def find_eggs_in_url(url):
"""Read a URL and find any links to egg files
Parameters
----------
url : string
the url to search
Returns
-------
dists : a list of pkg_resources.Distribution objects
"""
if url in cache:
return cache[url]
page = urlopen(url)
try:
info = page.info()
content = page.read()
finally:
page.close()
if info['Content-Type'].split(';')[0].strip() == 'text/html':
dists = []
for match in HREF.finditer(content):
ref = match.group(1)
schema, location, path, query, frag = urlsplit(ref)
if location == '':
ref = urljoin(url, ref)
dists += distros_for_url(ref)
cache[url] = dists
return dists
def process_url(self, url, retrieve=False):
"""Evaluate a URL as a possible download, and maybe retrieve it"""
if url in self.scanned_urls and not retrieve:
return
self.scanned_urls[url] = True
if not URL_SCHEME(url):
self.process_filename(url)
return
else:
dists = list(distros_for_url(url))
if dists:
if not self.url_ok(url):
return
self.debug("Found link: %s", url)
if dists or not retrieve or url in self.fetched_urls:
map(self.add, dists)
return # don't need the actual page
if not self.url_ok(url):
self.fetched_urls[url] = True
return
self.info("Reading %s", url)
f = self.open_url(url, "Download error on %s: %%s -- Some packages may not be found!" % url)
if f is None:
return
self.fetched_urls[url] = self.fetched_urls[f.url] = True
if "html" not in f.headers.get("content-type", "").lower():
f.close() # not html, we can't process it
return
base = f.url # handle redirects
page = f.read()
if not isinstance(page, str): # We are in Python 3 and got bytes. We want str.
if isinstance(f, urllib2.HTTPError):
# Errors have no charset, assume latin1:
charset = "latin-1"
else:
charset = f.headers.get_param("charset") or "latin-1"
page = page.decode(charset, "ignore")
f.close()
for match in HREF.finditer(page):
link = urlparse.urljoin(base, htmldecode(match.group(1)))
self.process_url(link)
for index_url in self.index_urls:
if url.startswith(index_url) and getattr(f, "code", None) != 404:
page = self.process_index(url, page)
def compute_version(filename):
match = WHEEL_RE.match(filename)
if match:
return match.group("ver")
try:
distro = next(distros_for_url(filename))
except StopIteration:
logger.info({
"event": "download_statitics.compute_version.ignore",
"filename": filename
})
return None
else:
return distro.version
def process_url(self, url, retrieve=False):
"""Evaluate a URL as a possible download, and maybe retrieve it"""
if url in self.scanned_urls and not retrieve:
return
self.scanned_urls[url] = True
if not URL_SCHEME(url):
self.process_filename(url)
return
else:
dists = list(distros_for_url(url))
if dists:
if not self.url_ok(url):
return
self.debug("Found link: %s", url)
if dists or not retrieve or url in self.fetched_urls:
list(map(self.add, dists))
return # don't need the actual page
if not self.url_ok(url):
self.fetched_urls[url] = True
return
self.info("Reading %s", url)
self.fetched_urls[url] = True # prevent multiple fetch attempts
tmpl = "Download error on %s: %%s -- Some packages may not be found!"
f = self.open_url(url, tmpl % url)
if f is None:
return
if isinstance(f, urllib.error.HTTPError) and f.code == 401:
self.info("Authentication error: %s" % f.msg)
self.fetched_urls[f.url] = True
if 'html' not in f.headers.get('content-type', '').lower():
f.close() # not html, we can't process it
return
base = f.url # handle redirects
page = f.read()
# --- LOCAL CHANGES MADE HERE: ---
if isinstance(page, six.text_type):
page = page.encode('utf8')
charset = 'utf8'
else:
if isinstance(f, urllib.error.HTTPError):
# Errors have no charset, assume latin1:
charset = 'latin-1'
else:
try:
charset = f.headers.get_param('charset') or 'latin-1'
except AttributeError:
# Python 2
charset = f.headers.getparam('charset') or 'latin-1'
try:
html_page = HTMLPage(page, charset, base, cache_link_parsing=False)
except TypeError:
html_page = HTMLPage(page, charset, base)
# https://github.com/buildout/buildout/issues/598
# use_deprecated_html5lib is a required addition in pip 22.
try:
plinks = parse_links(html_page, use_deprecated_html5lib=False)
except TypeError:
plinks = parse_links(html_page)
plinks = list(plinks)
pip_links = [l.url for l in plinks]
# --- END OF LOCAL CHANGES ---
if not isinstance(page, str):
# In Python 3 and got bytes but want str.
page = page.decode(charset, "ignore")
f.close()
# --- LOCAL CHANGES MADE HERE: ---
links = []
for match in HREF.finditer(page):
link = urllib.parse.urljoin(base, htmldecode(match.group(1)))
links.append(_clean_link(link))
