-# coding: utf-8
-
-from __future__ import unicode_literals
-
+import base64
import calendar
import copy
import datetime
-import functools
import hashlib
import itertools
import json
import random
import re
import sys
+import threading
import time
import traceback
-import threading
from .common import InfoExtractor, SearchInfoExtractor
+from ..compat import functools # isort: split
from ..compat import (
compat_chr,
compat_HTTPError,
)
from ..jsinterp import JSInterpreter
from ..utils import (
+ NO_DEFAULT,
+ ExtractorError,
bug_reports_message,
+ classproperty,
clean_html,
datetime_from_str,
dict_get,
error_to_compat_str,
- ExtractorError,
float_or_none,
format_field,
+ get_first,
int_or_none,
is_html,
join_nonempty,
js_to_json,
mimetype2ext,
network_exceptions,
- NO_DEFAULT,
orderedSet,
parse_codecs,
parse_count,
variadic,
)
-
-def get_first(obj, keys, **kwargs):
- return traverse_obj(obj, (..., *variadic(keys)), **kwargs, get_all=False)
-
-
# any clients starting with _ cannot be explicity requested by the user
INNERTUBE_CLIENTS = {
'web': {
}
},
'INNERTUBE_CONTEXT_CLIENT_NAME': 2
- }
+ },
+ # This client can access age restricted videos (unless the uploader has disabled the 'allow embedding' option)
+ # See: https://github.com/zerodytrash/YouTube-Internal-Clients
+ 'tv_embedded': {
+ 'INNERTUBE_API_KEY': 'AIzaSyAO_FJ2SlqU8Q4STEHLGCilw_Y9_11qcW8',
+ 'INNERTUBE_CONTEXT': {
+ 'client': {
+ 'clientName': 'TVHTML5_SIMPLY_EMBEDDED_PLAYER',
+ 'clientVersion': '2.0',
+ },
+ },
+ 'INNERTUBE_CONTEXT_CLIENT_NAME': 85
+ },
}
+def _split_innertube_client(client_name):
+ variant, *base = client_name.rsplit('.', 1)
+ if base:
+ return variant, base[0], variant
+ base, *variant = client_name.split('_', 1)
+ return client_name, base, variant[0] if variant else None
+
+
def build_innertube_clients():
THIRD_PARTY = {
- 'embedUrl': 'https://google.com', # Can be any valid URL
+ 'embedUrl': 'https://www.youtube.com/', # Can be any valid URL
}
- BASE_CLIENTS = ('android', 'web', 'ios', 'mweb')
+ BASE_CLIENTS = ('android', 'web', 'tv', 'ios', 'mweb')
priority = qualities(BASE_CLIENTS[::-1])
for client, ytcfg in tuple(INNERTUBE_CLIENTS.items()):
ytcfg.setdefault('REQUIRE_JS_PLAYER', True)
ytcfg['INNERTUBE_CONTEXT']['client'].setdefault('hl', 'en')
- base_client, *variant = client.split('_')
+ _, base_client, variant = _split_innertube_client(client)
ytcfg['priority'] = 10 * priority(base_client)
- if variant == ['embedded']:
+ if not variant:
+ INNERTUBE_CLIENTS[f'{client}_embedscreen'] = embedscreen = copy.deepcopy(ytcfg)
+ embedscreen['INNERTUBE_CONTEXT']['client']['clientScreen'] = 'EMBED'
+ embedscreen['INNERTUBE_CONTEXT']['thirdParty'] = THIRD_PARTY
+ embedscreen['priority'] -= 3
+ elif variant == 'embedded':
ytcfg['INNERTUBE_CONTEXT']['thirdParty'] = THIRD_PARTY
- INNERTUBE_CLIENTS[f'{base_client}_agegate'] = agegate_ytcfg = copy.deepcopy(ytcfg)
- agegate_ytcfg['INNERTUBE_CONTEXT']['client']['clientScreen'] = 'EMBED'
- agegate_ytcfg['priority'] -= 1
ytcfg['priority'] -= 2
- elif variant:
+ else:
ytcfg['priority'] -= 3
_PLAYLIST_ID_RE = r'(?:(?:PL|LL|EC|UU|FL|RD|UL|TL|PU|OLAK5uy_)[0-9A-Za-z-_]{10,}|RDMM|WL|LL|LM)'
- _NETRC_MACHINE = 'youtube'
+ # _NETRC_MACHINE = 'youtube'
# If True it will raise an error if no login info is provided
_LOGIN_REQUIRED = False
# invidious-redirect websites
r'(?:www\.)?redirect\.invidious\.io',
r'(?:(?:www|dev)\.)?invidio\.us',
- # Invidious instances taken from https://github.com/iv-org/documentation/blob/master/Invidious-Instances.md
+ # Invidious instances taken from https://github.com/iv-org/documentation/blob/master/docs/instances.md
r'(?:www\.)?invidious\.pussthecat\.org',
r'(?:www\.)?invidious\.zee\.li',
r'(?:www\.)?invidious\.ethibox\.fr',
r'(?:www\.)?kbjggqkzv65ivcqj6bumvp337z6264huv5kpkwuv6gu5yjiskvan7fad\.onion',
r'(?:www\.)?grwp24hodrefzvjjuccrkw3mjq4tzhaaq32amf33dzpmuxe7ilepcmad\.onion',
r'(?:www\.)?hpniueoejy4opn7bc4ftgazyqjoeqwlvh2uiku2xqku6zpoa4bf5ruid\.onion',
+ # piped instances from https://github.com/TeamPiped/Piped/wiki/Instances
+ r'(?:www\.)?piped\.kavin\.rocks',
+ r'(?:www\.)?piped\.silkky\.cloud',
+ r'(?:www\.)?piped\.tokhmi\.xyz',
+ r'(?:www\.)?piped\.moomoo\.me',
+ r'(?:www\.)?il\.ax',
+ r'(?:www\.)?piped\.syncpundit\.com',
+ r'(?:www\.)?piped\.mha\.fi',
+ r'(?:www\.)?piped\.mint\.lgbt',
+ r'(?:www\.)?piped\.privacy\.com\.de',
)
- def _login(self):
- """
- Attempt to log in to YouTube.
- If _LOGIN_REQUIRED is set and no authentication was provided, an error is raised.
- """
-
- if (self._LOGIN_REQUIRED
- and self.get_param('cookiefile') is None
- and self.get_param('cookiesfrombrowser') is None):
- self.raise_login_required(
- 'Login details are needed to download this content', method='cookies')
- username, password = self._get_login_info()
- if username:
- self.report_warning(f'Cannot login to YouTube using username and password. {self._LOGIN_HINTS["cookies"]}')
-
def _initialize_consent(self):
cookies = self._get_cookies('https://www.youtube.com/')
if cookies.get('__Secure-3PSID'):
def _real_initialize(self):
self._initialize_pref()
self._initialize_consent()
- self._login()
+ self._check_login_required()
- _YT_INITIAL_DATA_RE = r'(?:window\s*\[\s*["\']ytInitialData["\']\s*\]|ytInitialData)\s*=\s*({.+?})\s*;'
- _YT_INITIAL_PLAYER_RESPONSE_RE = r'ytInitialPlayerResponse\s*=\s*({.+?})\s*;'
- _YT_INITIAL_BOUNDARY_RE = r'(?:var\s+meta|</script|\n)'
+ def _check_login_required(self):
+ if self._LOGIN_REQUIRED and not self._cookies_passed:
+ self.raise_login_required('Login details are needed to download this content', method='cookies')
+
+ _YT_INITIAL_DATA_RE = r'(?:window\s*\[\s*["\']ytInitialData["\']\s*\]|ytInitialData)\s*='
+ _YT_INITIAL_PLAYER_RESPONSE_RE = r'ytInitialPlayerResponse\s*='
def _get_default_ytcfg(self, client='web'):
return copy.deepcopy(INNERTUBE_CLIENTS[client])
ytcfg, (lambda x: x['INNERTUBE_CLIENT_VERSION'],
lambda x: x['INNERTUBE_CONTEXT']['client']['clientVersion']), compat_str, default_client)
+ def _select_api_hostname(self, req_api_hostname, default_client=None):
+ return (self._configuration_arg('innertube_host', [''], ie_key=YoutubeIE.ie_key())[0]
+ or req_api_hostname or self._get_innertube_host(default_client or 'web'))
+
def _extract_api_key(self, ytcfg=None, default_client='web'):
return self._ytcfg_get_safe(ytcfg, lambda x: x['INNERTUBE_API_KEY'], compat_str, default_client)
return None
# SAPISIDHASH algorithm from https://stackoverflow.com/a/32065323
sapisidhash = hashlib.sha1(
- f'{time_now} {self._SAPISID} {origin}'.encode('utf-8')).hexdigest()
+ f'{time_now} {self._SAPISID} {origin}'.encode()).hexdigest()
return f'SAPISIDHASH {time_now}_{sapisidhash}'
def _call_api(self, ep, query, video_id, fatal=True, headers=None,
real_headers.update({'content-type': 'application/json'})
if headers:
real_headers.update(headers)
+ api_key = (self._configuration_arg('innertube_key', [''], ie_key=YoutubeIE.ie_key(), casesense=True)[0]
+ or api_key or self._extract_api_key(default_client=default_client))
return self._download_json(
- 'https://%s/youtubei/v1/%s' % (api_hostname or self._get_innertube_host(default_client), ep),
+ f'https://{self._select_api_hostname(api_hostname, default_client)}/youtubei/v1/{ep}',
video_id=video_id, fatal=fatal, note=note, errnote=errnote,
data=json.dumps(data).encode('utf8'), headers=real_headers,
- query={'key': api_key or self._extract_api_key()})
+ query={'key': api_key, 'prettyPrint': 'false'})
def extract_yt_initial_data(self, item_id, webpage, fatal=True):
- data = self._search_regex(
- (r'%s\s*%s' % (self._YT_INITIAL_DATA_RE, self._YT_INITIAL_BOUNDARY_RE),
- self._YT_INITIAL_DATA_RE), webpage, 'yt initial data', fatal=fatal)
- if data:
- return self._parse_json(data, item_id, fatal=fatal)
+ return self._search_json(self._YT_INITIAL_DATA_RE, webpage, 'yt initial data', item_id, fatal=fatal)
@staticmethod
def _extract_session_index(*data):
args, [('VISITOR_DATA', ('INNERTUBE_CONTEXT', 'client', 'visitorData'), ('responseContext', 'visitorData'))],
expected_type=str)
- @property
+ @functools.cached_property
