)
from ..jsinterp import JSInterpreter
from ..utils import (
- bool_or_none,
bytes_to_intlist,
clean_html,
- dict_get,
datetime_from_str,
+ dict_get,
error_to_compat_str,
ExtractorError,
- format_field,
float_or_none,
+ format_field,
int_or_none,
intlist_to_bytes,
mimetype2ext,
+ network_exceptions,
+ orderedSet,
parse_codecs,
parse_count,
parse_duration,
+ parse_iso8601,
qualities,
remove_start,
smuggle_url,
str_or_none,
str_to_int,
+ traverse_obj,
try_get,
unescapeHTML,
unified_strdate,
url_or_none,
urlencode_postdata,
urljoin,
- variadic
+ variadic,
)
self.report_warning(message)
# username+password login is broken
- if self._LOGIN_REQUIRED and self.get_param('cookiefile') is None:
+ 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()
},
'INNERTUBE_CONTEXT_CLIENT_NAME': 1
},
+ 'WEB_AGEGATE': {
+ 'INNERTUBE_API_VERSION': 'v1',
+ 'INNERTUBE_CLIENT_NAME': 'WEB',
+ 'INNERTUBE_CLIENT_VERSION': '2.20210622.10.00',
+ 'INNERTUBE_API_KEY': 'AIzaSyAO_FJ2SlqU8Q4STEHLGCilw_Y9_11qcW8',
+ 'INNERTUBE_CONTEXT': {
+ 'client': {
+ 'clientName': 'WEB',
+ 'clientVersion': '2.20210622.10.00',
+ 'clientScreen': 'EMBED',
+ 'hl': 'en',
+ }
+ },
+ 'INNERTUBE_CONTEXT_CLIENT_NAME': 1
+ },
'WEB_REMIX': {
'INNERTUBE_API_VERSION': 'v1',
'INNERTUBE_CLIENT_NAME': 'WEB_REMIX',
},
'INNERTUBE_CONTEXT_CLIENT_NAME': 3
},
+ 'ANDROID_AGEGATE': {
+ 'INNERTUBE_API_VERSION': 'v1',
+ 'INNERTUBE_CLIENT_NAME': 'ANDROID',
+ 'INNERTUBE_CLIENT_VERSION': '16.20',
+ 'INNERTUBE_API_KEY': 'AIzaSyAO_FJ2SlqU8Q4STEHLGCilw_Y9_11qcW8',
+ 'INNERTUBE_CONTEXT': {
+ 'client': {
+ 'clientName': 'ANDROID',
+ 'clientVersion': '16.20',
+ 'clientScreen': 'EMBED',
+ 'hl': 'en',
+ }
+ },
+ 'INNERTUBE_CONTEXT_CLIENT_NAME': 3
+ },
'ANDROID_EMBEDDED_PLAYER': {
'INNERTUBE_API_VERSION': 'v1',
'INNERTUBE_CLIENT_NAME': 'ANDROID_EMBEDDED_PLAYER',
}
},
'INNERTUBE_CONTEXT_CLIENT_NAME': 21
- }
+ },
+ 'IOS': {
+ 'INNERTUBE_API_VERSION': 'v1',
+ 'INNERTUBE_CLIENT_NAME': 'IOS',
+ 'INNERTUBE_CLIENT_VERSION': '16.20',
+ 'INNERTUBE_API_KEY': 'AIzaSyAO_FJ2SlqU8Q4STEHLGCilw_Y9_11qcW8',
+ 'INNERTUBE_CONTEXT': {
+ 'client': {
+ 'clientName': 'IOS',
+ 'clientVersion': '16.20',
+ 'hl': 'en',
+ }
+ },
+ 'INNERTUBE_CONTEXT_CLIENT_NAME': 5
+ },
+ 'IOS_AGEGATE': {
+ 'INNERTUBE_API_VERSION': 'v1',
+ 'INNERTUBE_CLIENT_NAME': 'IOS',
+ 'INNERTUBE_CLIENT_VERSION': '16.20',
+ 'INNERTUBE_API_KEY': 'AIzaSyAO_FJ2SlqU8Q4STEHLGCilw_Y9_11qcW8',
+ 'INNERTUBE_CONTEXT': {
+ 'client': {
+ 'clientName': 'IOS',
+ 'clientVersion': '16.20',
+ 'clientScreen': 'EMBED',
+ 'hl': 'en',
+ }
+ },
+ 'INNERTUBE_CONTEXT_CLIENT_NAME': 5
+ },
+ 'IOS_MUSIC': {
+ 'INNERTUBE_API_VERSION': 'v1',
+ 'INNERTUBE_CLIENT_NAME': 'IOS_MUSIC',
+ 'INNERTUBE_CLIENT_VERSION': '4.32',
+ 'INNERTUBE_API_KEY': 'AIzaSyDK3iBpDP9nHVTk2qL73FLJICfOC3c51Og',
+ 'INNERTUBE_CONTEXT': {
+ 'client': {
+ 'clientName': 'IOS_MUSIC',
+ 'clientVersion': '4.32',
+ 'hl': 'en',
+ }
+ },
+ 'INNERTUBE_CONTEXT_CLIENT_NAME': 26
+ },
+ 'IOS_MESSAGES_EXTENSION': {
+ 'INNERTUBE_API_VERSION': 'v1',
+ 'INNERTUBE_CLIENT_NAME': 'IOS_MESSAGES_EXTENSION',
+ 'INNERTUBE_CLIENT_VERSION': '16.20',
+ 'INNERTUBE_API_KEY': 'AIzaSyDCU8hByM-4DrUqRUYnGn-3llEO78bcxq8',
+ 'INNERTUBE_CONTEXT': {
+ 'client': {
+ 'clientName': 'IOS_MESSAGES_EXTENSION',
+ 'clientVersion': '16.20',
+ 'hl': 'en',
+ }
+ },
+ 'INNERTUBE_CONTEXT_CLIENT_NAME': 66
+ },
+ 'MWEB': {
+ 'INNERTUBE_API_VERSION': 'v1',
+ 'INNERTUBE_CLIENT_NAME': 'MWEB',
+ 'INNERTUBE_CLIENT_VERSION': '2.20210721.07.00',
+ 'INNERTUBE_API_KEY': 'AIzaSyDCU8hByM-4DrUqRUYnGn-3llEO78bcxq8',
+ 'INNERTUBE_CONTEXT': {
+ 'client': {
+ 'clientName': 'MWEB',
+ 'clientVersion': '2.20210721.07.00',
+ 'hl': 'en',
+ }
+ },
+ 'INNERTUBE_CONTEXT_CLIENT_NAME': 2
+ },
+ 'MWEB_AGEGATE': {
+ 'INNERTUBE_API_VERSION': 'v1',
+ 'INNERTUBE_CLIENT_NAME': 'MWEB',
+ 'INNERTUBE_CLIENT_VERSION': '2.20210721.07.00',
+ 'INNERTUBE_API_KEY': 'AIzaSyDCU8hByM-4DrUqRUYnGn-3llEO78bcxq8',
+ 'INNERTUBE_CONTEXT': {
+ 'client': {
+ 'clientName': 'MWEB',
+ 'clientVersion': '2.20210721.07.00',
+ 'clientScreen': 'EMBED',
+ 'hl': 'en',
+ }
+ },
+ 'INNERTUBE_CONTEXT_CLIENT_NAME': 2
+ },
}
_YT_DEFAULT_INNERTUBE_HOSTS = {
'ANDROID_MUSIC': 'music.youtube.com'
}
+ # clients starting with _ cannot be explicity requested by the user
+ _YT_CLIENTS = {
+ 'android': 'ANDROID',
+ 'android_music': 'ANDROID_MUSIC',
+ 'android_embedded': 'ANDROID_EMBEDDED_PLAYER',
+ 'android_agegate': 'ANDROID_AGEGATE',
+ 'ios': 'IOS',
+ 'ios_music': 'IOS_MUSIC',
+ 'ios_embedded': 'IOS_MESSAGES_EXTENSION',
+ 'ios_agegate': 'IOS_AGEGATE',
+ 'web': 'WEB',
+ 'web_music': 'WEB_REMIX',
+ 'web_embedded': 'WEB_EMBEDDED_PLAYER',
+ 'web_agegate': 'WEB_AGEGATE',
+ 'mweb': 'MWEB',
+ 'mweb_agegate': 'MWEB_AGEGATE',
+ }
+
def _get_default_ytcfg(self, client='WEB'):
if client in self._YT_DEFAULT_YTCFGS:
return copy.deepcopy(self._YT_DEFAULT_YTCFGS[client])
return self._ytcfg_get_safe(ytcfg, lambda x: x['INNERTUBE_CLIENT_NAME'], compat_str, default_client)
@staticmethod
- def _extract_session_index(ytcfg):
- return int_or_none(try_get(ytcfg, lambda x: x['SESSION_INDEX']))
+ def _extract_session_index(*data):
+ for ytcfg in data:
+ session_index = int_or_none(try_get(ytcfg, lambda x: x['SESSION_INDEX']))
+ if session_index is not None:
+ return session_index
def _extract_client_version(self, ytcfg, default_client='WEB'):
return self._ytcfg_get_safe(ytcfg, lambda x: x['INNERTUBE_CLIENT_VERSION'], compat_str, default_client)
yt_cookies = self._get_cookies('https://www.youtube.com')
sapisid_cookie = dict_get(
yt_cookies, ('__Secure-3PAPISID', 'SAPISID'))
- if sapisid_cookie is None:
+ if sapisid_cookie is None or not sapisid_cookie.value:
return
time_now = round(time.time())
# SAPISID cookie is required if not already present
if not yt_cookies.get('SAPISID'):
+ self.write_debug('Copying __Secure-3PAPISID cookie to SAPISID cookie', only_once=True)
self._set_cookie(
'.youtube.com', 'SAPISID', sapisid_cookie.value, secure=True, expire_time=time_now + 3600)
+ self.write_debug('Extracted SAPISID cookie', only_once=True)
# SAPISIDHASH algorithm from https://stackoverflow.com/a/32065323
sapisidhash = hashlib.sha1(
f'{time_now} {sapisid_cookie.value} {origin}'.encode('utf-8')).hexdigest()
data = {'context': context} if context else {'context': self._extract_context(default_client=default_client)}
data.update(query)
- real_headers = self._generate_api_headers(client=default_client)
+ real_headers = self.generate_api_headers(default_client=default_client)
real_headers.update({'content-type': 'application/json'})
if headers:
real_headers.update(headers)
data=json.dumps(data).encode('utf8'), headers=real_headers,
query={'key': api_key or self._extract_api_key()})
- def _extract_yt_initial_data(self, video_id, webpage):
+ def extract_yt_initial_data(self, video_id, webpage):
return self._parse_json(
self._search_regex(
(r'%s\s*%s' % (self._YT_INITIAL_DATA_RE, self._YT_INITIAL_BOUNDARY_RE),
video_id)
