mirror of
https://github.com/SickGear/SickGear.git
synced 2024-12-24 03:33:38 +00:00
271 lines
11 KiB
Python
271 lines
11 KiB
Python
|
# The public API for feedparser
|
||
|
# Copyright 2010-2020 Kurt McKee <contactme@kurtmckee.org>
|
||
|
# Copyright 2002-2008 Mark Pilgrim
|
||
|
# All rights reserved.
|
||
|
#
|
||
|
# This file is a part of feedparser.
|
||
|
#
|
||
|
# Redistribution and use in source and binary forms, with or without modification,
|
||
|
# are permitted provided that the following conditions are met:
|
||
|
#
|
||
|
# * Redistributions of source code must retain the above copyright notice,
|
||
|
# this list of conditions and the following disclaimer.
|
||
|
# * Redistributions in binary form must reproduce the above copyright notice,
|
||
|
# this list of conditions and the following disclaimer in the documentation
|
||
|
# and/or other materials provided with the distribution.
|
||
|
#
|
||
|
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS 'AS IS'
|
||
|
# AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
|
||
|
# IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
|
||
|
# ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
|
||
|
# LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
|
||
|
# CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
|
||
|
# SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
|
||
|
# INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
|
||
|
# CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
|
||
|
# ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
|
||
|
# POSSIBILITY OF SUCH DAMAGE.
|
||
|
|
||
|
import io
|
||
|
import urllib.parse
|
||
|
import xml.sax
|
||
|
|
||
|
import sgmllib3k as sgmllib
|
||
|
|
||
|
from .datetimes import registerDateHandler, _parse_date
|
||
|
from .encodings import convert_to_utf8
|
||
|
from .exceptions import *
|
||
|
from .html import _BaseHTMLProcessor
|
||
|
from . import http
|
||
|
from . import mixin
|
||
|
from .mixin import _FeedParserMixin
|
||
|
from .parsers.loose import _LooseFeedParser
|
||
|
from .parsers.strict import _StrictFeedParser
|
||
|
from .sanitizer import replace_doctype
|
||
|
from .urls import convert_to_idn, make_safe_absolute_uri
|
||
|
from .util import FeedParserDict
|
||
|
|
||
|
|
||
|
# List of preferred XML parsers, by SAX driver name. These will be tried first,
|
||
|
# but if they're not installed, Python will keep searching through its own list
|
||
|
# of pre-installed parsers until it finds one that supports everything we need.
|
||
|
PREFERRED_XML_PARSERS = ["drv_libxml2"]
|
||
|
|
||
|
_XML_AVAILABLE = True
|
||
|
|
||
|
SUPPORTED_VERSIONS = {
|
||
|
'': 'unknown',
|
||
|
'rss090': 'RSS 0.90',
|
||
|
'rss091n': 'RSS 0.91 (Netscape)',
|
||
|
'rss091u': 'RSS 0.91 (Userland)',
|
||
|
'rss092': 'RSS 0.92',
|
||
|
'rss093': 'RSS 0.93',
|
||
|
'rss094': 'RSS 0.94',
|
||
|
'rss20': 'RSS 2.0',
|
||
|
'rss10': 'RSS 1.0',
|
||
|
'rss': 'RSS (unknown version)',
|
||
|
'atom01': 'Atom 0.1',
|
||
|
'atom02': 'Atom 0.2',
|
||
|
'atom03': 'Atom 0.3',
|
||
|
'atom10': 'Atom 1.0',
|
||
|
'atom': 'Atom (unknown version)',
|
||
|
'cdf': 'CDF',
|
||
|
}
|
||
|
|
||
|
|
||
|
def _open_resource(url_file_stream_or_string, etag, modified, agent, referrer, handlers, request_headers, result):
|
||
|
"""URL, filename, or string --> stream
|
||
|
|
||
|
This function lets you define parsers that take any input source
|
||
|
(URL, pathname to local or network file, or actual data as a string)
|
||
|
and deal with it in a uniform manner. Returned object is guaranteed
|
||
|
to have all the basic stdio read methods (read, readline, readlines).
|
||
|
Just .close() the object when you're done with it.
|
||
|
|
||
|
If the etag argument is supplied, it will be used as the value of an
|
||
|
If-None-Match request header.
|
||
|
|
||
|
If the modified argument is supplied, it can be a tuple of 9 integers
|
||
|
(as returned by gmtime() in the standard Python time module) or a date
|
||
|
string in any format supported by feedparser. Regardless, it MUST
|
||
|
be in GMT (Greenwich Mean Time). It will be reformatted into an
|
||
|
RFC 1123-compliant date and used as the value of an If-Modified-Since
|
||
|
request header.