def is_authenticated(self):
return bool(self._generate_sapisidhash_header())
self, *, ytcfg=None, account_syncid=None, session_index=None,
visitor_data=None, identity_token=None, api_hostname=None, default_client='web'):
- origin = 'https://' + (api_hostname if api_hostname else self._get_innertube_host(default_client))
+ origin = 'https://' + (self._select_api_hostname(api_hostname, default_client))
headers = {
'X-YouTube-Client-Name': compat_str(
self._ytcfg_get_safe(ytcfg, lambda x: x['INNERTUBE_CONTEXT_CLIENT_NAME'], default_client=default_client)),
headers['X-Origin'] = origin
return {h: v for h, v in headers.items() if v is not None}
+ def _download_ytcfg(self, client, video_id):
+ url = {
+ 'web': 'https://www.youtube.com',
+ 'web_music': 'https://music.youtube.com',
+ 'web_embedded': f'https://www.youtube.com/embed/{video_id}?html5=1'
+ }.get(client)
+ if not url:
+ return {}
+ webpage = self._download_webpage(
+ url, video_id, fatal=False, note=f'Downloading {client.replace("_", " ").strip()} client config')
+ return self.extract_ytcfg(video_id, webpage) or {}
+
@staticmethod
def _build_api_continuation_query(continuation, ctp=None):
query = {
warnings.append([alert_type, alert_message])
for alert_type, alert_message in (warnings + errors[:-1]):
- self.report_warning('YouTube said: %s - %s' % (alert_type, alert_message), only_once=only_once)
+ self.report_warning(f'YouTube said: {alert_type} - {alert_message}', only_once=only_once)
if errors:
raise ExtractorError('YouTube said: %s' % errors[-1][1], expected=expected)
return None
def _extract_time_text(self, renderer, *path_list):
+ """@returns (timestamp, time_text)"""
text = self._get_text(renderer, *path_list) or ''
dt = self.extract_relative_time(text)
timestamp = None
timestamp = (
unified_timestamp(text) or unified_timestamp(
self._search_regex(
- (r'(?:.+|^)(?:live|premieres|ed|ing)(?:\s*on)?\s*(.+\d)', r'\w+[\s,\.-]*\w+[\s,\.-]+20\d{2}'),
+ (r'([a-z]+\s*\d{1,2},?\s*20\d{2})', r'(?:.+|^)(?:live|premieres|ed|ing)(?:\s*(?:on|for))?\s*(.+\d)'),
text.lower(), 'time text', default=None)))
if text and timestamp is None:
- self.report_warning('Cannot parse localized time text' + bug_reports_message(), only_once=True)
+ self.report_warning(f"Cannot parse localized time text '{text}'" + bug_reports_message(), only_once=True)
return timestamp, text
def _extract_response(self, item_id, query, note='Downloading API JSON', headers=None,
description = self._get_text(renderer, 'descriptionSnippet')
duration = parse_duration(self._get_text(
renderer, 'lengthText', ('thumbnailOverlays', ..., 'thumbnailOverlayTimeStatusRenderer', 'text')))
+ if duration is None:
+ duration = parse_duration(self._search_regex(
+ r'(?i)(ago)(?!.*\1)\s+(?P<duration>[a-z0-9 ,]+?)(?:\s+[\d,]+\s+views)?(?:\s+-\s+play\s+short)?$',
+ traverse_obj(renderer, ('title', 'accessibility', 'accessibilityData', 'label'), default='', expected_type=str),
+ video_id, default=None, group='duration'))
+
view_count = self._get_count(renderer, 'viewCountText')
uploader = self._get_text(renderer, 'ownerText', 'shortBylineText')
channel_id = traverse_obj(
- renderer, ('shortBylineText', 'runs', ..., 'navigationEndpoint', 'browseEndpoint', 'browseId'), expected_type=str, get_all=False)
+ renderer, ('shortBylineText', 'runs', ..., 'navigationEndpoint', 'browseEndpoint', 'browseId'),
+ expected_type=str, get_all=False)
timestamp, time_text = self._extract_time_text(renderer, 'publishedTimeText')
scheduled_timestamp = str_to_int(traverse_obj(renderer, ('upcomingEventData', 'startTime'), get_all=False))
overlay_style = traverse_obj(
- renderer, ('thumbnailOverlays', ..., 'thumbnailOverlayTimeStatusRenderer', 'style'), get_all=False, expected_type=str)
+ renderer, ('thumbnailOverlays', ..., 'thumbnailOverlayTimeStatusRenderer', 'style'),
+ get_all=False, expected_type=str)
badges = self._extract_badges(renderer)
thumbnails = self._extract_thumbnails(renderer, 'thumbnail')
+ navigation_url = urljoin('https://www.youtube.com/', traverse_obj(
+ renderer, ('navigationEndpoint', 'commandMetadata', 'webCommandMetadata', 'url'),
+ expected_type=str)) or ''
+ url = f'https://www.youtube.com/watch?v={video_id}'
+ if overlay_style == 'SHORTS' or '/shorts/' in navigation_url:
+ url = f'https://www.youtube.com/shorts/{video_id}'
return {
'_type': 'url',
'ie_key': YoutubeIE.ie_key(),
'id': video_id,
- 'url': f'https://www.youtube.com/watch?v={video_id}',
+ 'url': url,
'title': title,
'description': description,
'duration': duration,
'uploader': uploader,
'channel_id': channel_id,
'thumbnails': thumbnails,
- # 'upload_date': strftime_or_none(timestamp, '%Y%m%d'),
+ 'upload_date': (strftime_or_none(timestamp, '%Y%m%d')
+ if self._configuration_arg('approximate_date', ie_key='youtubetab')
+ else None),
'live_status': ('is_upcoming' if scheduled_timestamp is not None
else 'was_live' if 'streamed' in time_text.lower()
else 'is_live' if overlay_style is not None and overlay_style == 'LIVE' or 'live now' in badges
'id': 'Tq92D6wQ1mg',
'title': '[MMD] Adios - EVERGLOW [+Motion DL]',
'ext': 'mp4',
- 'upload_date': '20191227',
+ 'upload_date': '20191228',
'uploader_id': 'UC1yoRdFoFJaCY-AGfD9W0wQ',
'uploader': 'Projekt Melody',
'description': 'md5:17eccca93a786d51bc67646756894066',
'title': 'md5:e41008789470fc2533a3252216f1c1d1',
'description': 'md5:a677553cf0840649b731a3024aeff4cc',
'duration': 721,
- 'upload_date': '20150127',
+ 'upload_date': '20150128',
'uploader_id': 'BerkmanCenter',
'uploader_url': r're:https?://(?:www\.)?youtube\.com/user/BerkmanCenter',
'uploader': 'The Berkman Klein Center for Internet & Society',
'title': 'Democratic Socialism and Foreign Policy | Bernie Sanders',
'description': 'md5:13a2503d7b5904ef4b223aa101628f39',
'duration': 4060,
- 'upload_date': '20151119',
+ 'upload_date': '20151120',
'uploader': 'Bernie Sanders',
'uploader_id': 'UCH1dpzjCEiGAt8CXkryhkZg',
'uploader_url': r're:https?://(?:www\.)?youtube\.com/channel/UCH1dpzjCEiGAt8CXkryhkZg',
'age_limit': 0,
'channel_follower_count': int
}, 'params': {'format': 'mhtml', 'skip_download': True}
+ }, {
+ # Ensure video upload_date is in UTC timezone (video was uploaded 1641170939)
+ 'url': 'https://www.youtube.com/watch?v=2NUZ8W2llS4',
+ 'info_dict': {
+ 'id': '2NUZ8W2llS4',
+ 'ext': 'mp4',
+ 'title': 'The NP that test your phone performance 🙂',
+ 'description': 'md5:144494b24d4f9dfacb97c1bbef5de84d',
+ 'uploader': 'Leon Nguyen',
+ 'uploader_id': 'VNSXIII',
+ 'uploader_url': 'http://www.youtube.com/user/VNSXIII',
+ 'channel_id': 'UCRqNBSOHgilHfAczlUmlWHA',
+ 'channel_url': 'https://www.youtube.com/channel/UCRqNBSOHgilHfAczlUmlWHA',
+ 'duration': 21,
+ 'view_count': int,
+ 'age_limit': 0,
+ 'categories': ['Gaming'],
+ 'tags': 'count:23',
+ 'playable_in_embed': True,
+ 'live_status': 'not_live',
+ 'upload_date': '20220103',
+ 'like_count': int,
+ 'availability': 'public',
+ 'channel': 'Leon Nguyen',
+ 'thumbnail': 'https://i.ytimg.com/vi_webp/2NUZ8W2llS4/maxresdefault.webp',
+ 'channel_follower_count': int
+ }
+ }, {
+ # date text is premiered video, ensure upload date in UTC (published 1641172509)
+ 'url': 'https://www.youtube.com/watch?v=mzZzzBU6lrM',
+ 'info_dict': {
+ 'id': 'mzZzzBU6lrM',
+ 'ext': 'mp4',
+ 'title': 'I Met GeorgeNotFound In Real Life...',
+ 'description': 'md5:cca98a355c7184e750f711f3a1b22c84',
+ 'uploader': 'Quackity',
+ 'uploader_id': 'QuackityHQ',
+ 'uploader_url': 'http://www.youtube.com/user/QuackityHQ',
+ 'channel_id': 'UC_8NknAFiyhOUaZqHR3lq3Q',
+ 'channel_url': 'https://www.youtube.com/channel/UC_8NknAFiyhOUaZqHR3lq3Q',
+ 'duration': 955,
+ 'view_count': int,
+ 'age_limit': 0,
+ 'categories': ['Entertainment'],
+ 'tags': 'count:26',
+ 'playable_in_embed': True,
+ 'live_status': 'not_live',
+ 'release_timestamp': 1641172509,
+ 'release_date': '20220103',
+ 'upload_date': '20220103',
+ 'like_count': int,
+ 'availability': 'public',
+ 'channel': 'Quackity',
+ 'thumbnail': 'https://i.ytimg.com/vi/mzZzzBU6lrM/maxresdefault.jpg',
+ 'channel_follower_count': int
+ }
+ },
+ { # continuous livestream. Microformat upload date should be preferred.