def _extract_identity_token(self, webpage, item_id):
- ytcfg = self._extract_ytcfg(item_id, webpage)
+ if not webpage:
+ return None
+ ytcfg = self.extract_ytcfg(item_id, webpage)
if ytcfg:
token = try_get(ytcfg, lambda x: x['ID_TOKEN'], compat_str)
if token:
# and just "user_syncid||" for primary channel. We only want the channel_syncid
return sync_ids[0]
- def _extract_ytcfg(self, video_id, webpage):
+ def extract_ytcfg(self, video_id, webpage):
if not webpage:
return {}
return self._parse_json(
r'ytcfg\.set\s*\(\s*({.+?})\s*\)\s*;', webpage, 'ytcfg',
default='{}'), video_id, fatal=False) or {}
- def _generate_api_headers(self, ytcfg=None, identity_token=None, account_syncid=None,
- visitor_data=None, api_hostname=None, client='WEB', session_index=None):
- origin = 'https://' + (api_hostname if api_hostname else self._get_innertube_host(client))
+ def generate_api_headers(
+ self, ytcfg=None, identity_token=None, account_syncid=None,
+ visitor_data=None, api_hostname=None, default_client='WEB', session_index=None):
+ origin = 'https://' + (api_hostname if api_hostname else self._get_innertube_host(default_client))
headers = {
'X-YouTube-Client-Name': compat_str(
- self._ytcfg_get_safe(ytcfg, lambda x: x['INNERTUBE_CONTEXT_CLIENT_NAME'], default_client=client)),
- 'X-YouTube-Client-Version': self._extract_client_version(ytcfg, client),
+ self._ytcfg_get_safe(ytcfg, lambda x: x['INNERTUBE_CONTEXT_CLIENT_NAME'], default_client=default_client)),
+ 'X-YouTube-Client-Version': self._extract_client_version(ytcfg, default_client),
'Origin': origin
}
if not visitor_data and ytcfg:
visitor_data = try_get(
- self._extract_context(ytcfg, client), lambda x: x['client']['visitorData'], compat_str)
+ self._extract_context(ytcfg, default_client), lambda x: x['client']['visitorData'], compat_str)
if identity_token:
headers['X-Youtube-Identity-Token'] = identity_token
if account_syncid:
alert_type = alert.get('type')
if not alert_type:
continue
- message = cls._get_text(alert.get('text'))
+ message = cls._get_text(alert, 'text')
if message:
yield alert_type, message
return badges
@staticmethod
- def _get_text(data, getter=None, max_runs=None):
- for get in variadic(getter):
- d = try_get(data, get) if get is not None else data
- text = try_get(d, lambda x: x['simpleText'], compat_str)
- if text:
- return text
- runs = try_get(d, lambda x: x['runs'], list) or []
- if not runs and isinstance(d, list):
- runs = d
-
- def get_runs(runs):
- for run in runs[:min(len(runs), max_runs or len(runs))]:
- yield try_get(run, lambda x: x['text'], compat_str) or ''
-
- text = ''.join(get_runs(runs))
- if text:
- return text
+ def _get_text(data, *path_list, max_runs=None):
+ for path in path_list or [None]:
+ if path is None:
+ obj = [data]
+ else:
+ obj = traverse_obj(data, path, default=[])
+ if not any(key is ... or isinstance(key, (list, tuple)) for key in variadic(path)):
+ obj = [obj]
+ for item in obj:
+ text = try_get(item, lambda x: x['simpleText'], compat_str)
+ if text:
+ return text
+ runs = try_get(item, lambda x: x['runs'], list) or []
+ if not runs and isinstance(item, list):
+ runs = item
+
+ runs = runs[:min(len(runs), max_runs or len(runs))]
+ text = ''.join(traverse_obj(runs, (..., 'text'), expected_type=str, default=[]))
+ if text:
+ return text
def _extract_response(self, item_id, query, note='Downloading API JSON', headers=None,
ytcfg=None, check_get_keys=None, ep='browse', fatal=True, api_hostname=None,
api_hostname=api_hostname, default_client=default_client,
note='%s%s' % (note, ' (retry #%d)' % count if count else ''))
except ExtractorError as e:
- if isinstance(e.cause, compat_HTTPError) and e.cause.code in (500, 503, 404):
+ if isinstance(e.cause, network_exceptions):
# Downloading page may result in intermittent 5xx HTTP error
# Sometimes a 404 is also recieved. See: https://github.com/ytdl-org/youtube-dl/issues/28289
- last_error = 'HTTP Error %s' % e.cause.code
- if count < retries:
- continue
+ # We also want to catch all other network exceptions since errors in later pages can be troublesome
+ # See https://github.com/yt-dlp/yt-dlp/issues/507#issuecomment-880188210
+ if not isinstance(e.cause, compat_HTTPError) or e.cause.code not in (403, 429):
+ last_error = error_to_compat_str(e.cause or e)
+ if count < retries:
+ continue
if fatal:
raise
else:
def _extract_video(self, renderer):
video_id = renderer.get('videoId')
- title = self._get_text(renderer.get('title'))
- description = self._get_text(renderer.get('descriptionSnippet'))
- duration = parse_duration(self._get_text(renderer.get('lengthText')))
- view_count_text = self._get_text(renderer.get('viewCountText')) or ''
+ title = self._get_text(renderer, 'title')
+ description = self._get_text(renderer, 'descriptionSnippet')
+ duration = parse_duration(self._get_text(
+ renderer, 'lengthText', ('thumbnailOverlays', ..., 'thumbnailOverlayTimeStatusRenderer', 'text')))
+ view_count_text = self._get_text(renderer, 'viewCountText') or ''
view_count = str_to_int(self._search_regex(
r'^([\d,]+)', re.sub(r'\s', '', view_count_text),
'view count', default=None))
- uploader = self._get_text(renderer, (lambda x: x['ownerText'], lambda x: x['shortBylineText']))
+ uploader = self._get_text(renderer, 'ownerText', 'shortBylineText')
return {
'_type': 'url',
'format': '141/bestaudio[ext=m4a]',
},
},
- # Controversy video
- {
- 'url': 'https://www.youtube.com/watch?v=T4XJQO3qol8',
- 'info_dict': {
- 'id': 'T4XJQO3qol8',
- 'ext': 'mp4',
- 'duration': 219,
- 'upload_date': '20100909',
- 'uploader': 'Amazing Atheist',
- 'uploader_id': 'TheAmazingAtheist',
- 'uploader_url': r're:https?://(?:www\.)?youtube\.com/user/TheAmazingAtheist',
- 'title': 'Burning Everyone\'s Koran',
- 'description': 'SUBSCRIBE: http://www.youtube.com/saturninefilms \r\n\r\nEven Obama has taken a stand against freedom on this issue: http://www.huffingtonpost.com/2010/09/09/obama-gma-interview-quran_n_710282.html',
- }
- },
# Normal age-gate video (embed allowed)
{
'url': 'https://youtube.com/watch?v=HtVdAasjOgU',
'uploader_id': 'olympic',
'uploader_url': r're:https?://(?:www\.)?youtube\.com/user/olympic',
'description': 'HO09 - Women - GER-AUS - Hockey - 31 July 2012 - London 2012 Olympic Games',
- 'uploader': 'Olympic',
+ 'uploader': 'Olympics',
'title': 'Hockey - Women - GER-AUS - London 2012 Olympic Games',
},
'params': {
'id': 'lsguqyKfVQg',
'ext': 'mp4',
'title': '{dark walk}; Loki/AC/Dishonored; collab w/Elflover21',
- 'alt_title': 'Dark Walk - Position Music',
+ 'alt_title': 'Dark Walk',
'description': 'md5:8085699c11dc3f597ce0410b0dcbb34a',
'duration': 133,
'upload_date': '20151119',
'uploader_id': 'IronSoulElf',
'uploader_url': r're:https?://(?:www\.)?youtube\.com/user/IronSoulElf',
'uploader': 'IronSoulElf',
- 'creator': 'Todd Haberman, Daniel Law Heath and Aaron Kaplan',
- 'track': 'Dark Walk - Position Music',
- 'artist': 'Todd Haberman, Daniel Law Heath and Aaron Kaplan',
