def simple_package(package_name):
''' Given a package name, returns all the versions for downloading
that package.
If the package doesn't exists, then it will call PyPi (CheeseShop).
But if the package exists in the local path, then it will get all
the versions for the local package.
This will take into account if the egg is private or if it is a normal
egg that was uploaded to PyPi. This is important to take into account
the version of the eggs. For example, a proyect requires request==1.0.4
and another package uses request==1.0.3. Then the instalation of the
second package will fail because it wasn't downloaded an the **request**
folder only has the 1.0.4 version.
To solve this problem, the system uses 2 different kinds of eggs:
* private eggs: are the eggs that you uploaded to the private repo.
* normal eggs: are the eggs that are downloaded from pypi.
So the normal eggs will always get the simple page from the pypi repo,
will the private eggs will always be read from the filesystem.
:param package_name: the name of the egg package. This is only the
name of the package with the version or anything
else.
:return: a template with all the links to download the packages.
'''
app.logger.debug('Requesting index for: %s', package_name)
package_folder = get_package_path(package_name)
if (is_private(package_name) or (
exists(package_name) and app.config['SHOULD_USE_EXISTING'])):
app.logger.debug('Found information of package: %s in local repository',
package_name)
package_versions = []
template_data = dict(
source_letter=package_name[0],
package_name=package_name,
versions=package_versions
)
for filename in listdir(package_folder):
if not filename.endswith('.md5'):
# I only read .md5 files so I skip this egg (or tar,
# or zip) file
continue
with open(join(package_folder, filename)) as md5_file:
md5 = md5_file.read(-1)
# remove .md5 extension
name = filename[:-4]
data = VersionData(name, md5)
package_versions.append(data)
return render_template('simple_package.html', **template_data)
else:
app.logger.debug('Didnt found package: %s in local repository. '
'Using proxy.', package_name)
url = app.config['PYPI_URL'] + 'simple/%s' % package_name
response = get(url)
if response.status_code != 200:
app.logger.warning('Error while getting proxy info for: %s'
'Errors details: %s', package_name,
response.text)
abort(response.status_code)
content = response.content
p = PyQuery(content)
external_links = set()
for anchor in p("a"):
panchor = PyQuery(anchor)
href = panchor.attr('href')
# robin-jarry: modified the href to ../../packages/
# so that it works also for non-source packages (.egg, .exe and .msi)
parsed = urlparse.urlparse(href)
if parsed.hostname:
# the link is to an external server.
if parsed.hostname == 'pypi.python.org':
# we remove the hostname to make the URL relative
panchor.attr('href', parsed.path)
else:
if panchor.attr('rel') == 'download':
if url_is_egg_file(parsed.path):
# href points to a filename
external_links.add('<a href="%s">%s</a>' % (href, basename(parsed.path)))
else:
# href points to an external page where we will find
# links to package files
external_links.update(find_external_links(href))
# what ever happens, we remove the link for now
# we'll add the external_links after that we found after
panchor.remove()
else:
# local link to pypi.python.org
if not href.startswith('../../packages/'):
# ignore anything else than package links
panchor.remove()
# after collecting all external links, we insert them in the html page
for link in external_links:
plink = PyQuery(link)
href = plink.attr('href')
plink.attr('href', convert_to_internal_url(href, package_name, basename(href)))
p('a').after(plink)
content = p.outerHtml()
return content
def simple_package(package_name):
''' Given a package name, returns all the versions for downloading
that package.
If the package doesn't exists, then it will call PyPi (CheeseShop).
But if the package exists in the local path, then it will get all
the versions for the local package.
This will take into account if the egg is private or if it is a normal
egg that was uploaded to PyPi. This is important to take into account
the version of the eggs. For example, a proyect requires request==1.0.4
and another package uses request==1.0.3. Then the instalation of the
second package will fail because it wasn't downloaded an the **request**
folder only has the 1.0.4 version.
To solve this problem, the system uses 2 different kinds of eggs:
* private eggs: are the eggs that you uploaded to the private repo.
* normal eggs: are the eggs that are downloaded from pypi.
So the normal eggs will always get the simple page from the pypi repo,
will the private eggs will always be read from the filesystem.
:param package_name: the name of the egg package. This is only the
name of the package with the version or anything
else.
:return: a template with all the links to download the packages.