+ # Upload date was 2021-06-19 (not UTC), while stream start is 2021-11-27
+ 'url': 'https://www.youtube.com/watch?v=kgx4WGK0oNU',
+ 'info_dict': {
+ 'id': 'kgx4WGK0oNU',
+ 'title': r're:jazz\/lofi hip hop radio🌱chill beats to relax\/study to \[LIVE 24\/7\] \d{4}-\d{2}-\d{2} \d{2}:\d{2}',
+ 'ext': 'mp4',
+ 'channel_id': 'UC84whx2xxsiA1gXHXXqKGOA',
+ 'availability': 'public',
+ 'age_limit': 0,
+ 'release_timestamp': 1637975704,
+ 'upload_date': '20210619',
+ 'channel_url': 'https://www.youtube.com/channel/UC84whx2xxsiA1gXHXXqKGOA',
+ 'live_status': 'is_live',
+ 'thumbnail': 'https://i.ytimg.com/vi/kgx4WGK0oNU/maxresdefault.jpg',
+ 'uploader': '阿鲍Abao',
+ 'uploader_url': 'http://www.youtube.com/channel/UC84whx2xxsiA1gXHXXqKGOA',
+ 'channel': 'Abao in Tokyo',
+ 'channel_follower_count': int,
+ 'release_date': '20211127',
+ 'tags': 'count:39',
+ 'categories': ['People & Blogs'],
+ 'like_count': int,
+ 'uploader_id': 'UC84whx2xxsiA1gXHXXqKGOA',
+ 'view_count': int,
+ 'playable_in_embed': True,
+ 'description': 'md5:2ef1d002cad520f65825346e2084e49d',
+ },
+ 'params': {'skip_download': True}
+ }, {
+ # Story. Requires specific player params to work.
+ # Note: stories get removed after some period of time
+ 'url': 'https://www.youtube.com/watch?v=vv8qTUWmulI',
+ 'info_dict': {
+ 'id': 'vv8qTUWmulI',
+ 'ext': 'mp4',
+ 'availability': 'unlisted',
+ 'view_count': int,
+ 'channel_id': 'UCzIZ8HrzDgc-pNQDUG6avBA',
+ 'upload_date': '20220526',
+ 'categories': ['Education'],
+ 'title': 'Story',
+ 'channel': 'IT\'S HISTORY',
+ 'description': '',
+ 'uploader_id': 'BlastfromthePast',
+ 'duration': 12,
+ 'uploader': 'IT\'S HISTORY',
+ 'playable_in_embed': True,
+ 'age_limit': 0,
+ 'live_status': 'not_live',
+ 'tags': [],
+ 'thumbnail': 'https://i.ytimg.com/vi_webp/vv8qTUWmulI/maxresdefault.webp',
+ 'uploader_url': 'http://www.youtube.com/user/BlastfromthePast',
+ 'channel_url': 'https://www.youtube.com/channel/UCzIZ8HrzDgc-pNQDUG6avBA',
+ }
+ }, {
+ 'url': 'https://www.youtube.com/watch?v=tjjjtzRLHvA',
+ 'info_dict': {
+ 'id': 'tjjjtzRLHvA',
+ 'ext': 'mp4',
+ 'title': 'ハッシュタグ無し };if window.ytcsi',
+ 'upload_date': '20220323',
+ 'like_count': int,
+ 'availability': 'unlisted',
+ 'channel': 'nao20010128nao',
+ 'thumbnail': 'https://i.ytimg.com/vi_webp/tjjjtzRLHvA/maxresdefault.webp',
+ 'age_limit': 0,
+ 'uploader': 'nao20010128nao',
+ 'uploader_id': 'nao20010128nao',
+ 'categories': ['Music'],
+ 'view_count': int,
+ 'description': '',
+ 'channel_url': 'https://www.youtube.com/channel/UCdqltm_7iv1Vs6kp6Syke5A',
+ 'channel_id': 'UCdqltm_7iv1Vs6kp6Syke5A',
+ 'live_status': 'not_live',
+ 'playable_in_embed': True,
+ 'channel_follower_count': int,
+ 'duration': 6,
+ 'tags': [],
+ 'uploader_url': 'http://www.youtube.com/user/nao20010128nao',
+ }
}
]
qs = parse_qs(url)
if qs.get('list', [None])[0]:
return False
- return super(YoutubeIE, cls).suitable(url)
+ return super().suitable(url)
def __init__(self, *args, **kwargs):
- super(YoutubeIE, self).__init__(*args, **kwargs)
+ super().__init__(*args, **kwargs)
self._code_cache = {}
self._player_cache = {}
return f['manifest_url'], f['manifest_stream_number'], is_live
for f in formats:
+ f['is_live'] = True
f['protocol'] = 'http_dash_segments_generator'
f['fragments'] = functools.partial(
self._live_dash_fragments, f['format_id'], live_start_time, mpd_feed)
known_idx, no_fragment_score, last_segment_url = begin_index, 0, None
fragments, fragment_base_url = None, None
- def _extract_sequence_from_mpd(refresh_sequence):
+ def _extract_sequence_from_mpd(refresh_sequence, immediate):
nonlocal mpd_url, stream_number, is_live, no_fragment_score, fragments, fragment_base_url
# Obtain from MPD's maximum seq value
old_mpd_url = mpd_url
last_error = ctx.pop('last_error', None)
- expire_fast = last_error and isinstance(last_error, compat_HTTPError) and last_error.code == 403
+ expire_fast = immediate or last_error and isinstance(last_error, compat_HTTPError) and last_error.code == 403
mpd_url, stream_number, is_live = (mpd_feed(format_id, 5 if expire_fast else 18000)
or (mpd_url, stream_number, False))
if not refresh_sequence:
except ExtractorError:
fmts = None
if not fmts:
- no_fragment_score += 1
+ no_fragment_score += 2
return False, last_seq
fmt_info = next(x for x in fmts if x['manifest_stream_number'] == stream_number)
fragments = fmt_info['fragments']
urlh = None
last_seq = try_get(urlh, lambda x: int_or_none(x.headers['X-Head-Seqnum']))
if last_seq is None:
- no_fragment_score += 1
+ no_fragment_score += 2
last_segment_url = None
continue
else:
- should_continue, last_seq = _extract_sequence_from_mpd(True)
+ should_continue, last_seq = _extract_sequence_from_mpd(True, no_fragment_score > 15)
+ no_fragment_score += 2
if not should_continue:
continue
try:
for idx in range(known_idx, last_seq):
# do not update sequence here or you'll get skipped some part of it
- should_continue, _ = _extract_sequence_from_mpd(False)
+ should_continue, _ = _extract_sequence_from_mpd(False, False)
if not should_continue:
known_idx = idx - 1
raise ExtractorError('breaking out of outer loop')
last_segment_url = urljoin(fragment_base_url, 'sq/%d' % idx)
yield {
'url': last_segment_url,
+ 'fragment_count': last_seq,
}
if known_idx == last_seq:
no_fragment_score += 5
player_id = self._extract_player_info(player_url)
# Read from filesystem cache
- func_id = 'js_%s_%s' % (
- player_id, self._signature_cache_id(example_sig))
+ func_id = f'js_{player_id}_{self._signature_cache_id(example_sig)}'
assert os.path.basename(func_id) == func_id
cache_spec = self._downloader.cache.load('youtube-sigfuncs', func_id)
starts = '' if start == 0 else str(start)
ends = (':%d' % (end + step)) if end + step >= 0 else ':'
steps = '' if step == 1 else (':%d' % step)
- return 's[%s%s%s]' % (starts, ends, steps)
+ return f's[{starts}{ends}{steps}]'
step = None
# Quelch pyflakes warnings - start will be set when step is set
def _decrypt_signature(self, s, video_id, player_url):
"""Turn the encrypted s field into a working signature"""
-
- if player_url is None:
- raise ExtractorError('Cannot decrypt signature without player_url')
-
try:
player_id = (player_url, self._signature_cache_id(s))
if player_id not in self._player_cache:
- func = self._extract_signature_function(
- video_id, player_url, s
- )
+ func = self._extract_signature_function(video_id, player_url, s)
self._player_cache[player_id] = func
func = self._player_cache[player_id]
self._print_sig_code(func, s)
return func(s)
except Exception as e:
- raise ExtractorError('Signature extraction failed: ' + traceback.format_exc(), cause=e)
+ raise ExtractorError(traceback.format_exc(), cause=e, video_id=video_id)
def _decrypt_nsig(self, s, video_id, player_url):
"""Turn the encrypted n field into a working signature"""
if not idx:
return nfunc
return json.loads(js_to_json(self._search_regex(
- rf'var {nfunc}\s*=\s*(\[.+?\]);', jscode,
+ rf'var {re.escape(nfunc)}\s*=\s*(\[.+?\]);', jscode,
f'Initial JS player n function list ({nfunc}.{idx})')))[int(idx)]
def _extract_n_function(self, video_id, player_url):