+ 'creator': 'Todd Haberman;\nDaniel Law Heath and Aaron Kaplan',
+ 'track': 'Dark Walk',
+ 'artist': 'Todd Haberman;\nDaniel Law Heath and Aaron Kaplan',
'album': 'Position Music - Production Music Vol. 143 - Dark Walk',
},
'params': {
'url': 'https://www.youtube.com/watch?v=nGC3D_FkCmg',
'only_matching': True,
},
+ {
+ # controversial video, requires bpctr/contentCheckOk
+ 'url': 'https://www.youtube.com/watch?v=SZJvDhaSDnc',
+ 'info_dict': {
+ 'id': 'SZJvDhaSDnc',
+ 'ext': 'mp4',
+ 'title': 'San Diego teen commits suicide after bullying over embarrassing video',
+ 'channel_id': 'UC-SJ6nODDmufqBzPBwCvYvQ',
+ 'uploader': 'CBS This Morning',
+ 'uploader_id': 'CBSThisMorning',
+ 'upload_date': '20140716',
+ 'description': 'md5:acde3a73d3f133fc97e837a9f76b53b7'
+ }
+ },
{
# restricted location, https://github.com/ytdl-org/youtube-dl/issues/28685
'url': 'cBvYw8_A0vQ',
'url': 'https://www.youtube.com/watch?v=YOelRv7fMxY',
'info_dict': {
'id': 'YOelRv7fMxY',
- 'title': 'Digging a Secret Tunnel from my Workshop',
+ 'title': 'DIGGING A SECRET TUNNEL Part 1',
'ext': '3gp',
'upload_date': '20210624',
'channel_id': 'UCp68_FLety0O-n9QU6phsgw',
'uploader': 'colinfurze',
+ 'uploader_id': 'colinfurze',
'channel_url': r're:https?://(?:www\.)?youtube\.com/channel/UCp68_FLety0O-n9QU6phsgw',
- 'description': 'md5:ecb672623246d98c6c562eed6ae798c3'
+ 'description': 'md5:b5096f56af7ccd7a555c84db81738b22'
},
'params': {
'format': '17', # 3gp format available on android
def _extract_player_url(self, ytcfg=None, webpage=None):
player_url = try_get(ytcfg, (lambda x: x['PLAYER_JS_URL']), str)
- if not player_url:
+ if not player_url and webpage:
player_url = self._search_regex(
r'"(?:PLAYER_JS_URL|jsUrl)"\s*:\s*"([^"]+)"',
webpage, 'player URL', fatal=False)
+ if not player_url:
+ return None
if player_url.startswith('//'):
player_url = 'https:' + player_url
elif not re.match(r'https?://', player_url):
'JS player signature timestamp', group='sts', fatal=fatal))
return sts
- def _mark_watched(self, video_id, player_response):
- playback_url = url_or_none(try_get(
- player_response,
- lambda x: x['playbackTracking']['videostatsPlaybackUrl']['baseUrl']))
+ def _mark_watched(self, video_id, player_responses):
+ playback_url = traverse_obj(
+ player_responses, (..., 'playbackTracking', 'videostatsPlaybackUrl', 'baseUrl'),
+ expected_type=url_or_none, get_all=False)
if not playback_url:
+ self.report_warning('Unable to mark watched')
return
parsed_playback_url = compat_urlparse.urlparse(playback_url)
qs = compat_urlparse.parse_qs(parsed_playback_url.query)
video_id = mobj.group(2)
return video_id
- def _extract_chapters_from_json(self, data, video_id, duration):
- chapters_list = try_get(
+ def _extract_chapters_from_json(self, data, duration):
+ chapter_list = traverse_obj(
+ data, (
+ 'playerOverlays', 'playerOverlayRenderer', 'decoratedPlayerBarRenderer',
+ 'decoratedPlayerBarRenderer', 'playerBar', 'chapteredPlayerBarRenderer', 'chapters'
+ ), expected_type=list)
+
+ return self._extract_chapters(
+ chapter_list,
+ chapter_time=lambda chapter: float_or_none(
+ traverse_obj(chapter, ('chapterRenderer', 'timeRangeStartMillis')), scale=1000),
+ chapter_title=lambda chapter: traverse_obj(
+ chapter, ('chapterRenderer', 'title', 'simpleText'), expected_type=str),
+ duration=duration)
+
+ def _extract_chapters_from_engagement_panel(self, data, duration):
+ content_list = traverse_obj(
data,
- lambda x: x['playerOverlays']
- ['playerOverlayRenderer']
- ['decoratedPlayerBarRenderer']
- ['decoratedPlayerBarRenderer']
- ['playerBar']
- ['chapteredPlayerBarRenderer']
- ['chapters'],
- list)
- if not chapters_list:
- return
-
- def chapter_time(chapter):
- return float_or_none(
- try_get(
- chapter,
- lambda x: x['chapterRenderer']['timeRangeStartMillis'],
- int),
- scale=1000)
+ ('engagementPanels', ..., 'engagementPanelSectionListRenderer', 'content', 'macroMarkersListRenderer', 'contents'),
+ expected_type=list, default=[])
+ 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 = []
- for next_num, chapter in enumerate(chapters_list, start=1):
+ 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
- end_time = (chapter_time(chapters_list[next_num])
- if next_num < len(chapters_list) else duration)
- if end_time is None:
- continue
- title = try_get(
- chapter, lambda x: x['chapterRenderer']['title']['simpleText'],
- compat_str)
- chapters.append({
- 'start_time': start_time,
- 'end_time': end_time,
- 'title': title,
- })
+ 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
def _extract_yt_initial_variable(self, webpage, regex, video_id, name):
"""
time_text_split = time_text.split(' ')
if len(time_text_split) >= 3:
- return datetime_from_str('now-%s%s' % (time_text_split[0], time_text_split[1]), precision='auto')
+ try:
+ return datetime_from_str('now-%s%s' % (time_text_split[0], time_text_split[1]), precision='auto')
+ except ValueError:
+ return None
def _extract_comment(self, comment_renderer, parent=None):
comment_id = comment_renderer.get('commentId')
if not comment_id:
return
- text = self._get_text(comment_renderer.get('contentText'))
+ text = self._get_text(comment_renderer, 'contentText')
# note: timestamp is an estimate calculated from the current time and time_text
- time_text = self._get_text(comment_renderer.get('publishedTimeText')) or ''
+ time_text = self._get_text(comment_renderer, 'publishedTimeText') or ''
time_text_dt = self.parse_time_text(time_text)
if isinstance(time_text_dt, datetime.datetime):
timestamp = calendar.timegm(time_text_dt.timetuple())
- author = self._get_text(comment_renderer.get('authorText'))
+ author = self._get_text(comment_renderer, 'authorText')
author_id = try_get(comment_renderer,
lambda x: x['authorEndpoint']['browseEndpoint']['browseId'], compat_str)
for content in contents:
comments_header_renderer = try_get(content, lambda x: x['commentsHeaderRenderer'])
expected_comment_count = parse_count(self._get_text(
- comments_header_renderer, (lambda x: x['countText'], lambda x: x['commentsCount']), max_runs=1))
+ comments_header_renderer, 'countText', 'commentsCount', max_runs=1))
if expected_comment_count:
comment_counts[1] = expected_comment_count
for page_num in itertools.count(0):
if not continuation:
break
- headers = self._generate_api_headers(ytcfg, identity_token, account_syncid, visitor_data)
+ headers = self.generate_api_headers(ytcfg, identity_token, account_syncid, visitor_data)
comment_prog_str = '(%d/%d)' % (comment_counts[0], comment_counts[1])
if page_num == 0:
if is_first_continuation:
known_entry_comment_renderers = ('itemSectionRenderer',)
estimated_total = 0
max_comments = int_or_none(self._configuration_arg('max_comments', [''])[0]) or float('inf')
-
+ # Force English regardless of account setting to prevent parsing issues
+ # See: https://github.com/yt-dlp/yt-dlp/issues/532
+ ytcfg = copy.deepcopy(ytcfg)
+ traverse_obj(