|
||
|
|
||
|
If the agent argument is supplied, it will be used as the value of a
|
||
|
User-Agent request header.
|
||
|
|
||
|
If the referrer argument is supplied, it will be used as the value of a
|
||
|
Referer[sic] request header.
|
||
|
|
||
|
If handlers is supplied, it is a list of handlers used to build a
|
||
|
urllib2 opener.
|
||
|
|
||
|
if request_headers is supplied it is a dictionary of HTTP request headers
|
||
|
that will override the values generated by FeedParser.
|
||
|
|
||
|
:return: A bytes object.
|
||
|
"""
|
||
|
|
||
|
if hasattr(url_file_stream_or_string, 'read'):
|
||
|
return url_file_stream_or_string.read()
|
||
|
|
||
|
if isinstance(url_file_stream_or_string, str) \
|
||
|
and urllib.parse.urlparse(url_file_stream_or_string)[0] in ('http', 'https', 'ftp', 'file', 'feed'):
|
||
|
return http.get(url_file_stream_or_string, etag, modified, agent, referrer, handlers, request_headers, result)
|
||
|
|
||
|
# try to open with native open function (if url_file_stream_or_string is a filename)
|
||
|
try:
|
||
|
with open(url_file_stream_or_string, 'rb') as f:
|
||
|
data = f.read()
|
||
|
except (IOError, UnicodeEncodeError, TypeError, ValueError):
|
||
|
# if url_file_stream_or_string is a str object that
|
||
|
# cannot be converted to the encoding returned by
|
||
|
# sys.getfilesystemencoding(), a UnicodeEncodeError
|
||
|
# will be thrown
|
||
|
# If url_file_stream_or_string is a string that contains NULL
|
||
|
# (such as an XML document encoded in UTF-32), TypeError will
|
||
|
# be thrown.
|
||
|
pass
|
||
|
else:
|
||
|
return data
|
||
|
|
||
|
# treat url_file_stream_or_string as string
|
||
|
if not isinstance(url_file_stream_or_string, bytes):
|
||
|
return url_file_stream_or_string.encode('utf-8')
|
||
|
return url_file_stream_or_string
|
||
|
|
||
|
|
||
|
LooseFeedParser = type(
|
||
|
'LooseFeedParser',
|
||
|
(_LooseFeedParser, _FeedParserMixin, _BaseHTMLProcessor, object),
|
||
|
{},
|
||
|
)
|
||
|
|
||
|
StrictFeedParser = type(
|
||
|
'StrictFeedParser',
|
||
|
(_StrictFeedParser, _FeedParserMixin, xml.sax.handler.ContentHandler, object),
|
||
|
{},
|
||
|
)
|
||
|
|
||
|
|
||
|
def parse(url_file_stream_or_string, etag=None, modified=None, agent=None, referrer=None, handlers=None, request_headers=None, response_headers=None, resolve_relative_uris=None, sanitize_html=None):
|
||
|
"""Parse a feed from a URL, file, stream, or string.
|
||
|
|
||
|
:param url_file_stream_or_string:
|
||
|
File-like object, URL, file path, or string. Both byte and text strings
|
||
|
are accepted. If necessary, encoding will be derived from the response
|
||
|
headers or automatically detected.
|
||
|
|
||
|
Note that strings may trigger network I/O or filesystem access
|
||
|
depending on the value. Wrap an untrusted string in
|
||
|
a :class:`io.StringIO` or :class:`io.BytesIO` to avoid this. Do not
|
||
|
pass untrusted strings to this function.
|
||
|
|
||
|
When a URL is not passed the feed location to use in relative URL
|
||
|
resolution should be passed in the ``Content-Location`` response header
|
||
|
(see ``response_headers`` below).
|
||
|
|
||
|
:param str etag: HTTP ``ETag`` request header.
|
||
|
:param modified: HTTP ``Last-Modified`` request header.
|
||
|
:type modified: :class:`str`, :class:`time.struct_time` 9-tuple, or
|
||
|
:class:`datetime.datetime`
|
||
|
:param str agent: HTTP ``User-Agent`` request header, which defaults to
|
||
|
the value of :data:`feedparser.USER_AGENT`.
|
||
|
:param referrer: HTTP ``Referer`` [sic] request header.
|
||
|
:param request_headers:
|
||
|
A mapping of HTTP header name to HTTP header value to add to the
|
||
|
request, overriding internally generated values.
|
||
|
:type request_headers: :class:`dict` mapping :class:`str` to :class:`str`
|
||
|
:param response_headers:
|
||
|
A mapping of HTTP header name to HTTP header value. Multiple values may
|
||
|
be joined with a comma. If a HTTP request was made, these headers
|
||
|
override any matching headers in the response. Otherwise this specifies
|
||
|
the entirety of the response headers.