# cpn generation algorithm is reverse engineered from base.js.
# In fact it works even with dummy cpn.
CPN_ALPHABET = 'abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789-_'
- cpn = ''.join((CPN_ALPHABET[random.randint(0, 256) & 63] for _ in range(0, 16)))
+ cpn = ''.join(CPN_ALPHABET[random.randint(0, 256) & 63] for _ in range(0, 16))
qs.update({
'ver': ['2'],
chapter_time = lambda chapter: parse_duration(self._get_text(chapter, 'timeDescription'))
chapter_title = lambda chapter: self._get_text(chapter, 'title')
- return next((
- filter(None, (
- self._extract_chapters(
- traverse_obj(contents, (..., 'macroMarkersListItemRenderer')),
- chapter_time, chapter_title, duration)
- for contents in content_list
- ))), [])
-
- def _extract_chapters(self, chapter_list, chapter_time, chapter_title, duration):
- chapters = []
- last_chapter = {'start_time': 0}
- for idx, chapter in enumerate(chapter_list or []):
- title = chapter_title(chapter)
- start_time = chapter_time(chapter)
- if start_time is None:
- continue
- last_chapter['end_time'] = start_time
- if start_time < last_chapter['start_time']:
- if idx == 1:
- chapters.pop()
- self.report_warning('Invalid start time for chapter "%s"' % last_chapter['title'])
- else:
- self.report_warning(f'Invalid start time for chapter "{title}"')
- continue
- last_chapter = {'start_time': start_time, 'title': title}
- chapters.append(last_chapter)
- last_chapter['end_time'] = duration
- return chapters
+ return next(filter(None, (
+ self._extract_chapters(traverse_obj(contents, (..., 'macroMarkersListItemRenderer')),
+ chapter_time, chapter_title, duration)
+ for contents in content_list)), [])
- def _extract_yt_initial_variable(self, webpage, regex, video_id, name):
- return self._parse_json(self._search_regex(
- (r'%s\s*%s' % (regex, self._YT_INITIAL_BOUNDARY_RE),
- regex), webpage, name, default='{}'), video_id, fatal=False)
+ def _extract_chapters_from_description(self, description, duration):
+ return self._extract_chapters(
+ re.findall(r'(?m)^((?:\d+:)?\d{1,2}:\d{2})\b\W*\s(.+?)\s*$', description or ''),
+ chapter_time=lambda x: parse_duration(x[0]), chapter_title=lambda x: x[1],
+ duration=duration, strict=False)
+
+ def _extract_chapters(self, chapter_list, chapter_time, chapter_title, duration, strict=True):
+ if not duration:
+ return
+ chapter_list = [{
+ 'start_time': chapter_time(chapter),
+ 'title': chapter_title(chapter),
+ } for chapter in chapter_list or []]
+ if not strict:
+ chapter_list.sort(key=lambda c: c['start_time'] or 0)
+
+ chapters = [{'start_time': 0, 'title': '<Untitled>'}]
+ for idx, chapter in enumerate(chapter_list):
+ if chapter['start_time'] is None or not chapter['title']:
+ self.report_warning(f'Incomplete chapter {idx}')
+ elif chapters[-1]['start_time'] <= chapter['start_time'] <= duration:
+ chapters[-1]['end_time'] = chapter['start_time']
+ chapters.append(chapter)
+ else:
+ self.report_warning(f'Invalid start time for chapter "{chapter["title"]}"')
+ chapters[-1]['end_time'] = duration
+ return chapters if len(chapters) > 1 and chapters[1]['start_time'] else chapters[1:]
def _extract_comment(self, comment_renderer, parent=None):
comment_id = comment_renderer.get('commentId')
comment_entries_iter = self._comment_entries(
comment_replies_renderer, ytcfg, video_id,
parent=comment.get('id'), tracker=tracker)
- for reply_comment in itertools.islice(comment_entries_iter, min(max_replies_per_thread, max(0, max_replies - tracker['total_reply_comments']))):
- yield reply_comment
+ yield from itertools.islice(comment_entries_iter, min(
+ max_replies_per_thread, max(0, max_replies - tracker['total_reply_comments'])))
# Keeps track of counts across recursive calls
if not tracker:
lambda p: int_or_none(p, default=sys.maxsize), self._configuration_arg('max_comments', ) + [''] * 4)
continuation = self._extract_continuation(root_continuation_data)
- message = self._get_text(root_continuation_data, ('contents', ..., 'messageRenderer', 'text'), max_runs=1)
- if message and not parent:
- self.report_warning(message, video_id=video_id)
response = None
+ is_forced_continuation = False
is_first_continuation = parent is None
+ if is_first_continuation and not continuation:
+ # Sometimes you can get comments by generating the continuation yourself,
+ # even if YouTube initially reports them being disabled - e.g. stories comments.
+ # Note: if the comment section is actually disabled, YouTube may return a response with
+ # required check_get_keys missing. So we will disable that check initially in this case.
+ continuation = self._build_api_continuation_query(self._generate_comment_continuation(video_id))
+ is_forced_continuation = True
for page_num in itertools.count(0):
if not continuation:
response = self._extract_response(
item_id=None, query=continuation,
ep='next', ytcfg=ytcfg, headers=headers, note=note_prefix,
- check_get_keys='onResponseReceivedEndpoints')
-
+ check_get_keys='onResponseReceivedEndpoints' if not is_forced_continuation else None)
+ is_forced_continuation = False
continuation_contents = traverse_obj(
response, 'onResponseReceivedEndpoints', expected_type=list, default=[])
if continuation:
break
+ message = self._get_text(root_continuation_data, ('contents', ..., 'messageRenderer', 'text'), max_runs=1)
+ if message and not parent and tracker['running_total'] == 0:
+ self.report_warning(f'Youtube said: {message}', video_id=video_id, only_once=True)
+
+ @staticmethod
+ def _generate_comment_continuation(video_id):
+ """
+ Generates initial comment section continuation token from given video id
+ """
+ token = f'\x12\r\x12\x0b{video_id}\x18\x062\'"\x11"\x0b{video_id}0\x00x\x020\x00B\x10comments-section'
+ return base64.b64encode(token.encode()).decode()
+
def _get_comments(self, ytcfg, video_id, contents, webpage):
"""Entry for comment extraction"""
def _real_comment_extract(contents):
headers = self.generate_api_headers(
ytcfg=player_ytcfg, account_syncid=syncid, session_index=session_index, default_client=client)
- yt_query = {'videoId': video_id}
+ yt_query = {
+ 'videoId': video_id,
+ 'params': '8AEB' # enable stories
+ }
yt_query.update(self._generate_player_context(sts))
return self._extract_response(
item_id=video_id, ep='player', query=yt_query,
requested_clients = []
default = ['android', 'web']
allowed_clients = sorted(
- [client for client in INNERTUBE_CLIENTS.keys() if client[:1] != '_'],
+ (client for client in INNERTUBE_CLIENTS.keys() if client[:1] != '_'),
key=lambda client: INNERTUBE_CLIENTS[client]['priority'], reverse=True)
for client in self._configuration_arg('player_client'):
if client in allowed_clients:
return orderedSet(requested_clients)
- def _extract_player_ytcfg(self, client, video_id):
- url = {
- 'web_music': 'https://music.youtube.com',
- 'web_embedded': f'https://www.youtube.com/embed/{video_id}?html5=1'
- }.get(client)
- if not url:
- return {}
- webpage = self._download_webpage(url, video_id, fatal=False, note='Downloading %s config' % client.replace('_', ' ').strip())
- return self.extract_ytcfg(video_id, webpage) or {}
-
def _extract_player_responses(self, clients, video_id, webpage, master_ytcfg):
initial_pr = None
if webpage:
- initial_pr = self._extract_yt_initial_variable(
- webpage, self._YT_INITIAL_PLAYER_RESPONSE_RE,
- video_id, 'initial player response')
+ initial_pr = self._search_json(
+ self._YT_INITIAL_PLAYER_RESPONSE_RE, webpage, 'initial player response', video_id, fatal=False)
- original_clients = clients
+ all_clients = set(clients)
clients = clients[::-1]
prs = []
- def append_client(client_name):
- if client_name in INNERTUBE_CLIENTS and client_name not in original_clients:
- clients.append(client_name)
+ def append_client(*client_names):
+ """ Append the first client name that exists but not already used """
+ for client_name in client_names:
+ actual_client = _split_innertube_client(client_name)[0]
+ if actual_client in INNERTUBE_CLIENTS:
+ if actual_client not in all_clients:
+ clients.append(client_name)
+ all_clients.add(actual_client)
+ return
# Android player_response does not have microFormats which are needed for
# extraction of some data. So we return the initial_pr with formats
tried_iframe_fallback = False
player_url = None
while clients:
- client = clients.pop()
+ client, base_client, variant = _split_innertube_client(clients.pop())
player_ytcfg = master_ytcfg if client == 'web' else {}
- if 'configs' not in self._configuration_arg('player_skip'):
- player_ytcfg = self._extract_player_ytcfg(client, video_id) or player_ytcfg
+ if 'configs' not in self._configuration_arg('player_skip') and client != 'web':
+ player_ytcfg = self._download_ytcfg(client, video_id) or player_ytcfg
player_url = player_url or self._extract_player_url(master_ytcfg, player_ytcfg, webpage=webpage)
require_js_player = self._get_default_ytcfg(client).get('REQUIRE_JS_PLAYER')
prs.append(pr)
# creator clients can bypass AGE_VERIFICATION_REQUIRED if logged in
- if client.endswith('_agegate') and self._is_unplayable(pr) and self.is_authenticated:
- append_client(client.replace('_agegate', '_creator'))
+ if variant == 'embedded' and self._is_unplayable(pr) and self.is_authenticated:
+ append_client(f'{base_client}_creator')
elif self._is_agegated(pr):
- append_client(f'{client}_agegate')
+ if variant == 'tv_embedded':
+ append_client(f'{base_client}_embedded')
+ elif not variant:
+ append_client(f'tv_embedded.{base_client}', f'{base_client}_embedded')
if last_error:
if not len(prs):
self.report_warning(last_error)
return prs, player_url
- def _extract_formats(self, streaming_data, video_id, player_url, is_live):
+ def _extract_formats(self, streaming_data, video_id, player_url, is_live, duration):
itags, stream_ids = {}, []
itag_qualities, res_qualities = {}, {}
q = qualities([
'small', 'medium', 'large', 'hd720', 'hd1080', 'hd1440', 'hd2160', 'hd2880', 'highres'
])
streaming_formats = traverse_obj(streaming_data, (..., ('formats', 'adaptiveFormats'), ...), default=[])
- approx_duration = max(traverse_obj(streaming_formats, (..., 'approxDurationMs'), expected_type=float_or_none) or [0]) or None
for fmt in streaming_formats:
- if fmt.get('targetDurationSec') or fmt.get('drmFamilies'):
+ if fmt.get('targetDurationSec'):
continue
itag = str_or_none(fmt.get('itag'))
sc = compat_parse_qs(fmt.get('signatureCipher'))
fmt_url = url_or_none(try_get(sc, lambda x: x['url'][0]))
encrypted_sig = try_get(sc, lambda x: x['s'][0])
- if not (sc and fmt_url and encrypted_sig):
+ if not all((sc, fmt_url, player_url, encrypted_sig)):
continue
- if not player_url:
+ try:
+ fmt_url += '&%s=%s' % (
+ traverse_obj(sc, ('sp', -1)) or 'signature',
+ self._decrypt_signature(encrypted_sig, video_id, player_url)
+ )
+ except ExtractorError as e:
+ self.report_warning('Signature extraction failed: Some formats may be missing', only_once=True)
+ self.write_debug(e, only_once=True)
continue
- signature = self._decrypt_signature(sc['s'][0], video_id, player_url)
- sp = try_get(sc, lambda x: x['sp'][0]) or 'signature'
- fmt_url += '&' + sp + '=' + signature
query = parse_qs(fmt_url)
throttled = False
'n': self._decrypt_nsig(query['n'][0], video_id, player_url)})
except ExtractorError as e:
self.report_warning(
- f'nsig extraction failed: You may experience throttling for some formats\n'
- f'n = {query["n"][0]} ; player = {player_url}\n{e}', only_once=True)
+ 'nsig extraction failed: You may experience throttling for some formats\n'
+ f'n = {query["n"][0]} ; player = {player_url}', only_once=True)
+ self.write_debug(e, only_once=True)
throttled = True
if itag:
else -1)
# Some formats may have much smaller duration than others (possibly damaged during encoding)
# Eg: 2-nOtRESiUc Ref: https://github.com/yt-dlp/yt-dlp/issues/2823
- is_damaged = try_get(fmt, lambda x: float(x['approxDurationMs']) < approx_duration - 10000)
+ # Make sure to avoid false positives with small duration differences.