+ ytcfg, ('INNERTUBE_CONTEXT', 'client'), expected_type=dict, default={})['hl'] = 'en'
try:
for comment in _real_comment_extract(contents):
if len(comments) >= max_comments:
return {
'playbackContext': {
'contentPlaybackContext': context
- }
- }
-
- @staticmethod
- def _get_video_info_params(video_id, client='TVHTML5'):
- GVI_CLIENTS = {
- 'ANDROID': {
- 'c': 'ANDROID',
- 'cver': '16.20',
},
- 'TVHTML5': {
- 'c': 'TVHTML5',
- 'cver': '6.20180913',
- }
- }
- query = {
- 'video_id': video_id,
- 'eurl': 'https://youtube.googleapis.com/v/' + video_id,
- 'html5': '1'
+ 'contentCheckOk': True,
+ 'racyCheckOk': True
}
- query.update(GVI_CLIENTS.get(client))
- return query
- def _real_extract(self, url):
- url, smuggled_data = unsmuggle_url(url, {})
- video_id = self._match_id(url)
+ def _extract_player_response(self, client, video_id, master_ytcfg, player_ytcfg, identity_token, player_url, initial_pr):
- is_music_url = smuggled_data.get('is_music_url') or self.is_music_url(url)
+ session_index = self._extract_session_index(player_ytcfg, master_ytcfg)
+ syncid = self._extract_account_syncid(player_ytcfg, master_ytcfg, initial_pr)
+ sts = self._extract_signature_timestamp(video_id, player_url, master_ytcfg, fatal=False)
+ headers = self.generate_api_headers(
+ player_ytcfg, identity_token, syncid,
+ default_client=self._YT_CLIENTS[client], session_index=session_index)
- base_url = self.http_scheme() + '//www.youtube.com/'
- webpage_url = base_url + 'watch?v=' + video_id
- webpage = self._download_webpage(
- webpage_url + '&bpctr=9999999999&has_verified=1', video_id, fatal=False)
+ yt_query = {'videoId': video_id}
+ yt_query.update(self._generate_player_context(sts))
+ return self._extract_response(
+ item_id=video_id, ep='player', query=yt_query,
+ ytcfg=player_ytcfg, headers=headers, fatal=False,
+ default_client=self._YT_CLIENTS[client],
+ note='Downloading %s player API JSON' % client.replace('_', ' ').strip()
+ ) or None
+
+ def _get_requested_clients(self, url, smuggled_data):
+ requested_clients = []
+ allowed_clients = [client for client in self._YT_CLIENTS.keys() if client[:1] != '_']
+ for client in self._configuration_arg('player_client'):
+ if client in allowed_clients:
+ requested_clients.append(client)
+ elif client == 'all':
+ requested_clients.extend(allowed_clients)
+ else:
+ self.report_warning(f'Skipping unsupported client {client}')
+ if not requested_clients:
+ requested_clients = ['android', 'web']
+
+ if smuggled_data.get('is_music_url') or self.is_music_url(url):
+ requested_clients.extend(
+ f'{client}_music' for client in requested_clients if not client.endswith('_music'))
+
+ 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=f'Downloading {client} config')
+ return self.extract_ytcfg(video_id, webpage) or {}
- ytcfg = self._extract_ytcfg(video_id, webpage) or self._get_default_ytcfg()
- identity_token = self._extract_identity_token(webpage, video_id)
- session_index = self._extract_session_index(ytcfg)
- player_url = self._extract_player_url(ytcfg, webpage)
-
- player_client = self._configuration_arg('player_client', [''])[0]
- if player_client not in ('web', 'android', ''):
- self.report_warning(f'Invalid player_client {player_client} given. Falling back to android client.')
- force_mobile_client = player_client != 'web'
- player_skip = self._configuration_arg('player_skip')
- player_response = None
+ def _extract_player_responses(self, clients, video_id, webpage, master_ytcfg, player_url, identity_token):
+ initial_pr = None
if webpage:
- player_response = self._extract_yt_initial_variable(
+ initial_pr = self._extract_yt_initial_variable(
webpage, self._YT_INITIAL_PLAYER_RESPONSE_RE,
video_id, 'initial player response')
- syncid = self._extract_account_syncid(ytcfg, player_response)
- headers = self._generate_api_headers(ytcfg, identity_token, syncid, session_index=session_index)
-
- ytm_streaming_data = {}
- if is_music_url:
- ytm_webpage = None
- sts = self._extract_signature_timestamp(video_id, player_url, ytcfg, fatal=False)
- if sts and not force_mobile_client and 'configs' not in player_skip:
- ytm_webpage = self._download_webpage(
- 'https://music.youtube.com',
- video_id, fatal=False, note='Downloading remix client config')
-
- ytm_cfg = self._extract_ytcfg(video_id, ytm_webpage) or {}
- ytm_client = 'WEB_REMIX'
- if not sts or force_mobile_client:
- # Android client already has signature descrambled
- # See: https://github.com/TeamNewPipe/NewPipeExtractor/issues/562
- if not sts:
- self.report_warning('Falling back to android remix client for player API.')
- ytm_client = 'ANDROID_MUSIC'
- ytm_cfg = {}
-
- ytm_headers = self._generate_api_headers(
- ytm_cfg, identity_token, syncid,
- client=ytm_client, session_index=session_index)
- ytm_query = {'videoId': video_id}
- ytm_query.update(self._generate_player_context(sts))
-
- ytm_player_response = self._extract_response(
- item_id=video_id, ep='player', query=ytm_query,
- ytcfg=ytm_cfg, headers=ytm_headers, fatal=False,
- default_client=ytm_client,
- note='Downloading %sremix player API JSON' % ('android ' if force_mobile_client else ''))
- ytm_streaming_data = try_get(ytm_player_response, lambda x: x['streamingData'], dict) or {}
-
- if not player_response or force_mobile_client:
- sts = self._extract_signature_timestamp(video_id, player_url, ytcfg, fatal=False)
- yt_client = 'WEB'
- ytpcfg = ytcfg
- ytp_headers = headers
- if not sts or force_mobile_client:
- # Android client already has signature descrambled
- # See: https://github.com/TeamNewPipe/NewPipeExtractor/issues/562
- if not sts:
- self.report_warning('Falling back to android client for player API.')
- yt_client = 'ANDROID'
- ytpcfg = {}
- ytp_headers = self._generate_api_headers(ytpcfg, identity_token, syncid,
- client=yt_client, session_index=session_index)
-
- yt_query = {'videoId': video_id}
- yt_query.update(self._generate_player_context(sts))
- player_response = self._extract_response(
- item_id=video_id, ep='player', query=yt_query,
- ytcfg=ytpcfg, headers=ytp_headers, fatal=False,
- default_client=yt_client,
- note='Downloading %splayer API JSON' % ('android ' if force_mobile_client else '')
- ) or player_response
-
- # Age-gate workarounds
- playability_status = player_response.get('playabilityStatus') or {}
- if playability_status.get('reason') in self._AGE_GATE_REASONS:
- gvi_clients = ('ANDROID', 'TVHTML5') if force_mobile_client else ('TVHTML5', 'ANDROID')
- for gvi_client in gvi_clients:
- pr = self._parse_json(try_get(compat_parse_qs(
- self._download_webpage(
- base_url + 'get_video_info', video_id,
- 'Refetching age-gated %s info webpage' % gvi_client.lower(),
- 'unable to download video info webpage', fatal=False,
- query=self._get_video_info_params(video_id, client=gvi_client))),
- lambda x: x['player_response'][0],
- compat_str) or '{}', video_id)
- if pr:
- break
- if not pr:
- self.report_warning('Falling back to embedded-only age-gate workaround.')