'''
app.logger.debug('Requesting index for: %s', package_name)
package_folder = get_package_path(package_name)
if (is_private(package_name) or (
exists(package_name) and app.config['SHOULD_USE_EXISTING'])):
app.logger.debug('Found information of package: %s in local repository',
package_name)
package_versions = []
template_data = dict(
source_letter=package_name[0],
package_name=package_name,
versions=package_versions
)
for filename in listdir(package_folder):
if not filename.endswith('.md5'):
# I only read .md5 files so I skip this egg (or tar,
# or zip) file
continue
with open(join(package_folder, filename)) as md5_file:
md5 = md5_file.read(-1)
# remove .md5 extension
name = filename[:-4]
data = VersionData(name, md5, None)
package_versions.append(data)
return render_template('simple_package.html', **template_data)
else:
app.logger.debug('Didnt found package: %s in local repository. '
'Using proxy.', package_name)
url = app.config['PYPI_URL'] + 'simple/%s/' % package_name
response = get(url)
if response.status_code != 200:
app.logger.warning('Error while getting proxy info for: %s'
'Errors details: %s', package_name,
response.text)
abort(response.status_code)
if response.history:
app.logger.debug('The url was redirected')
# in this case, the request was redirect, so I should also
# take into account this change. For example, this happens
# when requesting flask-bcrypt and on Pypi the request is
# redirected to Flask-Bcrypt
package_name = urlparse.urlparse(response.url).path
package_name = package_name.replace('/simple/', '')
package_name = package_name.replace('/', '')
content = response.content
external_links = set()
# contains the list of pacges whih where checked because
# on the link they had the information of
visited_download_pages = set()
soup = BeautifulSoup(content)
package_versions = []
for panchor in soup.find_all('a'):
if panchor.get('rel') and panchor.get('rel')[0] == 'homepage':
# skip getting information on the project homepage
continue
href = panchor.get('href')
app.logger.debug('Found the link: %s', panchor.get('href'))
if href.startswith('../../packages/'):
# then the package is hosted on pypi.
pk_name = basename(href)
pk_name, md5_data = pk_name.split('#md5=')
pk_name = pk_name.replace('#md5=', '')
data = VersionData(pk_name, md5_data, None)
package_versions.append(data)
continue
parsed = urlparse.urlparse(href)
if parsed.hostname:
# then the package had a full path to the file
if parsed.hostname == 'pypi.python.org':
# then it is hosted on the pypi server, so I change
# it to make it a relative url
pk_name = basename(parsed.path)
if '#md5=' in parsed.path:
pk_name, md5_data = pk_name.split('#md5=')
pk_name = pk_name.replace('#md5=', '')
else:
md5_data = ''
data = VersionData(pk_name, md5_data, None)
package_versions.append(data)
else:
# the python package is hosted on another server
# that isn't pypi. The packages that doesn't have
# rel=download, then they are links to some pages
if panchor.get('rel') and panchor.get('rel')[0] == 'download':
if url_is_egg_file(parsed.path):
external_links.add(href)
else:
# href point to an external page where the links
# to download the package will be found
if href not in visited_download_pages:
visited_download_pages.add(href)
external_links.update(find_external_links(href))
# after collecting all external links, we insert them in the html page
for external_url in external_links:
package_version = basename(external_url)
existing_value = filter(lambda pv: pv.name == package_version,
package_versions)
external_link = urllib.urlencode({'remote': external_url})
if existing_value:
package_versions.remove(existing_value[0])
# check that the package version doens't override the one that
# already exists on pypi
existing_data = filter(lambda v: v.name == package_version,
package_versions)
if existing_data:
continue
data = VersionData(package_version, '', external_link)
package_versions.append(data)
template_data = dict(
source_letter=package_name[0],
package_name=package_name,
versions=package_versions
)
return render_template('simple_package.html', **template_data)
def simple_package(package_name):
''' Given a package name, returns all the versions for downloading
that package.
If the package doesn't exists, then it will call PyPI (CheeseShop).
But if the package exists in the local path, then it will get all
the versions for the local package.
This will take into account if the egg is private or if it is a normal
egg that was uploaded to PyPI. This is important to take into account
the version of the eggs. For example, a project requires requests==1.0.4
and another package uses requests==1.0.3. Then the instalation of the
second package will fail because it wasn't downloaded and the **requests**
folder only has the 1.0.4 version.
To solve this problem, the system uses 2 different kinds of eggs:
* private eggs: are the eggs that you uploaded to the private repo.
* normal eggs: are the eggs that are downloaded from PyPI.
So the normal eggs will always get the simple page from the PyPI repo,
will the private eggs will always be read from the filesystem.
