import netrc
import os
import random
+import re
import sys
import time
+import types
import urllib.parse
import urllib.request
import xml.etree.ElementTree
-from ..compat import functools, re # isort: split
+from ..compat import functools # isort: split
from ..compat import compat_etree_fromstring, compat_expanduser, compat_os_name
from ..downloader import FileDownloader
from ..downloader.f4m import get_base_url, remove_encrypted_media
from ..utils import (
+ IDENTITY,
JSON_LD_RE,
NO_DEFAULT,
ExtractorError,
GeoUtils,
LenientJSONDecoder,
RegexNotFoundError,
+ RetryManager,
UnsupportedError,
age_restricted,
base_url,
parse_m3u8_attributes,
parse_resolution,
sanitize_filename,
+ sanitize_url,
sanitized_Request,
str_or_none,
str_to_int,
live stream that goes on instead of a fixed-length video.
was_live: True, False, or None (=unknown). Whether this video was
originally a live stream.
- live_status: 'is_live', 'is_upcoming', 'was_live', 'not_live' or None (=unknown)
+ live_status: None (=unknown), 'is_live', 'is_upcoming', 'was_live', 'not_live'
+ or 'post_live' (was live, but VOD is not yet processed)
If absent, automatically set from is_live, was_live
start_time: Time in seconds where the reproduction should start, as
specified in the URL.
'private', 'premium_only', 'subscriber_only', 'needs_auth',
'unlisted' or 'public'. Use 'InfoExtractor._availability'
to set it
+ _old_archive_ids: A list of old archive ids needed for backward compatibility
__post_extractor: A function to be called just before the metadata is
written to either disk, logger or console. The function
must return a dict which will be added to the info_dict.
title, description etc.
- Subclasses of this should define a _VALID_URL regexp and, re-define the
- _real_extract() and (optionally) _real_initialize() methods.
- Probably, they should also be added to the list of extractors.
+ Subclasses of this should also be added to the list of extractors and
+ should define a _VALID_URL regexp and, re-define the _real_extract() and
+ (optionally) _real_initialize() methods.
Subclasses may also override suitable() if necessary, but ensure the function
signature is preserved and that this function imports everything it needs
(except other extractors), so that lazy_extractors works correctly.
+ Subclasses can define a list of _EMBED_REGEX, which will be searched for in
+ the HTML of Generic webpages. It may also override _extract_embed_urls
+ or _extract_from_webpage as necessary. While these are normally classmethods,
+ _extract_from_webpage is allowed to be an instance method.
+
+ _extract_from_webpage may raise self.StopExtraction() to stop further
+ processing of the webpage and obtain exclusive rights to it. This is useful
+ when the extractor cannot reliably be matched using just the URL.
+ Eg: invidious/peertube instances
+
+ Embed-only extractors can be defined by setting _VALID_URL = False.
+
To support username + password (or netrc) login, the extractor must define a
_NETRC_MACHINE and re-define _perform_login(username, password) and
(optionally) _initialize_pre_login() methods. The _perform_login method will
_NETRC_MACHINE = None
IE_DESC = None
SEARCH_KEY = None
+ _VALID_URL = None
+ _EMBED_REGEX = []
def _login_hint(self, method=NO_DEFAULT, netrc=None):
password_hint = f'--username and --password, or --netrc ({netrc or self._NETRC_MACHINE}) to provide account credentials'
@classmethod
def _match_valid_url(cls, url):
+ if cls._VALID_URL is False:
+ return None
# This does not use has/getattr intentionally - we want to know whether
# we have cached the regexp for *this* class, whereas getattr would also
# match the superclass
if '_VALID_URL_RE' not in cls.__dict__:
- if '_VALID_URL' not in cls.__dict__:
- cls._VALID_URL = cls._make_valid_url()
cls._VALID_URL_RE = re.compile(cls._VALID_URL)
return cls._VALID_URL_RE.match(url)
return None
if self._x_forwarded_for_ip:
ie_result['__x_forwarded_for_ip'] = self._x_forwarded_for_ip
- subtitles = ie_result.get('subtitles')
- if (subtitles and 'live_chat' in subtitles
- and 'no-live-chat' in self.get_param('compat_opts', [])):
- del subtitles['live_chat']
+ subtitles = ie_result.get('subtitles') or {}
+ if 'no-live-chat' in self.get_param('compat_opts'):
+ for lang in ('live_chat', 'comments', 'danmaku'):
+ subtitles.pop(lang, None)