- embed_webpage = None
- sts = self._extract_signature_timestamp(video_id, player_url, ytcfg, fatal=False)
- if sts and not force_mobile_client and 'configs' not in player_skip:
- embed_webpage = self._download_webpage(
- 'https://www.youtube.com/embed/%s?html5=1' % video_id,
- video_id=video_id, note='Downloading age-gated embed config')
-
- ytcfg_age = self._extract_ytcfg(video_id, embed_webpage) or {}
- # If we extracted the embed webpage, it'll tell us if we can view the video
- embedded_pr = self._parse_json(
- try_get(ytcfg_age, lambda x: x['PLAYER_VARS']['embedded_player_response'], str) or '{}',
- video_id=video_id)
- embedded_ps_reason = try_get(embedded_pr, lambda x: x['playabilityStatus']['reason'], str) or ''
- if embedded_ps_reason not in self._AGE_GATE_REASONS:
- yt_client = 'WEB_EMBEDDED_PLAYER'
- if not sts or force_mobile_client:
- # Android client already has signature descrambled
- # See: https://github.com/TeamNewPipe/NewPipeExtractor/issues/562
- if not sts:
- self.report_warning(
- 'Falling back to android embedded client for player API (note: some formats may be missing).')
- yt_client = 'ANDROID_EMBEDDED_PLAYER'
- ytcfg_age = {}
-
- ytage_headers = self._generate_api_headers(
- ytcfg_age, identity_token, syncid,
- client=yt_client, session_index=session_index)
- yt_age_query = {'videoId': video_id}
- yt_age_query.update(self._generate_player_context(sts))
- pr = self._extract_response(
- item_id=video_id, ep='player', query=yt_age_query,
- ytcfg=ytcfg_age, headers=ytage_headers, fatal=False,
- default_client=yt_client,
- note='Downloading %sage-gated player API JSON' % ('android ' if force_mobile_client else '')
- ) or {}
+ original_clients = clients
+ clients = clients[::-1]
+ while clients:
+ 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 client == 'web_embedded':
+ # If we extracted the embed webpage, it'll tell us if we can view the video
+ embedded_pr = self._parse_json(
+ traverse_obj(player_ytcfg, ('PLAYER_VARS', 'embedded_player_response'), expected_type=str) or '{}',
+ video_id=video_id)
+ embedded_ps_reason = traverse_obj(embedded_pr, ('playabilityStatus', 'reason'), expected_type=str) or ''
+ if embedded_ps_reason in self._AGE_GATE_REASONS:
+ self.report_warning(f'Youtube said: {embedded_ps_reason}')
+ continue
+ pr = (
+ initial_pr if client == 'web' and initial_pr
+ else self._extract_player_response(
+ client, video_id, player_ytcfg or master_ytcfg, player_ytcfg, identity_token, player_url, initial_pr))
if pr:
- player_response = pr
-
- trailer_video_id = try_get(
- playability_status,
- lambda x: x['errorScreen']['playerLegacyDesktopYpcTrailerRenderer']['trailerVideoId'],
- compat_str)
- if trailer_video_id:
- return self.url_result(
- trailer_video_id, self.ie_key(), trailer_video_id)
-
- search_meta = (
- lambda x: self._html_search_meta(x, webpage, default=None)) \
- if webpage else lambda x: None
-
- video_details = player_response.get('videoDetails') or {}
- microformat = try_get(
- player_response,
- lambda x: x['microformat']['playerMicroformatRenderer'],
- dict) or {}
- video_title = video_details.get('title') \
- or self._get_text(microformat.get('title')) \
- or search_meta(['og:title', 'twitter:title', 'title'])
- video_description = video_details.get('shortDescription')
-
- if not smuggled_data.get('force_singlefeed', False):
- if not self.get_param('noplaylist'):
- multifeed_metadata_list = try_get(
- player_response,
- lambda x: x['multicamera']['playerLegacyMulticameraRenderer']['metadataList'],
- compat_str)
- if multifeed_metadata_list:
- entries = []
- feed_ids = []
- for feed in multifeed_metadata_list.split(','):
- # Unquote should take place before split on comma (,) since textual
- # fields may contain comma as well (see
- # https://github.com/ytdl-org/youtube-dl/issues/8536)
- feed_data = compat_parse_qs(
- compat_urllib_parse_unquote_plus(feed))
-
- def feed_entry(name):
- return try_get(
- feed_data, lambda x: x[name][0], compat_str)
-
- feed_id = feed_entry('id')
- if not feed_id:
- continue
- feed_title = feed_entry('title')
- title = video_title
- if feed_title:
- title += ' (%s)' % feed_title
- entries.append({
- '_type': 'url_transparent',
- 'ie_key': 'Youtube',
- 'url': smuggle_url(
- base_url + 'watch?v=' + feed_data['id'][0],
- {'force_singlefeed': True}),
- 'title': title,
- })
- feed_ids.append(feed_id)
- self.to_screen(
- 'Downloading multifeed video (%s) - add --no-playlist to just download video %s'
- % (', '.join(feed_ids), video_id))
- return self.playlist_result(
- entries, video_id, video_title, video_description)
- else:
- self.to_screen('Downloading just video %s because of --no-playlist' % video_id)
-
- formats, itags, stream_ids = [], [], []
- itag_qualities = {}
+ yield pr
+
+ if traverse_obj(pr, ('playabilityStatus', 'reason')) in self._AGE_GATE_REASONS:
+ client = f'{client}_agegate'
+ if client in self._YT_CLIENTS and client not in original_clients:
+ clients.append(client)
+
+ # Android player_response does not have microFormats which are needed for
+ # extraction of some data. So we return the initial_pr with formats
+ # stripped out even if not requested by the user
+ # See: https://github.com/yt-dlp/yt-dlp/issues/501
+ if initial_pr and 'web' not in original_clients:
+ initial_pr['streamingData'] = None
+ yield initial_pr
+
+ def _extract_formats(self, streaming_data, video_id, player_url, is_live):
+ itags, stream_ids = [], []
+ itag_qualities, res_qualities = {}, {}
q = qualities([
- # "tiny" is the smallest video-only format. But some audio-only formats
- # was also labeled "tiny". It is not clear if such formats still exist
- 'tiny', 'audio_quality_low', 'audio_quality_medium', 'audio_quality_high', # Audio only formats
+ # Normally tiny is the smallest video-only formats. But
+ # audio-only formats with unknown quality may get tagged as tiny
+ 'tiny',
+ 'audio_quality_ultralow', 'audio_quality_low', 'audio_quality_medium', 'audio_quality_high', # Audio only formats
'small', 'medium', 'large', 'hd720', 'hd1080', 'hd1440', 'hd2160', 'hd2880', 'highres'
])
-
- streaming_data = player_response.get('streamingData') or {}
- streaming_formats = streaming_data.get('formats') or []
- streaming_formats.extend(streaming_data.get('adaptiveFormats') or [])
- streaming_formats.extend(ytm_streaming_data.get('formats') or [])
- streaming_formats.extend(ytm_streaming_data.get('adaptiveFormats') or [])
+ streaming_formats = traverse_obj(streaming_data, (..., ('formats', 'adaptiveFormats'), ...), default=[])
for fmt in streaming_formats:
if fmt.get('targetDurationSec') or fmt.get('drmFamilies'):
continue
quality = fmt.get('quality')
+ height = int_or_none(fmt.get('height'))
if quality == 'tiny' or not quality:
quality = fmt.get('audioQuality', '').lower() or quality
- if itag and quality:
- itag_qualities[itag] = quality
+ # The 3gp format (17) in android client has a quality of "small",
+ # but is actually worse than other formats
+ if itag == '17':
+ quality = 'tiny'
+ if quality:
+ if itag:
+ itag_qualities[itag] = quality
+ if height:
+ res_qualities[height] = quality
# FORMAT_STREAM_TYPE_OTF(otf=1) requires downloading the init fragment
# (adding `&sq=0` to the URL) and parsing emsg box to determine the
# number of fragment that would subsequently requested with (`&sq=N`)
'asr': int_or_none(fmt.get('audioSampleRate')),
'filesize': int_or_none(fmt.get('contentLength')),
'format_id': itag,
- 'format_note': audio_track.get('displayName') or fmt.get('qualityLabel') or quality,
+ 'format_note': ', '.join(filter(None, (
+ audio_track.get('displayName'),
+ fmt.get('qualityLabel') or quality.replace('audio_quality_', '')))),
'fps': int_or_none(fmt.get('fps')),
- 'height': int_or_none(fmt.get('height')),
+ 'height': height,
'quality': q(quality),
'tbr': tbr,
'url': fmt_url,
- 'width': fmt.get('width'),