|
||
|
:type response_headers: :class:`dict` mapping :class:`str` to :class:`str`
|
||
|
|
||
|
:param bool resolve_relative_uris:
|
||
|
Should feedparser attempt to resolve relative URIs absolute ones within
|
||
|
HTML content? Defaults to the value of
|
||
|
:data:`feedparser.RESOLVE_RELATIVE_URIS`, which is ``True``.
|
||
|
:param bool sanitize_html:
|
||
|
Should feedparser skip HTML sanitization? Only disable this if you know
|
||
|
what you are doing! Defaults to the value of
|
||
|
:data:`feedparser.SANITIZE_HTML`, which is ``True``.
|
||
|
|
||
|
:return: A :class:`FeedParserDict`.
|
||
|
"""
|
||
|
|
||
|
if not agent or sanitize_html is None or resolve_relative_uris is None:
|
||
|
import feedparser
|
||
|
if not agent:
|
||
|
agent = feedparser.USER_AGENT
|
||
|
if sanitize_html is None:
|
||
|
sanitize_html = feedparser.SANITIZE_HTML
|
||
|
if resolve_relative_uris is None:
|
||
|
resolve_relative_uris = feedparser.RESOLVE_RELATIVE_URIS
|
||
|
|
||
|
result = FeedParserDict(
|
||
|
bozo=False,
|
||
|
entries=[],
|
||
|
feed=FeedParserDict(),
|
||
|
headers={},
|
||
|
)
|
||
|
|
||
|
data = _open_resource(url_file_stream_or_string, etag, modified, agent, referrer, handlers, request_headers, result)
|
||
|
|
||
|
if not data:
|
||
|
return result
|
||
|
|
||
|
# overwrite existing headers using response_headers
|
||
|
result['headers'].update(response_headers or {})
|
||
|
|
||
|
data = convert_to_utf8(result['headers'], data, result)
|
||
|
use_strict_parser = result['encoding'] and True or False
|
||
|
|
||
|
result['version'], data, entities = replace_doctype(data)
|
||
|
|
||
|
# Ensure that baseuri is an absolute URI using an acceptable URI scheme.
|
||
|
contentloc = result['headers'].get('content-location', '')
|
||
|
href = result.get('href', '')
|
||
|
baseuri = make_safe_absolute_uri(href, contentloc) or make_safe_absolute_uri(contentloc) or href
|
||
|
|
||
|
baselang = result['headers'].get('content-language', None)
|
||
|
if isinstance(baselang, bytes) and baselang is not None:
|
||
|
baselang = baselang.decode('utf-8', 'ignore')
|
||
|
|
||
|
if not _XML_AVAILABLE:
|
||
|
use_strict_parser = 0
|
||
|
if use_strict_parser:
|
||
|
# initialize the SAX parser
|
||
|
feedparser = StrictFeedParser(baseuri, baselang, 'utf-8')
|
||
|
feedparser.resolve_relative_uris = resolve_relative_uris
|
||
|
feedparser.sanitize_html = sanitize_html
|
||
|
saxparser = xml.sax.make_parser(PREFERRED_XML_PARSERS)
|
||
|
saxparser.setFeature(xml.sax.handler.feature_namespaces, 1)
|
||
|
try:
|
||
|
# disable downloading external doctype references, if possible
|
||
|
saxparser.setFeature(xml.sax.handler.feature_external_ges, 0)
|
||
|
except xml.sax.SAXNotSupportedException:
|
||
|
pass
|
||
|
saxparser.setContentHandler(feedparser)
|
||
|
saxparser.setErrorHandler(feedparser)
|
||
|
source = xml.sax.xmlreader.InputSource()
|
||
|
source.setByteStream(io.BytesIO(data))
|
||
|
try:
|
||
|
saxparser.parse(source)
|
||
|
except xml.sax.SAXException as e:
|
||
|
result['bozo'] = 1
|
||
|
result['bozo_exception'] = feedparser.exc or e
|
||
|
use_strict_parser = 0
|
||
|
if not use_strict_parser:
|
||
|
feedparser = LooseFeedParser(baseuri, baselang, 'utf-8', entities)
|
||
|
feedparser.resolve_relative_uris = resolve_relative_uris
|
||
|
feedparser.sanitize_html = sanitize_html
|
||
|
feedparser.feed(data.decode('utf-8', 'replace'))
|
||
|
result['feed'] = feedparser.feeddata
|
||
|
result['entries'] = feedparser.entries
|
||
|
result['version'] = result['version'] or feedparser.version
|
||
|
result['namespaces'] = feedparser.namespaces_in_use
|
||
|
return result
|