return ie_result
except GeoRestrictedError as e:
if self.__maybe_fake_ip_and_retry(e.countries):
return content
- def _parse_xml(self, xml_string, video_id, transform_source=None, fatal=True):
+ def __print_error(self, errnote, fatal, video_id, err):
+ if fatal:
+ raise ExtractorError(f'{video_id}: {errnote}', cause=err)
+ elif errnote:
+ self.report_warning(f'{video_id}: {errnote}: {err}')
+
+ def _parse_xml(self, xml_string, video_id, transform_source=None, fatal=True, errnote=None):
if transform_source:
xml_string = transform_source(xml_string)
try:
return compat_etree_fromstring(xml_string.encode('utf-8'))
except xml.etree.ElementTree.ParseError as ve:
- errmsg = '%s: Failed to parse XML ' % video_id
- if fatal:
- raise ExtractorError(errmsg, cause=ve)
- else:
- self.report_warning(errmsg + str(ve))
+ self.__print_error('Failed to parse XML' if errnote is None else errnote, fatal, video_id, ve)
- def _parse_json(self, json_string, video_id, transform_source=None, fatal=True, **parser_kwargs):
+ def _parse_json(self, json_string, video_id, transform_source=None, fatal=True, errnote=None, **parser_kwargs):
try:
return json.loads(
json_string, cls=LenientJSONDecoder, strict=False, transform_source=transform_source, **parser_kwargs)
except ValueError as ve:
- errmsg = f'{video_id}: Failed to parse JSON'
- if fatal:
- raise ExtractorError(errmsg, cause=ve)
- else:
- self.report_warning(f'{errmsg}: {ve}')
+ self.__print_error('Failed to parse JSON' if errnote is None else errnote, fatal, video_id, ve)
- def _parse_socket_response_as_json(self, data, video_id, transform_source=None, fatal=True):
- return self._parse_json(
- data[data.find('{'):data.rfind('}') + 1],
- video_id, transform_source, fatal)
+ def _parse_socket_response_as_json(self, data, *args, **kwargs):
+ return self._parse_json(data[data.find('{'):data.rfind('}') + 1], *args, **kwargs)
def __create_download_methods(name, parser, note, errnote, return_value):
- def parse(ie, content, *args, **kwargs):
+ def parse(ie, content, *args, errnote=errnote, **kwargs):
if parser is None:
return content
+ if errnote is False:
+ kwargs['errnote'] = errnote
# parser is fetched by name so subclasses can override it
return getattr(ie, parser)(content, *args, **kwargs)
if res is False:
return res
content, urlh = res
- return parse(self, content, video_id, transform_source=transform_source, fatal=fatal), urlh
+ return parse(self, content, video_id, transform_source=transform_source, fatal=fatal, errnote=errnote), urlh
def download_content(self, url_or_request, video_id, note=note, errnote=errnote, transform_source=None,
fatal=True, encoding=None, data=None, headers={}, query={}, expected_status=None):
self.report_warning(f'Unable to load request from disk: {e}')
else:
content = self.__decode_webpage(webpage_bytes, encoding, url_or_request.headers)
- return parse(self, content, video_id, transform_source, fatal)
+ return parse(self, content, video_id, transform_source=transform_source, fatal=fatal, errnote=errnote)
kwargs = {
'note': note,
'errnote': errnote,
'url': url,
}
- def playlist_from_matches(self, matches, playlist_id=None, playlist_title=None, getter=None, ie=None, video_kwargs=None, **kwargs):
- urls = (self.url_result(self._proto_relative_url(m), ie, **(video_kwargs or {}))
- for m in orderedSet(map(getter, matches) if getter else matches))
- return self.playlist_result(urls, playlist_id, playlist_title, **kwargs)
+ @classmethod
+ def playlist_from_matches(cls, matches, playlist_id=None, playlist_title=None,
+ getter=IDENTITY, ie=None, video_kwargs=None, **kwargs):
+ return cls.playlist_result(
+ (cls.url_result(m, ie, **(video_kwargs or {})) for m in orderedSet(map(getter, matches), lazy=True)),
+ playlist_id, playlist_title, **kwargs)
@staticmethod
def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None, *, multi_video=False, **kwargs):
def _dc_search_uploader(self, html):
return self._html_search_meta('dc.creator', html, 'uploader')
- def _rta_search(self, html):
+ @staticmethod
+ def _rta_search(html):
# See http://www.rtalabel.org/index.php?content=howtofaq#single
if re.search(r'(?ix)<meta\s+name="rating"\s+'
r' content="RTA-5042-1996-1400-1577-RTA"',
html):
return 18
+
+ # And then there are the jokers who advertise that they use RTA, but actually don't.