+ 'width': int_or_none(fmt.get('width')),
'language': audio_track.get('id', '').split('.')[0],
}
mime_mobj = re.match(
if mime_mobj:
dct['ext'] = mimetype2ext(mime_mobj.group(1))
dct.update(parse_codecs(mime_mobj.group(2)))
- # The 3gp format in android client has a quality of "small",
- # but is actually worse than all other formats
- if dct['ext'] == '3gp':
- dct['quality'] = q('tiny')
no_audio = dct.get('acodec') == 'none'
no_video = dct.get('vcodec') == 'none'
if no_audio:
}
if dct.get('ext'):
dct['container'] = dct['ext'] + '_dash'
- formats.append(dct)
+ yield dct
skip_manifests = self._configuration_arg('skip')
- get_dash = 'dash' not in skip_manifests and self.get_param('youtube_include_dash_manifest', True)
+ get_dash = not is_live and 'dash' not in skip_manifests and self.get_param('youtube_include_dash_manifest', True)
get_hls = 'hls' not in skip_manifests and self.get_param('youtube_include_hls_manifest', True)
- for sd in (streaming_data, ytm_streaming_data):
+ def guess_quality(f):
+ for val, qdict in ((f.get('format_id'), itag_qualities), (f.get('height'), res_qualities)):
+ if val in qdict:
+ return q(qdict[val])
+ return -1
+
+ for sd in streaming_data:
hls_manifest_url = get_hls and sd.get('hlsManifestUrl')
if hls_manifest_url:
- for f in self._extract_m3u8_formats(
- hls_manifest_url, video_id, 'mp4', fatal=False):
+ for f in self._extract_m3u8_formats(hls_manifest_url, video_id, 'mp4', fatal=False):
itag = self._search_regex(
r'/itag/(\d+)', f['url'], 'itag', default=None)
+ if itag in itags:
+ continue
if itag:
f['format_id'] = itag
- formats.append(f)
+ itags.append(itag)
+ f['quality'] = guess_quality(f)
+ yield f
dash_manifest_url = get_dash and sd.get('dashManifestUrl')
if dash_manifest_url:
- for f in self._extract_mpd_formats(
- dash_manifest_url, video_id, fatal=False):
+ for f in self._extract_mpd_formats(dash_manifest_url, video_id, fatal=False):
itag = f['format_id']
if itag in itags:
continue
- if itag in itag_qualities:
- f['quality'] = q(itag_qualities[itag])
+ if itag:
+ itags.append(itag)
+ f['quality'] = guess_quality(f)
filesize = int_or_none(self._search_regex(
r'/clen/(\d+)', f.get('fragment_base_url')
or f['url'], 'file size', default=None))
if filesize:
f['filesize'] = filesize
- formats.append(f)
+ yield f
+
+ def _real_extract(self, url):
+ url, smuggled_data = unsmuggle_url(url, {})
+ video_id = self._match_id(url)
+
+ base_url = self.http_scheme() + '//www.youtube.com/'
+ webpage_url = base_url + 'watch?v=' + video_id
+ webpage = self._download_webpage(
+ webpage_url + '&bpctr=9999999999&has_verified=1', video_id, fatal=False)
+
+ master_ytcfg = self.extract_ytcfg(video_id, webpage) or self._get_default_ytcfg()
+ player_url = self._extract_player_url(master_ytcfg, webpage)
+ identity_token = self._extract_identity_token(webpage, video_id)
+
+ player_responses = list(self._extract_player_responses(
+ self._get_requested_clients(url, smuggled_data),
+ video_id, webpage, master_ytcfg, player_url, identity_token))
+
+ get_first = lambda obj, keys, **kwargs: traverse_obj(obj, (..., *variadic(keys)), **kwargs, get_all=False)
+
+ playability_statuses = traverse_obj(
+ player_responses, (..., 'playabilityStatus'), expected_type=dict, default=[])
+
+ trailer_video_id = get_first(
+ playability_statuses,
+ ('errorScreen', 'playerLegacyDesktopYpcTrailerRenderer', 'trailerVideoId'),
+ expected_type=str)
+ if trailer_video_id:
+ return self.url_result(
+ trailer_video_id, self.ie_key(), trailer_video_id)
+
+ search_meta = ((lambda x: self._html_search_meta(x, webpage, default=None))
+ if webpage else (lambda x: None))
+
+ video_details = traverse_obj(
+ player_responses, (..., 'videoDetails'), expected_type=dict, default=[])
+ microformats = traverse_obj(
+ player_responses, (..., 'microformat', 'playerMicroformatRenderer'),
+ expected_type=dict, default=[])
+ video_title = (
+ get_first(video_details, 'title')
+ or self._get_text(microformats, (..., 'title'))
+ or search_meta(['og:title', 'twitter:title', 'title']))
+ video_description = get_first(video_details, 'shortDescription')
+
+ if not smuggled_data.get('force_singlefeed', False):
+ if not self.get_param('noplaylist'):
+ multifeed_metadata_list = get_first(
+ player_responses,
+ ('multicamera', 'playerLegacyMulticameraRenderer', 'metadataList'),
+ expected_type=str)
+ if multifeed_metadata_list:
+ entries = []
+ feed_ids = []
+ for feed in multifeed_metadata_list.split(','):
+ # Unquote should take place before split on comma (,) since textual
+ # fields may contain comma as well (see
+ # https://github.com/ytdl-org/youtube-dl/issues/8536)
+ feed_data = compat_parse_qs(
+ compat_urllib_parse_unquote_plus(feed))
+
+ def feed_entry(name):
+ return try_get(
+ feed_data, lambda x: x[name][0], compat_str)
+
+ feed_id = feed_entry('id')
+ if not feed_id:
+ continue
+ feed_title = feed_entry('title')
+ title = video_title
+ if feed_title:
+ title += ' (%s)' % feed_title
+ entries.append({
+ '_type': 'url_transparent',
+ 'ie_key': 'Youtube',
+ 'url': smuggle_url(
+ '%swatch?v=%s' % (base_url, feed_data['id'][0]),
+ {'force_singlefeed': True}),
+ 'title': title,
+ })
+ feed_ids.append(feed_id)
+ self.to_screen(
+ 'Downloading multifeed video (%s) - add --no-playlist to just download video %s'
+ % (', '.join(feed_ids), video_id))
+ return self.playlist_result(
+ entries, video_id, video_title, video_description)
+ else:
+ self.to_screen('Downloading just video %s because of --no-playlist' % video_id)
+
+ 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))
if not formats:
- if not self.get_param('allow_unplayable_formats') and streaming_data.get('licenseInfos'):
+ if not self.get_param('allow_unplayable_formats') and traverse_obj(streaming_data, (..., 'licenseInfos')):
self.raise_no_formats(
'This video is DRM protected.', expected=True)
- pemr = try_get(
- playability_status,
- lambda x: x['errorScreen']['playerErrorMessageRenderer'],
- dict) or {}
- reason = self._get_text(pemr.get('reason')) or playability_status.get('reason')
- subreason = pemr.get('subreason')
+ pemr = get_first(
+ playability_statuses,
+ ('errorScreen', 'playerErrorMessageRenderer'), expected_type=dict) or {}
+ reason = self._get_text(pemr, 'reason') or get_first(playability_statuses, 'reason')
+ subreason = clean_html(self._get_text(pemr, 'subreason') or '')
if subreason:
- subreason = clean_html(self._get_text(subreason))
if subreason == 'The uploader has not made this video available in your country.':
- countries = microformat.get('availableCountries')
+ countries = get_first(microformats, 'availableCountries')
if not countries:
regions_allowed = search_meta('regionsAllowed')
countries = regions_allowed.split(',') if regions_allowed else None
self.raise_geo_restricted(subreason, countries, metadata_available=True)
- reason += '\n' + subreason
+ reason += f'. {subreason}'
if reason:
self.raise_no_formats(reason, expected=True)
- self._sort_formats(formats)
+ for f in formats:
+ if '&c=WEB&' in f['url'] and '&ratebypass=yes&' not in f['url']: # throttled
+ f['source_preference'] = -10
+ note = f.get('format_note')
+ f['format_note'] = f'{note} (throttled)' if note else '(throttled)'
+
+ # Source is given priority since formats that throttle are given lower source_preference
+ # When throttling issue is fully fixed, remove this
+ self._sort_formats(formats, ('quality', 'height', 'fps', 'source'))
- keywords = video_details.get('keywords') or []
+ keywords = get_first(video_details, 'keywords', expected_type=list) or []
if not keywords and webpage:
keywords = [
unescapeHTML(m.group('content'))
f['stretched_ratio'] = ratio
break
- category = microformat.get('category') or search_meta('genre')
- channel_id = video_details.get('channelId') \
- or microformat.get('externalChannelId') \
- or search_meta('channelId')
- duration = int_or_none(