+ # Eg: __2ABJjxzNo, ySuUZEjARPY
+ is_damaged = try_get(fmt, lambda x: float(x['approxDurationMs']) / duration < 500)
+ if is_damaged:
+ self.report_warning(
+ f'{video_id}: Some formats are possibly damaged. They will be deprioritized', only_once=True)
dct = {
'asr': int_or_none(fmt.get('audioSampleRate')),
'filesize': int_or_none(fmt.get('contentLength')),
' (default)' if language_preference > 0 else ''),
fmt.get('qualityLabel') or quality.replace('audio_quality_', ''),
throttled and 'THROTTLED', is_damaged and 'DAMAGED', delim=', '),
- 'source_preference': -10 if throttled else -1,
+ # Format 22 is likely to be damaged. See https://github.com/yt-dlp/yt-dlp/issues/3372
+ 'source_preference': -10 if throttled else -5 if itag == '22' else -1,
'fps': int_or_none(fmt.get('fps')) or None,
'height': height,
'quality': q(quality),
+ 'has_drm': bool(fmt.get('drmFamilies')),
'tbr': tbr,
'url': fmt_url,
'width': int_or_none(fmt.get('width')),
'language': join_nonempty(audio_track.get('id', '').split('.')[0],
'desc' if language_preference < -1 else ''),
'language_preference': language_preference,
- 'preference': -10 if is_damaged else None,
+ # Strictly de-prioritize damaged and 3gp formats
+ 'preference': -10 if is_damaged else -2 if itag == '17' else None,
}
mime_mobj = re.match(
r'((?:[^/]+)/(?:[^;]+))(?:;\s*codecs="([^"]+)")?', fmt.get('mimeType') or '')
skip_manifests = self._configuration_arg('skip')
if not self.get_param('youtube_include_hls_manifest', True):
skip_manifests.append('hls')
+ if not self.get_param('youtube_include_dash_manifest', True):
+ skip_manifests.append('dash')
get_dash = 'dash' not in skip_manifests and (
not is_live or live_from_start or self._configuration_arg('include_live_dash'))
get_hls = not live_from_start and 'hls' not in skip_manifests
'width': width,
'height': height,
'fragments': [{
- 'path': url.replace('$M', str(j)),
+ 'url': url.replace('$M', str(j)),
'duration': min(fragment_duration, duration - (j * fragment_duration)),
} for j in range(math.ceil(fragment_count))],
}
webpage = None
if 'webpage' not in self._configuration_arg('player_skip'):
webpage = self._download_webpage(
- webpage_url + '&bpctr=9999999999&has_verified=1', video_id, fatal=False)
+ webpage_url + '&bpctr=9999999999&has_verified=1&pp=8AEB', video_id, fatal=False)
master_ytcfg = self.extract_ytcfg(video_id, webpage) or self._get_default_ytcfg()
return webpage, master_ytcfg, player_responses, player_url
- def _list_formats(self, video_id, microformats, video_details, player_responses, player_url):
+ def _list_formats(self, video_id, microformats, video_details, player_responses, player_url, duration=None):
live_broadcast_details = traverse_obj(microformats, (..., 'liveBroadcastDetails'))
is_live = get_first(video_details, 'isLive')
if is_live is None:
is_live = get_first(live_broadcast_details, 'isLiveNow')
streaming_data = traverse_obj(player_responses, (..., 'streamingData'), default=[])
- formats = list(self._extract_formats(streaming_data, video_id, player_url, is_live))
+ formats = list(self._extract_formats(streaming_data, video_id, player_url, is_live, duration))
return live_broadcast_details, is_live, streaming_data, formats
return self.playlist_result(
entries, video_id, video_title, video_description)
- live_broadcast_details, is_live, streaming_data, formats = self._list_formats(video_id, microformats, video_details, player_responses, player_url)
+ duration = int_or_none(
+ get_first(video_details, 'lengthSeconds')
+ or get_first(microformats, 'lengthSeconds')
+ or parse_duration(search_meta('duration'))) or None
+
+ if get_first(video_details, 'isPostLiveDvr'):
+ self.write_debug('Video is in Post-Live Manifestless mode')
+ if duration or 0 > 4 * 3600:
+ self.report_warning(
+ 'The livestream has not finished processing. Only 4 hours of the video can be currently downloaded. '
+ 'This is a known issue and patches are welcome')
+
+ live_broadcast_details, is_live, streaming_data, formats = self._list_formats(
+ video_id, microformats, video_details, player_responses, player_url, duration)
if not formats:
if not self.get_param('allow_unplayable_formats') and traverse_obj(streaming_data, (..., 'licenseInfos')):
original_thumbnails = thumbnails.copy()
# The best resolution thumbnails sometimes does not appear in the webpage
- # See: https://github.com/ytdl-org/youtube-dl/issues/29049, https://github.com/yt-dlp/yt-dlp/issues/340
+ # See: https://github.com/yt-dlp/yt-dlp/issues/340
# List of possible thumbnails - Ref: <https://stackoverflow.com/a/20542029>
thumbnail_names = [
- 'maxresdefault', 'hq720', 'sddefault', 'sd1', 'sd2', 'sd3',
- 'hqdefault', 'hq1', 'hq2', 'hq3', '0',
- 'mqdefault', 'mq1', 'mq2', 'mq3',
- 'default', '1', '2', '3'
+ # While the *1,*2,*3 thumbnails are just below their correspnding "*default" variants
+ # in resolution, these are not the custom thumbnail. So de-prioritize them
+ 'maxresdefault', 'hq720', 'sddefault', 'hqdefault', '0', 'mqdefault', 'default',
+ 'sd1', 'sd2', 'sd3', 'hq1', 'hq2', 'hq3', 'mq1', 'mq2', 'mq3', '1', '2', '3'
]
n_thumbnail_names = len(thumbnail_names)
thumbnails.extend({
get_first(video_details, 'channelId')
or get_first(microformats, 'externalChannelId')
or search_meta('channelId'))
- duration = int_or_none(
- get_first(video_details, 'lengthSeconds')
- or get_first(microformats, 'lengthSeconds')
- or parse_duration(search_meta('duration'))) or None
owner_profile_url = get_first(microformats, 'ownerProfileUrl')
live_content = get_first(video_details, 'isLiveContent')
# URL checking if user don't care about getting the best possible thumbnail
'thumbnail': traverse_obj(original_thumbnails, (-1, 'url')),
'description': video_description,
- 'upload_date': unified_strdate(
- get_first(microformats, 'uploadDate')
- or search_meta('uploadDate')),
'uploader': get_first(video_details, 'author'),
'uploader_id': self._search_regex(r'/(?:channel|user)/([^/?&#]+)', owner_profile_url, 'uploader id') if owner_profile_url else None,
'uploader_url': owner_profile_url,
subtitles, automatic_captions = {}, {}
for lang_code, caption_track in captions.items():
base_url = caption_track.get('baseUrl')
+ orig_lang = parse_qs(base_url).get('lang', [None])[-1]
if not base_url:
continue
lang_name = self._get_text(caption_track, 'name', max_runs=1)
for trans_code, trans_name in translation_languages.items():
if not trans_code:
continue
+ orig_trans_code = trans_code
if caption_track.get('kind') != 'asr':
+ if 'translated_subs' in self._configuration_arg('skip'):
+ continue
trans_code += f'-{lang_code}'
trans_name += format_field(lang_name, template=' from %s')
- process_language(
- automatic_captions, base_url, trans_code, trans_name, {'tlang': trans_code})
- if lang_code == f'a-{trans_code}':
+ # Add an "-orig" label to the original language so that it can be distinguished.
+ # The subs are returned without "-orig" as well for compatibility
+ if lang_code == f'a-{orig_trans_code}':
process_language(
- automatic_captions, base_url, f'{trans_code}-orig', f'{trans_name} (Original)', {'tlang': trans_code})
+ automatic_captions, base_url, f'{trans_code}-orig', f'{trans_name} (Original)', {})
+ # Setting tlang=lang returns damaged subtitles.
+ process_language(automatic_captions, base_url, trans_code, trans_name,
+ {} if orig_lang == orig_trans_code else {'tlang': trans_code})
info['automatic_captions'] = automatic_captions
info['subtitles'] = subtitles
# Youtube Music Auto-generated description
if video_description:
- mobj = re.search(r'(?s)(?P<track>[^·\n]+)·(?P<artist>[^\n]+)\n+(?P<album>[^\n]+)(?:.+?℗\s*(?P<release_year>\d{4})(?!\d))?(?:.+?Released on\s*:\s*(?P<release_date>\d{4}-\d{2}-\d{2}))?(.+?\nArtist\s*:\s*(?P<clean_artist>[^\n]+))?.+\nAuto-generated by YouTube\.\s*$', video_description)
+ mobj = re.search(
+ r'''(?xs)
+ (?P<track>[^·\n]+)·(?P<artist>[^\n]+)\n+
+ (?P<album>[^\n]+)
+ (?:.+?℗\s*(?P<release_year>\d{4})(?!\d))?
+ (?:.+?Released on\s*:\s*(?P<release_date>\d{4}-\d{2}-\d{2}))?
+ (.+?\nArtist\s*:\s*(?P<clean_artist>[^\n]+))?
