(paths))
# This is an intentional priority ordering
return file_candidates + page_candidates
def find_all_candidates(self, project_name):
# type: (str) -> List[InstallationCandidate]
"""Find all available InstallationCandidate for project_name
This checks index_urls and find_links.
All versions found are returned as an InstallationCandidate list.
See LinkEvaluator.evaluate_link() for details on which files
are accepted.
"""
collected_links = self._link_collector.collect_links(project_name)
link_evaluator = self.make_link_evaluator(project_name)
find_links_versions = self.evaluate_links(
link_evaluator,
links=collected_links.find_links,
)
page_versions = []
for project_url in collected_links.project_urls:
package_links = self.process_project_url(
project_url,
link_evaluator=link_evaluator,
)
page_versions.extend(package_links)
file_versions = self.evaluate_links(
link_evaluator,
links=collected_links.files,
)
if file_versions:
file_versions.sort(reverse=True)
logger.debug(
"Local files found: %s",
", ".join([
url_to_path(candidate.link.url)
for candidate in file_versions
]),
)
# This is an intentional priority ordering
return file_versions + find_links_versions + page_versions
def editable_project_location(self) -> Optional[str]:
"""The project location for editable distributions.
This is the directory where pyproject.toml or setup.py is located.
None if the distribution is not installed in editable mode.
"""
# TODO: this property is relatively costly to compute, memoize it ?
direct_url = self.direct_url
if direct_url:
if direct_url.is_local_editable():
return url_to_path(direct_url.url)
else:
# Search for an .egg-link file by walking sys.path, as it was
# done before by dist_is_editable().
egg_link_path = egg_link_path_from_sys_path(self.raw_name)
if egg_link_path:
# TODO: get project location from second line of egg_link file
# (https://github.com/pypa/pip/issues/10243)
return self.location
return None
<<<<<<< HEAD
=======
:param comes_from: Origin description of requirements.
>>>>>>> 74c061954d5e927be4caafbd793e96a50563c265
"""
scheme = get_url_scheme(url)
if scheme in ['http', 'https']:
# FIXME: catch some errors
resp = session.get(url)
raise_for_status(resp)
return resp.url, resp.text
elif scheme == 'file':
<<<<<<< HEAD
url = url_to_path(url)
=======
if comes_from and comes_from.startswith('http'):
raise InstallationError(
'Requirements file {} references URL {}, '
'which is local'.format(comes_from, url)
)
path = url.split(':', 1)[1]
path = path.replace('\\', '/')
match = _url_slash_drive_re.match(path)
if match:
path = match.group(1) + ':' + path.split('|', 1)[1]
path = urllib_parse.unquote(path)
if path.startswith('/'):
path = '/' + path.lstrip('/')
def test_url_to_path_path_to_url_symmetry_win() -> None:
path = r"C:\tmp\file"
assert url_to_path(path_to_url(path)) == path
unc_path = r"\\unc\share\path"
assert url_to_path(path_to_url(unc_path)) == unc_path
def test_url_to_path_path_to_url_symmetry_win():
path = r'C:\tmp\file'
assert url_to_path(path_to_url(path)) == path
unc_path = r'\\unc\share\path'
assert url_to_path(path_to_url(unc_path)) == unc_path
def file_path(self):
# type: () -> str
return url_to_path(self.url)
def file_path(self) -> str:
return url_to_path(self.url)
def prepare_linked_requirement(
self,
req, # type: InstallRequirement
session, # type: PipSession
finder, # type: PackageFinder
require_hashes, # type: bool
):
# type: (...) -> AbstractDistribution
"""Prepare a requirement that would be obtained from req.link
"""
assert req.link
link = req.link
# TODO: Breakup into smaller functions
if link.scheme == 'file':
path = url_to_path(link.url)
logger.info('Processing %s', display_path(path))
else:
logger.info('Collecting %s', req.req or req)
with indent_log():
# @@ if filesystem packages are not marked
# editable in a req, a non deterministic error
# occurs when the script attempts to unpack the
# build directory
req.ensure_has_source_dir(self.build_dir)
# If a checkout exists, it's unwise to keep going. version
# inconsistencies are logged later, but do not fail the
# installation.
# FIXME: this won't upgrade when there's an existing
# package unpacked in `req.source_dir`
if os.path.exists(os.path.join(req.source_dir, 'setup.py')):
raise PreviousBuildDirError(
"pip can't proceed with requirements '%s' due to a"
" pre-existing build directory (%s). This is "
"likely due to a previous installation that failed"
". pip is being responsible and not assuming it "
"can delete this. Please delete it and try again."
% (req, req.source_dir)
)
# Now that we have the real link, we can tell what kind of
# requirements we have and raise some more informative errors
# than otherwise. (For example, we can raise VcsHashUnsupported
# for a VCS URL rather than HashMissing.)
if require_hashes:
# We could check these first 2 conditions inside
# unpack_url and save repetition of conditions, but then
# we would report less-useful error messages for
# unhashable requirements, complaining that there's no
# hash provided.
if link.is_vcs:
raise VcsHashUnsupported()
elif is_file_url(link) and is_dir_url(link):
raise DirectoryUrlHashUnsupported()
if not req.original_link and not req.is_pinned:
# Unpinned packages are asking for trouble when a new
# version is uploaded. This isn't a security check, but
# it saves users a surprising hash mismatch in the
# future.
#
# file:/// URLs aren't pinnable, so don't complain
# about them not being pinned.
raise HashUnpinned()
hashes = req.hashes(trust_internet=not require_hashes)
if require_hashes and not hashes:
# Known-good hashes are missing for this requirement, so
# shim it with a facade object that will provoke hash
# computation and then raise a HashMissing exception
# showing the user what the hash should be.
hashes = MissingHashes()
download_dir = self.download_dir
if link.is_wheel and self.wheel_download_dir:
# when doing 'pip wheel` we download wheels to a
# dedicated dir.
download_dir = self.wheel_download_dir
try:
unpack_url(
link, req.source_dir, download_dir,
session=session, hashes=hashes,
progress_bar=self.progress_bar
)
except requests.HTTPError as exc:
logger.critical(
'Could not install requirement %s because of error %s',
req,
exc,
)
raise InstallationError(
'Could not install requirement %s because of HTTP '
'error %s for URL %s' %
(req, exc, link)
)
if link.is_wheel:
if download_dir:
# When downloading, we only unpack wheels to get
# metadata.
autodelete_unpacked = True
else:
# When installing a wheel, we use the unpacked
# wheel.
autodelete_unpacked = False
else:
# We always delete unpacked sdists after pip runs.
autodelete_unpacked = True
if autodelete_unpacked:
write_delete_marker_file(req.source_dir)
abstract_dist = _get_prepared_distribution(
req, self.req_tracker, finder, self.build_isolation,
)
if self._download_should_save:
# Make a .zip of the source_dir we already created.
if not link.is_artifact:
req.archive(self.download_dir)
return abstract_dist