- video_details.get('lengthSeconds')
- or microformat.get('lengthSeconds')) \
- or parse_duration(search_meta('duration'))
- is_live = video_details.get('isLive')
- is_upcoming = video_details.get('isUpcoming')
- owner_profile_url = microformat.get('ownerProfileUrl')
-
thumbnails = []
- thumbnail_types = ['maxresdefault', 'sddefault', 'hqdefault', '0', 'mqdefault', 'default', '1', '2', '3']
-
- for container in (video_details, microformat):
- for thumbnail in (try_get(
- container,
- lambda x: x['thumbnail']['thumbnails'], list) or []):
- thumbnail_url = thumbnail.get('url')
- if not thumbnail_url:
- continue
- # Sometimes youtube gives a wrong thumbnail URL. See:
- # https://github.com/yt-dlp/yt-dlp/issues/233
- # https://github.com/ytdl-org/youtube-dl/issues/28023
- if 'maxresdefault' in thumbnail_url:
- thumbnail_url = thumbnail_url.split('?')[0]
- thumbnails.append({
- 'url': thumbnail_url,
- 'height': int_or_none(thumbnail.get('height')),
- 'width': int_or_none(thumbnail.get('width')),
- })
+ thumbnail_dicts = traverse_obj(
+ (video_details, microformats), (..., ..., 'thumbnail', 'thumbnails', ...),
+ expected_type=dict, default=[])
+ for thumbnail in thumbnail_dicts:
+ thumbnail_url = thumbnail.get('url')
+ if not thumbnail_url:
+ continue
+ # Sometimes youtube gives a wrong thumbnail URL. See:
+ # https://github.com/yt-dlp/yt-dlp/issues/233
+ # https://github.com/ytdl-org/youtube-dl/issues/28023
+ if 'maxresdefault' in thumbnail_url:
+ thumbnail_url = thumbnail_url.split('?')[0]
+ thumbnails.append({
+ 'url': thumbnail_url,
+ 'height': int_or_none(thumbnail.get('height')),
+ 'width': int_or_none(thumbnail.get('width')),
+ })
thumbnail_url = search_meta(['og:image', 'twitter:image'])
if thumbnail_url:
thumbnails.append({
})
# 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
+ # List of possible thumbnails - Ref: <https://stackoverflow.com/a/20542029>
+ hq_thumbnail_names = ['maxresdefault', 'hq720', 'sddefault', 'sd1', 'sd2', 'sd3']
+ # TODO: Test them also? - For some videos, even these don't exist
+ guaranteed_thumbnail_names = [
+ 'hqdefault', 'hq1', 'hq2', 'hq3', '0',
+ 'mqdefault', 'mq1', 'mq2', 'mq3',
+ 'default', '1', '2', '3'
+ ]
+ thumbnail_names = hq_thumbnail_names + guaranteed_thumbnail_names
+ n_thumbnail_names = len(thumbnail_names)
+
thumbnails.extend({
'url': 'https://i.ytimg.com/vi{webp}/{video_id}/{name}{live}.{ext}'.format(
video_id=video_id, name=name, ext=ext,
webp='_webp' if ext == 'webp' else '', live='_live' if is_live else ''),
- '_test_url': True,
- } for name in thumbnail_types for ext in ('webp', 'jpg'))
+ '_test_url': name in hq_thumbnail_names,
+ } for name in thumbnail_names for ext in ('webp', 'jpg'))
for thumb in thumbnails:
- i = next((i for i, t in enumerate(thumbnail_types) if f'/{video_id}/{t}' in thumb['url']), 20)
+ i = next((i for i, t in enumerate(thumbnail_names) if f'/{video_id}/{t}' in thumb['url']), n_thumbnail_names)
thumb['preference'] = (0 if '.webp' in thumb['url'] else -1) - (2 * i)
self._remove_duplicate_formats(thumbnails)
+ category = get_first(microformats, 'category') or search_meta('genre')
+ channel_id = str_or_none(
+ 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')
+ is_upcoming = get_first(video_details, 'isUpcoming')
+ if is_live is None:
+ if is_upcoming or live_content is False:
+ is_live = False
+ if is_upcoming is None and (live_content or is_live):
+ is_upcoming = False
+ live_starttime = parse_iso8601(get_first(live_broadcast_details, 'startTimestamp'))
+ live_endtime = parse_iso8601(get_first(live_broadcast_details, 'endTimestamp'))
+ if not duration and live_endtime and live_starttime:
+ duration = live_endtime - live_starttime
+
info = {
'id': video_id,
'title': self._live_title(video_title) if is_live else video_title,
'thumbnails': thumbnails,
'description': video_description,
'upload_date': unified_strdate(
- microformat.get('uploadDate')
+ get_first(microformats, 'uploadDate')
or search_meta('uploadDate')),
- 'uploader': video_details['author'],
+ '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,
'channel_id': channel_id,
- 'channel_url': 'https://www.youtube.com/channel/' + channel_id if channel_id else None,
+ 'channel_url': f'https://www.youtube.com/channel/{channel_id}' if channel_id else None,
'duration': duration,
'view_count': int_or_none(
- video_details.get('viewCount')
- or microformat.get('viewCount')
+ get_first((video_details, microformats), (..., 'viewCount'))
or search_meta('interactionCount')),
- 'average_rating': float_or_none(video_details.get('averageRating')),
+ 'average_rating': float_or_none(get_first(video_details, 'averageRating')),
'age_limit': 18 if (
- microformat.get('isFamilySafe') is False
+ get_first(microformats, 'isFamilySafe') is False
or search_meta('isFamilyFriendly') == 'false'
or search_meta('og:restrictions:age') == '18+') else 0,
'webpage_url': webpage_url,
'categories': [category] if category else None,
'tags': keywords,
+ 'playable_in_embed': get_first(playability_statuses, 'playableInEmbed'),
'is_live': is_live,
- 'playable_in_embed': playability_status.get('playableInEmbed'),
- 'was_live': video_details.get('isLiveContent'),
+ 'was_live': (False if is_live or is_upcoming or live_content is False
+ else None if is_live is None or is_upcoming is None
+ else live_content),
+ 'live_status': 'is_upcoming' if is_upcoming else None, # rest will be set by YoutubeDL
+ 'release_timestamp': live_starttime,
}
- pctr = try_get(
- player_response,
- lambda x: x['captions']['playerCaptionsTracklistRenderer'], dict)
+ pctr = traverse_obj(player_responses, (..., 'captions', 'playerCaptionsTracklistRenderer'), expected_type=dict)
+ # Converted into dicts to remove duplicates
+ captions = {
+ sub.get('baseUrl'): sub
+ for sub in traverse_obj(pctr, (..., 'captionTracks', ...), default=[])}
+ translation_languages = {
+ lang.get('languageCode'): lang.get('languageName')
+ for lang in traverse_obj(pctr, (..., 'translationLanguages', ...), default=[])}
subtitles = {}
if pctr:
def process_language(container, base_url, lang_code, sub_name, query):
'name': sub_name,
})
- for caption_track in (pctr.get('captionTracks') or []):
- base_url = caption_track.get('baseUrl')
+ for base_url, caption_track in captions.items():
if not base_url:
continue
if caption_track.get('kind') != 'asr':
continue
process_language(
subtitles, base_url, lang_code,
- try_get(caption_track, lambda x: x['name']['simpleText']),
+ traverse_obj(caption_track, ('name', 'simpleText')),
{})
continue
automatic_captions = {}
- for translation_language in (pctr.get('translationLanguages') or []):
- translation_language_code = translation_language.get('languageCode')
- if not translation_language_code:
+ for trans_code, trans_name in translation_languages.items():
+ if not trans_code:
continue
process_language(
- automatic_captions, base_url, translation_language_code,
- self._get_text(translation_language.get('languageName'), max_runs=1),
- {'tlang': translation_language_code})
+ automatic_captions, base_url, trans_code,
+ self._get_text(trans_name, max_runs=1),
+ {'tlang': trans_code})
info['automatic_captions'] = automatic_captions
info['subtitles'] = subtitles
webpage, self._YT_INITIAL_DATA_RE, video_id,
'yt initial data')
if not initial_data:
+ headers = self.generate_api_headers(
+ master_ytcfg, identity_token, self._extract_account_syncid(master_ytcfg),
+ session_index=self._extract_session_index(master_ytcfg))
+
initial_data = self._extract_response(
item_id=video_id, ep='next', fatal=False,
- ytcfg=ytcfg, headers=headers, query={'videoId': video_id},
+ ytcfg=master_ytcfg, headers=headers, query={'videoId': video_id},
note='Downloading initial data API JSON')
try:
pass
if initial_data:
- chapters = self._extract_chapters_from_json(
- initial_data, video_id, duration)
- if not chapters:
- for engagment_pannel in (initial_data.get('engagementPanels') or []):
- contents = try_get(
- engagment_pannel, lambda x: x['engagementPanelSectionListRenderer']['content']['macroMarkersListRenderer']['contents'],
- list)
- if not contents:
- continue
-
- def chapter_time(mmlir):
- return parse_duration(
- self._get_text(mmlir.get('timeDescription')))
-
- chapters = []
- for next_num, content in enumerate(contents, start=1):
- mmlir = content.get('macroMarkersListItemRenderer') or {}
- start_time = chapter_time(mmlir)
- end_time = chapter_time(try_get(
- contents, lambda x: x[next_num]['macroMarkersListItemRenderer'])) \
- if next_num < len(contents) else duration
- if start_time is None or end_time is None:
- continue
- chapters.append({
- 'start_time': start_time,
- 'end_time': end_time,
- 'title': self._get_text(mmlir.get('title')),
- })
- if chapters:
- break
- if chapters:
- info['chapters'] = chapters
+ info['chapters'] = (
+ self._extract_chapters_from_json(initial_data, duration)
+ or self._extract_chapters_from_engagement_panel(initial_data, duration)
+ or None)
contents = try_get(
initial_data,
})
vsir = content.get('videoSecondaryInfoRenderer')
if vsir:
- info['channel'] = self._get_text(try_get(
- vsir,
- lambda x: x['owner']['videoOwnerRenderer']['title'],
- dict))
+ info['channel'] = self._get_text(vsir, ('owner', 'videoOwnerRenderer', 'title'))
rows = try_get(
vsir,
lambda x: x['metadataRowContainer']['metadataRowContainerRenderer']['rows'],
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])
+ 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 v:
info[d_k] = v
- is_private = bool_or_none(video_details.get('isPrivate'))
- is_unlisted = bool_or_none(microformat.get('isUnlisted'))
+ is_private = get_first(video_details, 'isPrivate', expected_type=bool)
+ is_unlisted = get_first(microformats, 'isUnlisted', expected_type=bool)
is_membersonly = None
is_premium = None
if initial_data and is_private is not None:
get_comments = self.get_param('getcomments', False)
if get_annotations or get_comments:
xsrf_token = None
- ytcfg = self._extract_ytcfg(video_id, webpage)
- if ytcfg:
- xsrf_token = try_get(ytcfg, lambda x: x['XSRF_TOKEN'], compat_str)
+ if master_ytcfg:
+ xsrf_token = try_get(master_ytcfg, lambda x: x['XSRF_TOKEN'], compat_str)
if not xsrf_token:
xsrf_token = self._search_regex(
r'([\'"])XSRF_TOKEN\1\s*:\s*([\'"])(?P<xsrf_token>(?:(?!\2).)+)\2',
# annotations
if get_annotations:
- invideo_url = try_get(
- player_response, lambda x: x['annotations'][0]['playerAnnotationsUrlsRenderer']['invideoUrl'], compat_str)
+ invideo_url = get_first(
+ player_responses,
+ ('annotations', 0, 'playerAnnotationsUrlsRenderer', 'invideoUrl'),
+ expected_type=str)
if xsrf_token and invideo_url:
xsrf_field_name = None
- if ytcfg:
- xsrf_field_name = try_get(ytcfg, lambda x: x['XSRF_FIELD_NAME'], compat_str)
+ if master_ytcfg:
+ xsrf_field_name = try_get(master_ytcfg, lambda x: x['XSRF_FIELD_NAME'], compat_str)
if not xsrf_field_name:
xsrf_field_name = self._search_regex(
r'([\'"])XSRF_FIELD_NAME\1\s*:\s*([\'"])(?P<xsrf_field_name>\w+)\2',
data=urlencode_postdata({xsrf_field_name: xsrf_token}))
if get_comments:
- info['__post_extractor'] = lambda: self._extract_comments(ytcfg, video_id, contents, webpage)
+ info['__post_extractor'] = lambda: self._extract_comments(master_ytcfg, video_id, contents, webpage)
- self.mark_watched(video_id, player_response)
+ self.mark_watched(video_id, player_responses)
return info
}, {
'url': 'https://www.youtube.com/channel/UCoMdktPbSTixAyNGwb-UYkQ/live',
'info_dict': {
- 'id': 'X1whbWASnNQ', # This will keep changing
+ 'id': 'FMtPN8yp5LU', # This will keep changing
'ext': 'mp4',
'title': compat_str,
'uploader': 'Sky News',
renderer = self._extract_basic_item_renderer(item)
if not isinstance(renderer, dict):
continue
- title = self._get_text(renderer.get('title'))
+ title = self._get_text(renderer, 'title')
# playlist
playlist_id = renderer.get('playlistId')
# will not work
if skip_channels and '/channels?' in shelf_url:
return
- title = self._get_text(shelf_renderer, lambda x: x['title'])
+ 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):
for page_num in itertools.count(1):
if not continuation:
break
- headers = self._generate_api_headers(ytcfg, identity_token, account_syncid, visitor_data)
+ headers = self.generate_api_headers(ytcfg, identity_token, account_syncid, visitor_data)
response = self._extract_response(
item_id='%s page %s' % (item_id, page_num),
query=continuation, headers=headers, ytcfg=ytcfg,
known_renderers = {
'gridPlaylistRenderer': (self._grid_entries, 'items'),
'gridVideoRenderer': (self._grid_entries, 'items'),
+ 'gridChannelRenderer': (self._grid_entries, 'items'),
'playlistVideoRenderer': (self._playlist_entries, 'contents'),
'itemSectionRenderer': (extract_entries, 'contents'), # for feeds
'richItemRenderer': (extract_entries, 'contents'), # for hashtag
'channel': metadata['uploader'],
'channel_id': metadata['uploader_id'],
'channel_url': metadata['uploader_url']})
- ytcfg = self._extract_ytcfg(item_id, webpage)
+ ytcfg = self.extract_ytcfg(item_id, webpage)
return self.playlist_result(
self._entries(
selected_tab, playlist_id,
def _extract_mix_playlist(self, playlist, playlist_id, data, webpage):
first_id = last_id = None
- ytcfg = self._extract_ytcfg(playlist_id, webpage)
- headers = self._generate_api_headers(
+ ytcfg = self.extract_ytcfg(playlist_id, webpage)
+ headers = self.generate_api_headers(
ytcfg, account_syncid=self._extract_account_syncid(ytcfg, data),
identity_token=self._extract_identity_token(webpage, item_id=playlist_id))
for page_num in itertools.count(1):
renderer_dict, lambda x: x['privacyDropdownItemRenderer']['isSelected'], bool) or False
if not is_selected:
continue
- label = self._get_text(
- try_get(renderer_dict, lambda x: x['privacyDropdownItemRenderer']['label'], dict) or [])
+ label = self._get_text(renderer_dict, ('privacyDropdownItemRenderer', 'label'))
if label:
badge_labels.add(label.lower())
break
params = browse_endpoint.get('params')
break
- ytcfg = self._extract_ytcfg(item_id, webpage)
- headers = self._generate_api_headers(
+ ytcfg = self.extract_ytcfg(item_id, webpage)
+ headers = self.generate_api_headers(
ytcfg, account_syncid=self._extract_account_syncid(ytcfg, data),
identity_token=self._extract_identity_token(webpage, item_id=item_id),
visitor_data=try_get(
webpage = self._download_webpage(
url, item_id,
'Downloading webpage%s' % (' (retry #%d)' % count if count else ''))
- data = self._extract_yt_initial_data(item_id, webpage)
+ data = self.extract_yt_initial_data(item_id, webpage)
if data.get('contents') or data.get('currentVideoEndpoint'):
break
# Extract alerts here only when there is error
'id': 'PLBB231211A4F62143',
'uploader': 'Wickydoo',
'uploader_id': 'UCKSpbfbl5kRQpTdL7kMc-1Q',
+ 'description': 'md5:8fa6f52abb47a9552002fa3ddfc57fc2',
},
'playlist_mincount': 29,
}, {
}
}, {
'url': 'http://www.youtube.com/embed/_xDOZElKyNU?list=PLsyOSbh5bs16vubvKePAQ1x3PhKavfBIl',
- 'playlist_mincount': 982,
+ 'playlist_mincount': 654,
'info_dict': {
'title': '2018 Chinese New Singles (11/6 updated)',
'id': 'PLsyOSbh5bs16vubvKePAQ1x3PhKavfBIl',
'uploader': 'LBK',
'uploader_id': 'UC21nz3_MesPLqtDqwdvnoxA',
+ 'description': 'md5:da521864744d60a198e3a88af4db0d9d',
}
}, {
'url': 'TLGGrESM50VT6acwMjAyMjAxNw',
def _get_n_results(self, query, n):
"""Get a specified number of results for a query"""
- return self.playlist_result(self._entries(query, n), query)
+ return self.playlist_result(self._entries(query, n), query, query)
class YoutubeSearchDateIE(YoutubeSearchIE):
'url': 'https://www.youtube.com/results?baz=bar&search_query=youtube-dl+test+video&filters=video&lclk=video',
'playlist_mincount': 5,
'info_dict': {
+ 'id': 'youtube-dl test video',
'title': 'youtube-dl test video',
}
}, {
projects / yt-dlp.git / blobdiff