+ .+\nAuto-generated\ by\ YouTube\.\s*$
+ ''', video_description)
if mobj:
release_year = mobj.group('release_year')
release_date = mobj.group('release_date')
initial_data = None
if webpage:
- initial_data = self._extract_yt_initial_variable(
- webpage, self._YT_INITIAL_DATA_RE, video_id,
- 'yt initial data')
+ initial_data = self._search_json(
+ self._YT_INITIAL_DATA_RE, webpage, 'yt initial data', video_id, fatal=False)
if not initial_data:
query = {'videoId': video_id}
query.update(self._get_checkok_params())
headers=self.generate_api_headers(ytcfg=master_ytcfg),
note='Downloading initial data API JSON')
- try:
- # This will error if there is no livechat
+ try: # This will error if there is no livechat
initial_data['contents']['twoColumnWatchNextResults']['conversationBar']['liveChatRenderer']['continuations'][0]['reloadContinuationData']['continuation']
+ except (KeyError, IndexError, TypeError):
+ pass
+ else:
info.setdefault('subtitles', {})['live_chat'] = [{
- 'url': 'https://www.youtube.com/watch?v=%s' % video_id, # url is needed to set cookies
+ 'url': f'https://www.youtube.com/watch?v={video_id}', # url is needed to set cookies
'video_id': video_id,
'ext': 'json',
'protocol': 'youtube_live_chat' if is_live or is_upcoming else 'youtube_live_chat_replay',
}]
- except (KeyError, IndexError, TypeError):
- pass
if initial_data:
info['chapters'] = (
self._extract_chapters_from_json(initial_data, duration)
or self._extract_chapters_from_engagement_panel(initial_data, duration)
+ or self._extract_chapters_from_description(video_description, duration)
or None)
- contents = try_get(
- initial_data,
- lambda x: x['contents']['twoColumnWatchNextResults']['results']['results']['contents'],
- list) or []
- for content in contents:
- vpir = content.get('videoPrimaryInfoRenderer')
- if vpir:
- stl = vpir.get('superTitleLink')
- if stl:
- stl = self._get_text(stl)
- if try_get(
- vpir,
- lambda x: x['superTitleIcon']['iconType']) == 'LOCATION_PIN':
- info['location'] = stl
- else:
- mobj = re.search(r'(.+?)\s*S(\d+)\s*•\s*E(\d+)', stl)
- if mobj:
- info.update({
- 'series': mobj.group(1),
- 'season_number': int(mobj.group(2)),
- 'episode_number': int(mobj.group(3)),
- })
- for tlb in (try_get(
- vpir,
- lambda x: x['videoActions']['menuRenderer']['topLevelButtons'],
- list) or []):
- tbr = tlb.get('toggleButtonRenderer') or {}
- for getter, regex in [(
- lambda x: x['defaultText']['accessibility']['accessibilityData'],
- r'(?P<count>[\d,]+)\s*(?P<type>(?:dis)?like)'), ([
- lambda x: x['accessibility'],
- lambda x: x['accessibilityData']['accessibilityData'],
- ], r'(?P<type>(?:dis)?like) this video along with (?P<count>[\d,]+) other people')]:
- label = (try_get(tbr, getter, dict) or {}).get('label')
- if label:
- mobj = re.match(regex, label)
- if mobj:
- info[mobj.group('type') + '_count'] = str_to_int(mobj.group('count'))
- break
- sbr_tooltip = try_get(
- vpir, lambda x: x['sentimentBar']['sentimentBarRenderer']['tooltip'])
- if sbr_tooltip:
- like_count, dislike_count = sbr_tooltip.split(' / ')
+ contents = traverse_obj(
+ initial_data, ('contents', 'twoColumnWatchNextResults', 'results', 'results', 'contents'),
+ expected_type=list, default=[])
+
+ vpir = get_first(contents, 'videoPrimaryInfoRenderer')
+ if vpir:
+ stl = vpir.get('superTitleLink')
+ if stl:
+ stl = self._get_text(stl)
+ if try_get(
+ vpir,
+ lambda x: x['superTitleIcon']['iconType']) == 'LOCATION_PIN':
+ info['location'] = stl
+ else:
+ mobj = re.search(r'(.+?)\s*S(\d+)\s*•?\s*E(\d+)', stl)
+ if mobj:
info.update({
- 'like_count': str_to_int(like_count),
- 'dislike_count': str_to_int(dislike_count),
+ 'series': mobj.group(1),
+ 'season_number': int(mobj.group(2)),
+ 'episode_number': int(mobj.group(3)),
})
- vsir = content.get('videoSecondaryInfoRenderer')
- if vsir:
- vor = traverse_obj(vsir, ('owner', 'videoOwnerRenderer'))
- info.update({
- 'channel': self._get_text(vor, 'title'),
- 'channel_follower_count': self._get_count(vor, 'subscriberCountText')})
-
- rows = try_get(
- vsir,
- lambda x: x['metadataRowContainer']['metadataRowContainerRenderer']['rows'],
- list) or []
- multiple_songs = False
- for row in rows:
- if try_get(row, lambda x: x['metadataRowRenderer']['hasDividerLine']) is True:
- multiple_songs = True
+ for tlb in (try_get(
+ vpir,
+ lambda x: x['videoActions']['menuRenderer']['topLevelButtons'],
+ list) or []):
+ tbr = tlb.get('toggleButtonRenderer') or {}
+ for getter, regex in [(
+ lambda x: x['defaultText']['accessibility']['accessibilityData'],
+ r'(?P<count>[\d,]+)\s*(?P<type>(?:dis)?like)'), ([
+ lambda x: x['accessibility'],
+ lambda x: x['accessibilityData']['accessibilityData'],
+ ], r'(?P<type>(?:dis)?like) this video along with (?P<count>[\d,]+) other people')]:
+ label = (try_get(tbr, getter, dict) or {}).get('label')
+ if label:
+ mobj = re.match(regex, label)
+ if mobj:
+ info[mobj.group('type') + '_count'] = str_to_int(mobj.group('count'))
break
- for row in rows:
- mrr = row.get('metadataRowRenderer') or {}
- mrr_title = mrr.get('title')
- if not mrr_title:
- continue
- mrr_title = self._get_text(mrr, 'title')
- mrr_contents_text = self._get_text(mrr, ('contents', 0))
- if mrr_title == 'License':
- info['license'] = mrr_contents_text
- elif not multiple_songs:
- if mrr_title == 'Album':
- info['album'] = mrr_contents_text
- elif mrr_title == 'Artist':
- info['artist'] = mrr_contents_text
- elif mrr_title == 'Song':
- info['track'] = mrr_contents_text
+ sbr_tooltip = try_get(
+ vpir, lambda x: x['sentimentBar']['sentimentBarRenderer']['tooltip'])
+ if sbr_tooltip:
+ like_count, dislike_count = sbr_tooltip.split(' / ')
+ info.update({
+ 'like_count': str_to_int(like_count),
+ 'dislike_count': str_to_int(dislike_count),
+ })
+ vsir = get_first(contents, 'videoSecondaryInfoRenderer')
+ if vsir:
+ vor = traverse_obj(vsir, ('owner', 'videoOwnerRenderer'))
+ info.update({
+ 'channel': self._get_text(vor, 'title'),
+ 'channel_follower_count': self._get_count(vor, 'subscriberCountText')})
+
+ rows = try_get(
+ vsir,
+ lambda x: x['metadataRowContainer']['metadataRowContainerRenderer']['rows'],
+ list) or []
+ multiple_songs = False
+ for row in rows:
+ if try_get(row, lambda x: x['metadataRowRenderer']['hasDividerLine']) is True:
+ multiple_songs = True
+ break
+ for row in rows:
+ mrr = row.get('metadataRowRenderer') or {}
+ mrr_title = mrr.get('title')
+ if not mrr_title:
+ continue
+ mrr_title = self._get_text(mrr, 'title')
+ mrr_contents_text = self._get_text(mrr, ('contents', 0))
+ if mrr_title == 'License':
+ info['license'] = mrr_contents_text
+ elif not multiple_songs:
+ if mrr_title == 'Album':
+ info['album'] = mrr_contents_text
+ elif mrr_title == 'Artist':
+ info['artist'] = mrr_contents_text
+ elif mrr_title == 'Song':
+ info['track'] = mrr_contents_text
fallbacks = {
'channel': 'uploader',
'channel_id': 'uploader_id',
'channel_url': 'uploader_url',
}
+
+ # The upload date for scheduled, live and past live streams / premieres in microformats
+ # may be different from the stream date. Although not in UTC, we will prefer it in this case.
+ # See: https://github.com/yt-dlp/yt-dlp/pull/2223#issuecomment-1008485139
+ upload_date = (
+ unified_strdate(get_first(microformats, 'uploadDate'))
+ or unified_strdate(search_meta('uploadDate')))
+ if not upload_date or (not info.get('is_live') and not info.get('was_live') and info.get('live_status') != 'is_upcoming'):
+ upload_date = strftime_or_none(self._extract_time_text(vpir, 'dateText')[0], '%Y%m%d') or upload_date
+ info['upload_date'] = upload_date
+
for to, frm in fallbacks.items():
if not info.get(to):
info[to] = info.get(frm)
def _extract_basic_item_renderer(item):
# Modified from _extract_grid_item_renderer
known_basic_renderers = (
- 'playlistRenderer', 'videoRenderer', 'channelRenderer', 'showRenderer'
+ 'playlistRenderer', 'videoRenderer', 'channelRenderer', 'showRenderer', 'reelItemRenderer'
)
for key, renderer in item.items():
if not isinstance(renderer, dict):
# TODO: add support for nested playlists so each shelf is processed
# as separate playlist
# TODO: this includes only first N items
- for entry in self._grid_entries(renderer):
- yield entry
+ yield from self._grid_entries(renderer)
renderer = content.get('horizontalListRenderer')
if renderer:
# TODO
title = self._get_text(shelf_renderer, 'title')
yield self.url_result(shelf_url, video_title=title)
# Shelf may not contain shelf URL, fallback to extraction from content
- for entry in self._shelf_entries_from_content(shelf_renderer):
- yield entry
+ yield from self._shelf_entries_from_content(shelf_renderer)
def _playlist_entries(self, video_list_renderer):
for content in video_list_renderer['contents']:
if video_id:
return self._extract_video(video_renderer)
+ def _hashtag_tile_entry(self, hashtag_tile_renderer):
+ url = urljoin('https://youtube.com', traverse_obj(
+ hashtag_tile_renderer, ('onTapCommand', 'commandMetadata', 'webCommandMetadata', 'url')))
+ if url:
+ return self.url_result(
+ url, ie=YoutubeTabIE.ie_key(), title=self._get_text(hashtag_tile_renderer, 'hashtag'))
+
def _post_thread_entries(self, post_thread_renderer):
post_renderer = try_get(
post_thread_renderer, lambda x: x['post']['backstagePostRenderer'], dict)
return
for content in contents:
renderer = content.get('backstagePostThreadRenderer')