:param package_name: the name of the egg package. This is only the
name of the package with the version or anything
else.
:return: a template with all the links to download the packages.
'''
app.logger.debug('Requesting index for: %s', package_name)
package_folder = get_package_path(package_name)
if (is_private(package_name)
or (exists(package_folder) and app.config['SHOULD_USE_EXISTING'])):
app.logger.debug(
'Found information of package: %s in local repository',
package_name)
package_versions = []
template_data = dict(source_letter=package_name[0],
package_name=package_name,
versions=package_versions)
for filename in listdir(package_folder):
if not filename.endswith('.md5'):
# I only read .md5 files so I skip this egg (or tar,
# or zip) file
continue
with open(join(package_folder, filename)) as md5_file:
md5 = md5_file.read(-1)
# remove .md5 extension
name = filename[:-4]
data = VersionData('source', name, md5, None)
package_versions.append(data)
return render_template('simple_package.html', **template_data)
else:
app.logger.debug(
'Didnt found package: %s in local repository. '
'Using proxy.', package_name)
url = app.config['PYPI_URL'] + 'simple/%s/' % package_name
response = get(url)
if response.status_code != 200:
app.logger.warning(
'Error while getting proxy info for: %s'
'Errors details: %s', package_name, response.text)
abort(response.status_code)
if response.history:
app.logger.debug('The url was redirected')
# in this case, the request was redirect, so I should also
# take into account this change. For example, this happens
# when requesting flask-bcrypt and on Pypi the request is
# redirected to Flask-Bcrypt
package_name = urlparse.urlparse(response.url).path
package_name = package_name.replace('/simple/', '')
package_name = package_name.replace('/', '')
content = response.content
external_links = set()
# contains the list of pacges whih where checked because
# on the link they had the information of
visited_download_pages = set()
soup = BeautifulSoup(content)
package_versions = []
for panchor in soup.find_all('a'):
if panchor.get('rel') and panchor.get('rel')[0] == 'homepage':
# skip getting information on the project homepage
continue
href = panchor.get('href')
app.logger.debug('Found the link: %s', href)
if href.startswith('../../packages/'):
# then the package is hosted on PyPI.
pk_name = basename(href)
pk_name, md5_data = pk_name.split('#md5=')
pk_name = pk_name.replace('#md5=', '')
# remove md5 part to make the url shorter.
split_data = urlparse.urlsplit(href)
absolute_url = urlparse.urljoin(url, split_data.path)
external_link = urllib.urlencode({'remote': absolute_url})
pk_type = href.split('/')[3]
data = VersionData(pk_type, pk_name, md5_data, external_link)
package_versions.append(data)
app.logger.debug('pk_type: %s', pk_type)
continue
parsed = urlparse.urlparse(href)
if parsed.hostname:
# then the package had a full path to the file
if parsed.hostname == 'pypi.python.org':
# then it is hosted on the PyPI server, so I change
# it to make it a relative url
pk_name = basename(parsed.path)
if '#md5=' in parsed.path:
pk_name, md5_data = pk_name.split('#md5=')
pk_name = pk_name.replace('#md5=', '')
else:
md5_data = ''
absolute_url = urlparse.urljoin(url, parsed.path)
external_link = urllib.urlencode({'remote': absolute_url})
pk_type = parsed.path.split('/')[2]
data = VersionData(pk_type, pk_name, md5_data,
external_link)
package_versions.append(data)
else:
# the python package is hosted on another server
# that isn't PyPI. The packages that don't have
# rel=download are links to some pages
if panchor.get('rel') and panchor.get(
'rel')[0] == 'download':
if url_is_egg_file(parsed.path):
external_links.add(href)
else:
# href points to an external page where the links
# to download the package will be found
if href not in visited_download_pages:
visited_download_pages.add(href)
external_links.update(
find_external_links(href))
# after collecting all external links, we insert them in the html page
for external_url in external_links:
package_version = basename(external_url)
existing_value = filter(lambda pv: pv.name == package_version,
package_versions)
if existing_value:
# if the package already exists on PyPI, then
# use its version instead of using the one that is
# hosted on a remote server
continue
external_link = urllib.urlencode({'remote': external_url})
data = VersionData('source', package_version, '', external_link)
package_versions.append(data)
package_versions.sort(key=lambda v: v.name)
template_data = dict(source_letter=package_name[0],
package_name=package_name,
versions=package_versions)
return render_template('simple_package.html', **template_data)