+ AGE_LIMIT_MARKERS = [
+ r'Proudly Labeled <a href="http://www\.rtalabel\.org/" title="Restricted to Adults">RTA</a>',
+ ]
+ if any(re.search(marker, html) for marker in AGE_LIMIT_MARKERS):
+ return 18
return 0
def _media_rating_search(self, html):
else 'https:')
def _proto_relative_url(self, url, scheme=None):
- if url is None:
- return url
- if url.startswith('//'):
- if scheme is None:
- scheme = self.http_scheme()
- return scheme + url
- else:
- return url
+ scheme = scheme or self.http_scheme()
+ assert scheme.endswith(':')
+ return sanitize_url(url, scheme=scheme[:-1])
def _sleep(self, timeout, video_id, msg_template=None):
if msg_template is None:
t['name'] = cls.ie_key()
yield t
+ @classmethod
+ def get_webpage_testcases(cls):
+ tests = getattr(cls, '_WEBPAGE_TESTS', [])
+ for t in tests:
+ t['name'] = cls.ie_key()
+ return tests
+
@classproperty
def age_limit(cls):
"""Get age limit from the testcases"""
return max(traverse_obj(
- tuple(cls.get_testcases(include_onlymatching=False)),
+ (*cls.get_testcases(include_onlymatching=False), *cls.get_webpage_testcases()),
(..., (('playlist', 0), None), 'info_dict', 'age_limit')) or [0])
@classmethod
headers['Ytdl-request-proxy'] = geo_verification_proxy
return headers
- def _generic_id(self, url):
+ @staticmethod
+ def _generic_id(url):
return urllib.parse.unquote(os.path.splitext(url.rstrip('/').split('/')[-1])[0])
- def _generic_title(self, url):
+ @staticmethod
+ def _generic_title(url):
return urllib.parse.unquote(os.path.splitext(url_basename(url))[0])
@staticmethod
self.to_screen(f'Downloading {playlist_label}{playlist_id} - add --no-playlist to download just the {video_label}{video_id}')
return True
+ def _error_or_warning(self, err, _count=None, _retries=0, *, fatal=True):
+ RetryManager.report_retry(err, _count or int(fatal), _retries, info=self.to_screen, warn=self.report_warning,
+ sleep_func=self.get_param('retry_sleep_functions', {}).get('extractor'))
+
+ def RetryManager(self, **kwargs):
+ return RetryManager(self.get_param('extractor_retries', 3), self._error_or_warning, **kwargs)
+
+ @classmethod
+ def extract_from_webpage(cls, ydl, url, webpage):
+ ie = (cls if isinstance(cls._extract_from_webpage, types.MethodType)
+ else ydl.get_info_extractor(cls.ie_key()))
+ for info in ie._extract_from_webpage(url, webpage) or []:
+ # url = None since we do not want to set (webpage/original)_url
+ ydl.add_default_extra_info(info, ie, None)
+ yield info
+
+ @classmethod
+ def _extract_from_webpage(cls, url, webpage):
+ for embed_url in orderedSet(
+ cls._extract_embed_urls(url, webpage) or [], lazy=True):
+ yield cls.url_result(embed_url, cls)
+
+ @classmethod
+ def _extract_embed_urls(cls, url, webpage):
+ """@returns all the embed urls on the webpage"""
+ if '_EMBED_URL_RE' not in cls.__dict__:
+ assert isinstance(cls._EMBED_REGEX, (list, tuple))
+ for idx, regex in enumerate(cls._EMBED_REGEX):
+ assert regex.count('(?P<url>') == 1, \
+ f'{cls.__name__}._EMBED_REGEX[{idx}] must have exactly 1 url group\n\t{regex}'
+ cls._EMBED_URL_RE = tuple(map(re.compile, cls._EMBED_REGEX))
+
+ for regex in cls._EMBED_URL_RE:
+ for mobj in regex.finditer(webpage):
+ embed_url = urllib.parse.urljoin(url, unescapeHTML(mobj.group('url')))
+ if cls._VALID_URL is False or cls.suitable(embed_url):
+ yield embed_url
+
+ class StopExtraction(Exception):
+ pass
+
+ @classmethod
+ def _extract_url(cls, webpage): # TODO: Remove
+ """Only for compatibility with some older extractors"""
+ return next(iter(cls._extract_embed_urls(None, webpage) or []), None)
+
class SearchInfoExtractor(InfoExtractor):
"""
_MAX_RESULTS = float('inf')
- @classmethod
- def _make_valid_url(cls):
+ @classproperty
+ def _VALID_URL(cls):
return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
def _real_extract(self, query):