- if not isinstance(renderer, dict):
+ if isinstance(renderer, dict):
+ yield from self._post_thread_entries(renderer)
continue
- for entry in self._post_thread_entries(renderer):
- yield entry
+ renderer = content.get('videoRenderer')
+ if isinstance(renderer, dict):
+ yield self._video_entry(renderer)
r''' # unused
def _rich_grid_entries(self, contents):
if entry:
yield entry
'''
+
def _extract_entries(self, parent_renderer, continuation_list):
# continuation_list is modified in-place with continuation_list = [continuation_token]
continuation_list[:] = [None]
known_renderers = {
'playlistVideoListRenderer': self._playlist_entries,
'gridRenderer': self._grid_entries,
- 'shelfRenderer': lambda x: self._shelf_entries(x),
+ 'reelShelfRenderer': self._grid_entries,
+ 'shelfRenderer': self._shelf_entries,
'musicResponsiveListItemRenderer': lambda x: [self._music_reponsive_list_entry(x)],
'backstagePostThreadRenderer': self._post_thread_entries,
'videoRenderer': lambda x: [self._video_entry(x)],
'playlistRenderer': lambda x: self._grid_entries({'items': [{'playlistRenderer': x}]}),
'channelRenderer': lambda x: self._grid_entries({'items': [{'channelRenderer': x}]}),
+ 'hashtagTileRenderer': lambda x: [self._hashtag_tile_entry(x)]
}
for key, renderer in isr_content.items():
if key not in known_renderers:
parent_renderer = (
try_get(tab_content, lambda x: x['sectionListRenderer'], dict)
or try_get(tab_content, lambda x: x['richGridRenderer'], dict) or {})
- for entry in extract_entries(parent_renderer):
- yield entry
+ yield from extract_entries(parent_renderer)
continuation = continuation_list[0]
for page_num in itertools.count(1):
headers = self.generate_api_headers(
ytcfg=ytcfg, account_syncid=account_syncid, visitor_data=visitor_data)
response = self._extract_response(
- item_id='%s page %s' % (item_id, page_num),
+ item_id=f'{item_id} page {page_num}',
query=continuation, headers=headers, ytcfg=ytcfg,
check_get_keys=('continuationContents', 'onResponseReceivedActions', 'onResponseReceivedEndpoints'))
continue
continuation_renderer = value
continuation_list = [None]
- for entry in known_continuation_renderers[key](continuation_renderer):
- yield entry
+ yield from known_continuation_renderers[key](continuation_renderer)
continuation = continuation_list[0] or self._extract_continuation(continuation_renderer)
break
if continuation_renderer:
continue
known_renderers = {
+ 'videoRenderer': (self._grid_entries, 'items'), # for membership tab
'gridPlaylistRenderer': (self._grid_entries, 'items'),
'gridVideoRenderer': (self._grid_entries, 'items'),
'gridChannelRenderer': (self._grid_entries, 'items'),
continue
video_items_renderer = {known_renderers[key][1]: continuation_items}
continuation_list = [None]
- for entry in known_renderers[key][0](video_items_renderer):
- yield entry
+ yield from known_renderers[key][0](video_items_renderer)
continuation = continuation_list[0] or self._extract_continuation(video_items_renderer)
break
if video_items_renderer:
break
@staticmethod
- def _extract_selected_tab(tabs):
+ def _extract_selected_tab(tabs, fatal=True):
for tab in tabs:
renderer = dict_get(tab, ('tabRenderer', 'expandableTabRenderer')) or {}
if renderer.get('selected') is True:
return renderer
else:
- raise ExtractorError('Unable to find selected tab')
+ if fatal:
+ raise ExtractorError('Unable to find selected tab')
- @classmethod
- def _extract_uploader(cls, data):
+ def _extract_uploader(self, data):
uploader = {}
- renderer = cls._extract_sidebar_info_renderer(data, 'playlistSidebarSecondaryInfoRenderer') or {}
+ renderer = self._extract_sidebar_info_renderer(data, 'playlistSidebarSecondaryInfoRenderer') or {}
owner = try_get(
renderer, lambda x: x['videoOwner']['videoOwnerRenderer']['title']['runs'][0], dict)
if owner:
- uploader['uploader'] = owner.get('text')
+ owner_text = owner.get('text')
+ uploader['uploader'] = self._search_regex(
+ r'^by (.+) and \d+ others?$', owner_text, 'uploader', default=owner_text)
uploader['uploader_id'] = try_get(
owner, lambda x: x['navigationEndpoint']['browseEndpoint']['browseId'], compat_str)
uploader['uploader_url'] = urljoin(
})
primary_thumbnails = self._extract_thumbnails(
- primary_sidebar_renderer, ('thumbnailRenderer', 'playlistVideoThumbnailRenderer', 'thumbnail'))
+ primary_sidebar_renderer, ('thumbnailRenderer', ('playlistVideoThumbnailRenderer', 'playlistCustomThumbnailRenderer'), 'thumbnail'))
if playlist_id is None:
playlist_id = item_id
self._extract_visitor_data(data, ytcfg)),
**metadata)
- def _extract_mix_playlist(self, playlist, playlist_id, data, ytcfg):
+ def _extract_inline_playlist(self, playlist, playlist_id, data, ytcfg):
first_id = last_id = response = None
for page_num in itertools.count(1):
videos = list(self._playlist_entries(playlist))
start = next((i for i, v in enumerate(videos) if v['id'] == last_id), -1) + 1
if start >= len(videos):
return
- for video in videos[start:]:
- if video['id'] == first_id:
- self.to_screen('First video %s found again; Assuming end of Mix' % first_id)
- return
- yield video
+ yield from videos[start:]
first_id = first_id or videos[0]['id']
last_id = videos[-1]['id']
watch_endpoint = try_get(
playlist_url = urljoin(url, try_get(
playlist, lambda x: x['endpoint']['commandMetadata']['webCommandMetadata']['url'],
compat_str))
- if playlist_url and playlist_url != url:
+
+ # Some playlists are unviewable but YouTube still provides a link to the (broken) playlist page [1]
+ # [1] MLCT, RLTDwFCb4jeqaKWnciAYM-ZVHg
+ is_known_unviewable = re.fullmatch(r'MLCT|RLTD[\w-]{22}', playlist_id)
+
+ if playlist_url and playlist_url != url and not is_known_unviewable:
return self.url_result(
playlist_url, ie=YoutubeTabIE.ie_key(), video_id=playlist_id,
video_title=title)
return self.playlist_result(
- self._extract_mix_playlist(playlist, playlist_id, data, ytcfg),
+ self._extract_inline_playlist(playlist, playlist_id, data, ytcfg),
playlist_id=playlist_id, playlist_title=title)
def _extract_availability(self, data):
check_get_keys='contents', fatal=False, ytcfg=ytcfg,
note='Downloading API JSON with unavailable videos')
+ @functools.cached_property
+ def skip_webpage(self):
+ return 'webpage' in self._configuration_arg('skip', ie_key=YoutubeTabIE.ie_key())
+
def _extract_webpage(self, url, item_id, fatal=True):
retries = self.get_param('extractor_retries', 3)
count = -1
self.report_warning(error_to_compat_str(e))
break
- if dict_get(data, ('contents', 'currentVideoEndpoint')):
+ if dict_get(data, ('contents', 'currentVideoEndpoint', 'onResponseReceivedActions')):
break
last_error = 'Incomplete yt initial data received'
return webpage, data
+ def _report_playlist_authcheck(self, ytcfg, fatal=True):
+ """Use if failed to extract ytcfg (and data) from initial webpage"""
+ if not ytcfg and self.is_authenticated:
+ msg = 'Playlists that require authentication may not extract correctly without a successful webpage download'
+ if 'authcheck' not in self._configuration_arg('skip', ie_key=YoutubeTabIE.ie_key()) and fatal:
+ raise ExtractorError(
+ f'{msg}. If you are not downloading private content, or '
+ 'your cookies are only for the first account and channel,'
+ ' pass "--extractor-args youtubetab:skip=authcheck" to skip this check',
+ expected=True)
+ self.report_warning(msg, only_once=True)
+
def _extract_data(self, url, item_id, ytcfg=None, fatal=True, webpage_fatal=False, default_client='web'):
data = None
- if 'webpage' not in self._configuration_arg('skip'):
+ if not self.skip_webpage:
webpage, data = self._extract_webpage(url, item_id, fatal=webpage_fatal)
ytcfg = ytcfg or self.extract_ytcfg(item_id, webpage)
- if not data:
- if not ytcfg and self.is_authenticated:
- msg = 'Playlists that require authentication may not extract correctly without a successful webpage download.'
- if 'authcheck' not in self._configuration_arg('skip') and fatal:
- raise ExtractorError(
- msg + ' If you are not downloading private content, or your cookies are only for the first account and channel,'
- ' pass "--extractor-args youtubetab:skip=authcheck" to skip this check',
- expected=True)
+ # Reject webpage data if redirected to home page without explicitly requesting
+ selected_tab = self._extract_selected_tab(traverse_obj(
+ data, ('contents', 'twoColumnBrowseResultsRenderer', 'tabs'), expected_type=list, default=[]), fatal=False) or {}
+ if (url != 'https://www.youtube.com/feed/recommended'
+ and selected_tab.get('tabIdentifier') == 'FEwhat_to_watch' # Home page
+ and 'no-youtube-channel-redirect' not in self.get_param('compat_opts', [])):
+ msg = 'The channel/playlist does not exist and the URL redirected to youtube.com home page'
+ if fatal:
+ raise ExtractorError(msg, expected=True)
self.report_warning(msg, only_once=True)
+ if not data:
+ self._report_playlist_authcheck(ytcfg, fatal=fatal)
data = self._extract_tab_endpoint(url, item_id, ytcfg, fatal=fatal, default_client=default_client)
return data, ytcfg
return self._extract_response(
item_id=item_id, query=params, ep=ep, headers=headers,
ytcfg=ytcfg, fatal=fatal, default_client=default_client,
- check_get_keys=('contents', 'currentVideoEndpoint'))
+ check_get_keys=('contents', 'currentVideoEndpoint', 'onResponseReceivedActions'))
err_note = 'Failed to resolve url (does the playlist exist?)'