# TODO: remove assertion and double index page parsing before releasing.
assert set(pip_links) == set(links)
for link in plinks:
if _check_link_requires_python(link, PY_VERSION_INFO):
self.process_url(link.url)
# --- END OF LOCAL CHANGES ---
if url.startswith(self.index_url) and getattr(f, 'code', None) != 404:
page = self.process_index(url, page)
class Package(models.Model):
index = models.ForeignKey(PackageIndex)
name = models.CharField(max_length=255, unique=True, primary_key=True)
auto_hide = models.BooleanField(default=True, blank=False)
updated_from_remote_at = models.DateTimeField(null=True, blank=True)
parsed_external_links_at = models.DateTimeField(null=True, blank=True)
class Meta:
verbose_name = _(u"package")
verbose_name_plural = _(u"packages")
get_latest_by = "releases__latest"
ordering = [
'name',
]
def __unicode__(self):
return self.name
@models.permalink
def get_absolute_url(self):
return ('packageindex-package', (), {'package': self.name})
@property
def latest(self):
try:
return self.releases.latest()
except Release.DoesNotExist:
return None
def get_release(self, version):
"""Return the release object for version, or None"""
try:
return self.releases.get(version=version)
except Release.DoesNotExist:
return None
def update_release_metadata(self, update_distribution_metadata=True):
now = datetime.datetime.now()
try:
name = self.name.encode('ascii')
except UnicodeEncodeError:
print "illegal package name!"
return
for release_string in self.index.client.package_releases(
self.name, True): # True -> show hidden
data = self.index.client.release_data(self.name, release_string)
kwargs = {
'hidden': data.get('_pypi_hidden', False),
'package_info': MultiValueDict(),
'is_from_external': False,
}
for key, value in data.items():
kwargs['package_info'][key] = value
release, created = Release.objects.get_or_create(
package=self, version=release_string, defaults=kwargs)
if not created:
for key, value in kwargs.items():
setattr(release, key, value)
release.save()
if update_distribution_metadata:
release.update_distribution_metatdata()
self.updated_from_remote_at = now
self.save()
def update_external_release_metadata(self,
update_distribution_metadata=True):
try:
name = self.name.encode('ascii')
except UnicodeEncodeError:
print "illegal package name!"
return
mpackage = mirror.Package(package_name=name,
pypi_base_url=self.index.simple_url)
try:
files = mpackage.ls(filename_matches='*',
external_links=True,
follow_external_index_pages=True)
except (PackageError, ), e:
print type(e), e
files = []
for (dist_url, file_name, md5sum) in files:
if dist_url.startswith('../../'):
# Ignore relative urls, as they are files hosted on pypi and have already been fetched over the xml-rpc
# api
continue
i = 1
for dist in distros_for_url(dist_url):
if not dist.project_name == self.name or not dist.version:
continue
release = Release.objects.get_or_create(
package=self,
version=dist.version,
defaults={'is_from_external': True})[0]
pyversion = dist.py_version or 'any'
f, ext = os.path.splitext(file_name)
if ext.startswith('.egg'):
filetype = 'bdist_egg'
elif ext in ('.exe', ):
filetype = 'bdist_wininst'
elif ext in ('.dmg', '.pgk'):
filetype = 'bdist_dmg'
elif ext in ('.rpm', ):
filetype = 'bdist_rpm'
elif ext in ('.tar.gz', '.zip', '.bz2'):
filetype = 'sdist'
else:
continue
defaults = {
'filename': file_name,
'url': dist_url,
'is_from_external': True
}
distribution = Distribution.objects.get_or_create(
release=release,
pyversion=pyversion,
filetype=filetype,
defaults=defaults)[0]
if distribution.is_from_external and not distribution.file:
# we only overwrite the url if the package has not been mirrored yet and it is not a real pypi
# hosted package
distribution.filename = file_name
distribution.url = dist_url
distribution.save()
print i, dist.project_name, dist.py_version, dist.version, distribution
i += 1
self.parsed_external_links_at = datetime.datetime.now()
self.save()
def version_for_url(project, url):
normalized = safe_name(project).lower()
return [dist for dist in distros_for_url(url) if safe_name(dist.project_name).lower() == normalized][0].version
def installable(project, url):
normalized = safe_name(project).lower()
return bool([dist for dist in distros_for_url(url) if safe_name(dist.project_name).lower() == normalized])
def get_distro(url):
return next(distros_for_url(url))
def version_for_url(project, url):
normalized = safe_name(project).lower()
return [
dist for dist in distros_for_url(url)
if safe_name(dist.project_name).lower() == normalized
][0].version
def installable(project, url):
normalized = safe_name(project).lower()
return bool([
dist for dist in distros_for_url(url)
if safe_name(dist.project_name).lower() == normalized
])
def get_distro(url):
return next(distros_for_url(url))
