def iter_version_links(html, name):
"""
Iterate through version links (in order) within HTML.
Filtering out links that don't "look" like versions.
Either yields hrefs to be recursively searches or tuples of (name, href)
that match the given name.
"""
soup = BeautifulSoup(html)
for node in soup.findAll("a"):
if node.get("href") is None:
continue
try:
guessed_name, _ = guess_name_and_version(node.text)
except ValueError:
href = node["href"]
for extension in [".tar.gz", ".zip"]:
if href.endswith(extension):
yield basename(href), href
break
else:
if node.get("rel") == "download":
# Might be a recursive link.
yield href
# else couldn't parse name and version, probably the wrong kind of link
else:
if guessed_name.replace("_", "-").lower() != name.replace("_", "-").lower():
continue
yield node.text, node["href"]
def _get_metadata(self, path, filename):
metadata = read_metadata(path)
# make sure metadata validates
if not self.validate_metadata(metadata):
abort(codes.bad_request)
# make sure nothing fishy is going on
if metadata.get(Version.FILENAME) not in [None, filename]:
abort(codes.bad_request)
# make sure metadata is consistent with filename
expected_name, expected_version = guess_name_and_version(filename)
if metadata["name"] != expected_name or metadata["version"] != expected_version:
self.logger.warn("Aborting upload of: {}; conflicting filename and metadata".format(filename))
abort(codes.bad_request)
# include local path in metadata
metadata[Version.FILENAME] = filename
# add upload timestamp
metadata["_uploaded_timestamp"] = time()
return metadata
def validate_guess(basename, expected_name, expected_version):
name, version = guess_name_and_version(basename)
eq_(name, expected_name)
eq_(version, expected_version)