if fatal:
raise ExtractorError(err_note, expected=True)
('contents', 'tabbedSearchResultsRenderer', 'tabs', 0, 'tabRenderer', 'content', 'sectionListRenderer', 'contents'),
('continuationContents', ),
)
- check_get_keys = tuple(set(keys[0] for keys in content_keys))
+ display_id = f'query "{query}"'
+ check_get_keys = tuple({keys[0] for keys in content_keys})
+ ytcfg = self._download_ytcfg(default_client, display_id) if not self.skip_webpage else {}
+ self._report_playlist_authcheck(ytcfg, fatal=False)
continuation_list = [None]
+ search = None
for page_num in itertools.count(1):
data.update(continuation_list[0] or {})
+ headers = self.generate_api_headers(
+ ytcfg=ytcfg, visitor_data=self._extract_visitor_data(search), default_client=default_client)
search = self._extract_response(
- item_id='query "%s" page %s' % (query, page_num), ep='search', query=data,
- default_client=default_client, check_get_keys=check_get_keys)
+ item_id=f'{display_id} page {page_num}', ep='search', query=data,
+ default_client=default_client, check_get_keys=check_get_keys, ytcfg=ytcfg, headers=headers)
slr_contents = traverse_obj(search, *content_keys)
yield from self._extract_entries({'contents': list(variadic(slr_contents))}, continuation_list)
if not continuation_list[0]:
'skip_download': True,
'extractor_args': {'youtubetab': {'skip': ['webpage']}}
},
+ }, {
+ 'note': 'non-standard redirect to regional channel',
+ 'url': 'https://www.youtube.com/channel/UCwVVpHQ2Cs9iGJfpdFngePQ',
+ 'only_matching': True
+ }, {
+ 'note': 'collaborative playlist (uploader name in the form "by <uploader> and x other(s)")',
+ 'url': 'https://www.youtube.com/playlist?list=PLx-_-Kk4c89oOHEDQAojOXzEzemXxoqx6',
+ 'info_dict': {
+ 'id': 'PLx-_-Kk4c89oOHEDQAojOXzEzemXxoqx6',
+ 'modified_date': '20220407',
+ 'channel_url': 'https://www.youtube.com/channel/UCKcqXmCcyqnhgpA5P0oHH_Q',
+ 'tags': [],
+ 'uploader_id': 'UCKcqXmCcyqnhgpA5P0oHH_Q',
+ 'uploader': 'pukkandan',
+ 'availability': 'unlisted',
+ 'channel_id': 'UCKcqXmCcyqnhgpA5P0oHH_Q',
+ 'channel': 'pukkandan',
+ 'description': 'Test for collaborative playlist',
+ 'title': 'yt-dlp test - collaborative playlist',
+ 'uploader_url': 'https://www.youtube.com/channel/UCKcqXmCcyqnhgpA5P0oHH_Q',
+ },
+ 'playlist_mincount': 2
}]
@classmethod
def suitable(cls, url):
- return False if YoutubeIE.suitable(url) else super(
- YoutubeTabIE, cls).suitable(url)
+ return False if YoutubeIE.suitable(url) else super().suitable(url)
_URL_RE = re.compile(rf'(?P<pre>{_VALID_URL})(?(not_channel)|(?P<tab>/\w+))?(?P<post>.*)$')
# Handle both video/playlist URLs
qs = parse_qs(url)
- video_id, playlist_id = [qs.get(key, [None])[0] for key in ('v', 'list')]
+ video_id, playlist_id = (qs.get(key, [None])[0] for key in ('v', 'list'))
if not video_id and mobj['not_channel'].startswith('watch'):
if not playlist_id:
data, ytcfg = self._extract_data(url, item_id)
+ # YouTube may provide a non-standard redirect to the regional channel
+ # See: https://github.com/yt-dlp/yt-dlp/issues/2694
+ redirect_url = traverse_obj(
+ data, ('onResponseReceivedActions', ..., 'navigateAction', 'endpoint', 'commandMetadata', 'webCommandMetadata', 'url'), get_all=False)
+ if redirect_url and 'no-youtube-channel-redirect' not in compat_opts:
+ redirect_url = ''.join((
+ urljoin('https://www.youtube.com', redirect_url), mobj['tab'], mobj['post']))
+ self.to_screen(f'This playlist is likely not available in your region. Following redirect to regional playlist {redirect_url}')
+ return self.url_result(redirect_url, ie=YoutubeTabIE.ie_key())
+
tabs = traverse_obj(data, ('contents', 'twoColumnBrowseResultsRenderer', 'tabs'), expected_type=list)
if tabs:
selected_tab = self._extract_selected_tab(tabs)
qs = parse_qs(url)
if qs.get('v', [None])[0]:
return False
- return super(YoutubePlaylistIE, cls).suitable(url)
+ return super().suitable(url)
def _real_extract(self, url):
playlist_id = self._match_id(url)
ie=YoutubeTabIE.ie_key())
+class YoutubeNotificationsIE(YoutubeTabBaseInfoExtractor):
+ IE_NAME = 'youtube:notif'
+ IE_DESC = 'YouTube notifications; ":ytnotif" keyword (requires cookies)'
+ _VALID_URL = r':ytnotif(?:ication)?s?'
+ _LOGIN_REQUIRED = True
+ _TESTS = [{
+ 'url': ':ytnotif',
+ 'only_matching': True,
+ }, {
+ 'url': ':ytnotifications',
+ 'only_matching': True,
+ }]
+
+ def _extract_notification_menu(self, response, continuation_list):
+ notification_list = traverse_obj(
+ response,
+ ('actions', 0, 'openPopupAction', 'popup', 'multiPageMenuRenderer', 'sections', 0, 'multiPageMenuNotificationSectionRenderer', 'items'),
+ ('actions', 0, 'appendContinuationItemsAction', 'continuationItems'),
+ expected_type=list) or []
+ continuation_list[0] = None
+ for item in notification_list:
+ entry = self._extract_notification_renderer(item.get('notificationRenderer'))
+ if entry:
+ yield entry
+ continuation = item.get('continuationItemRenderer')
+ if continuation:
+ continuation_list[0] = continuation
+
+ def _extract_notification_renderer(self, notification):
+ video_id = traverse_obj(
+ notification, ('navigationEndpoint', 'watchEndpoint', 'videoId'), expected_type=str)
+ url = f'https://www.youtube.com/watch?v={video_id}'
+ channel_id = None
+ if not video_id:
+ browse_ep = traverse_obj(
+ notification, ('navigationEndpoint', 'browseEndpoint'), expected_type=dict)
+ channel_id = traverse_obj(browse_ep, 'browseId', expected_type=str)
+ post_id = self._search_regex(
+ r'/post/(.+)', traverse_obj(browse_ep, 'canonicalBaseUrl', expected_type=str),
+ 'post id', default=None)
+ if not channel_id or not post_id:
+ return
+ # The direct /post url redirects to this in the browser
+ url = f'https://www.youtube.com/channel/{channel_id}/community?lb={post_id}'
+
+ channel = traverse_obj(
+ notification, ('contextualMenu', 'menuRenderer', 'items', 1, 'menuServiceItemRenderer', 'text', 'runs', 1, 'text'),
+ expected_type=str)
+ notification_title = self._get_text(notification, 'shortMessage')
+ if notification_title:
+ notification_title = notification_title.replace('\xad', '') # remove soft hyphens
+ # TODO: handle recommended videos
+ title = self._search_regex(
+ rf'{re.escape(channel or "")}[^:]+: (.+)', notification_title,
+ 'video title', default=None)
+ upload_date = (strftime_or_none(self._extract_time_text(notification, 'sentTimeText')[0], '%Y%m%d')
+ if self._configuration_arg('approximate_date', ie_key=YoutubeTabIE.ie_key())
+ else None)
+ return {
+ '_type': 'url',
+ 'url': url,
+ 'ie_key': (YoutubeIE if video_id else YoutubeTabIE).ie_key(),
+ 'video_id': video_id,
+ 'title': title,
+ 'channel_id': channel_id,
+ 'channel': channel,
+ 'thumbnails': self._extract_thumbnails(notification, 'videoThumbnail'),
+ 'upload_date': upload_date,
+ }
+
+ def _notification_menu_entries(self, ytcfg):
+ continuation_list = [None]
+ response = None
+ for page in itertools.count(1):
+ ctoken = traverse_obj(
+ continuation_list, (0, 'continuationEndpoint', 'getNotificationMenuEndpoint', 'ctoken'), expected_type=str)
+ response = self._extract_response(
+ item_id=f'page {page}', query={'ctoken': ctoken} if ctoken else {}, ytcfg=ytcfg,
+ ep='notification/get_notification_menu', check_get_keys='actions',
+ headers=self.generate_api_headers(ytcfg=ytcfg, visitor_data=self._extract_visitor_data(response)))
+ yield from self._extract_notification_menu(response, continuation_list)
+ if not continuation_list[0]:
+ break
+
+ def _real_extract(self, url):
+ display_id = 'notifications'
+ ytcfg = self._download_ytcfg('web', display_id) if not self.skip_webpage else {}
+ self._report_playlist_authcheck(ytcfg)
+ return self.playlist_result(self._notification_menu_entries(ytcfg), display_id, display_id)
+
+
class YoutubeSearchIE(YoutubeTabBaseInfoExtractor, SearchInfoExtractor):
IE_DESC = 'YouTube search'
IE_NAME = 'youtube:search'
'id': 'python',
'title': 'python',
}
-
+ }, {
+ 'url': 'https://www.youtube.com/results?search_query=%23cats',
+ 'playlist_mincount': 1,
+ 'info_dict': {
+ 'id': '#cats',
+ 'title': '#cats',
+ 'entries': [{
+ 'url': r're:https://(www\.)?youtube\.com/hashtag/cats',
+ 'title': '#cats',
+ }],
+ },
}, {
'url': 'https://www.youtube.com/results?q=test&sp=EgQIBBgB',
'only_matching': True,
class YoutubeFeedsInfoExtractor(InfoExtractor):
"""
Base class for feed extractors
- Subclasses must define the _FEED_NAME property.
+ Subclasses must re-define the _FEED_NAME property.
"""
_LOGIN_REQUIRED = True
- _TESTS = []
+ _FEED_NAME = 'feeds'
- @property
+ def _real_initialize(self):
+ YoutubeBaseInfoExtractor._check_login_required(self)
+
+ @classproperty
def IE_NAME(self):
- return 'youtube:%s' % self._FEED_NAME
+ return f'youtube:{self._FEED_NAME}'
def _real_extract(self, url):
return self.url_result(
}]
+class YoutubeStoriesIE(InfoExtractor):
+ IE_DESC = 'YouTube channel stories; "ytstories:" prefix'
+ IE_NAME = 'youtube:stories'
+ _VALID_URL = r'ytstories:UC(?P<id>[A-Za-z0-9_-]{21}[AQgw])$'
+ _TESTS = [{
+ 'url': 'ytstories:UCwFCb4jeqaKWnciAYM-ZVHg',
+ 'only_matching': True,
+ }]
+
+ def _real_extract(self, url):
+ playlist_id = f'RLTD{self._match_id(url)}'
+ return self.url_result(
+ f'https://www.youtube.com/playlist?list={playlist_id}&playnext=1',
+ ie=YoutubeTabIE, video_id=playlist_id)
+
+
class YoutubeTruncatedURLIE(InfoExtractor):
IE_NAME = 'youtube:truncated_url'
IE_DESC = False # Do not list
def _real_extract(self, url):
video_id = self._match_id(url)
raise ExtractorError(
- 'Incomplete YouTube ID %s. URL %s looks truncated.' % (video_id, url),
+ f'Incomplete YouTube ID {video_id}. URL {url} looks truncated.',
expected=True)
projects / yt-dlp.git / blobdiff