import threading
import time
import traceback
+import urllib.error
+import urllib.parse
from .common import InfoExtractor, SearchInfoExtractor
-from ..compat import functools # isort: split
-from ..compat import (
- compat_chr,
- compat_HTTPError,
- compat_parse_qs,
- compat_str,
- compat_urllib_parse_unquote_plus,
- compat_urllib_parse_urlencode,
- compat_urllib_parse_urlparse,
- compat_urlparse,
-)
+from ..compat import functools
from ..jsinterp import JSInterpreter
from ..utils import (
NO_DEFAULT,
ExtractorError,
+ UserNotLive,
bug_reports_message,
classproperty,
clean_html,
variadic,
)
-# any clients starting with _ cannot be explicity requested by the user
+# any clients starting with _ cannot be explicitly requested by the user
INNERTUBE_CLIENTS = {
'web': {
'INNERTUBE_API_KEY': 'AIzaSyAO_FJ2SlqU8Q4STEHLGCilw_Y9_11qcW8',
pref = {}
if pref_cookie:
try:
- pref = dict(compat_urlparse.parse_qsl(pref_cookie.value))
+ pref = dict(urllib.parse.parse_qsl(pref_cookie.value))
except ValueError:
self.report_warning('Failed to parse user PREF cookie' + bug_reports_message())
pref.update({'hl': 'en', 'tz': 'UTC'})
- self._set_cookie('.youtube.com', name='PREF', value=compat_urllib_parse_urlencode(pref))
+ self._set_cookie('.youtube.com', name='PREF', value=urllib.parse.urlencode(pref))
def _real_initialize(self):
self._initialize_pref()
def _extract_client_name(self, ytcfg, default_client='web'):
return self._ytcfg_get_safe(
ytcfg, (lambda x: x['INNERTUBE_CLIENT_NAME'],
- lambda x: x['INNERTUBE_CONTEXT']['client']['clientName']), compat_str, default_client)
+ lambda x: x['INNERTUBE_CONTEXT']['client']['clientName']), str, default_client)
def _extract_client_version(self, ytcfg, default_client='web'):
return self._ytcfg_get_safe(
ytcfg, (lambda x: x['INNERTUBE_CLIENT_VERSION'],
- lambda x: x['INNERTUBE_CONTEXT']['client']['clientVersion']), compat_str, default_client)
+ lambda x: x['INNERTUBE_CONTEXT']['client']['clientVersion']), 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 self._ytcfg_get_safe(ytcfg, lambda x: x['INNERTUBE_API_KEY'], str, default_client)
def _extract_context(self, ytcfg=None, default_client='web'):
context = get_first(
# Deprecated?
def _extract_identity_token(self, ytcfg=None, webpage=None):
if ytcfg:
- token = try_get(ytcfg, lambda x: x['ID_TOKEN'], compat_str)
+ token = try_get(ytcfg, lambda x: x['ID_TOKEN'], str)
if token:
return token
if webpage:
"""
for data in args:
# ytcfg includes channel_syncid if on secondary channel
- delegated_sid = try_get(data, lambda x: x['DELEGATED_SESSION_ID'], compat_str)
+ delegated_sid = try_get(data, lambda x: x['DELEGATED_SESSION_ID'], str)
if delegated_sid:
return delegated_sid
sync_ids = (try_get(
data, (lambda x: x['responseContext']['mainAppWebResponseContext']['datasyncId'],
- lambda x: x['DATASYNC_ID']), compat_str) or '').split('||')
+ lambda x: x['DATASYNC_ID']), str) or '').split('||')
if len(sync_ids) >= 2 and sync_ids[1]:
# datasyncid is of the form "channel_syncid||user_syncid" for secondary channel
# and just "user_syncid||" for primary channel. We only want the channel_syncid
origin = 'https://' + (self._select_api_hostname(api_hostname, default_client))
headers = {
- 'X-YouTube-Client-Name': compat_str(
+ 'X-YouTube-Client-Name': str(
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,
def _extract_continuation_ep_data(cls, continuation_ep: dict):
if isinstance(continuation_ep, dict):
continuation = try_get(
- continuation_ep, lambda x: x['continuationCommand']['token'], compat_str)
+ continuation_ep, lambda x: x['continuationCommand']['token'], str)
if not continuation:
return
ctp = continuation_ep.get('clickTrackingParams')
def _extract_badges(self, renderer: dict):
badges = set()
for badge in try_get(renderer, lambda x: x['badges'], list) or []:
- label = try_get(badge, lambda x: x['metadataBadgeRenderer']['label'], compat_str)
+ label = try_get(badge, lambda x: x['metadataBadgeRenderer']['label'], str)
if label:
badges.add(label.lower())
return badges
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)
+ text = try_get(item, lambda x: x['simpleText'], str)
if text:
return text
runs = try_get(item, lambda x: x['runs'], list) or []
note='%s%s' % (note, ' (retry #%d)' % count if count else ''))
except ExtractorError as e:
if isinstance(e.cause, network_exceptions):
- if isinstance(e.cause, compat_HTTPError):
+ if isinstance(e.cause, urllib.error.HTTPError):
first_bytes = e.cause.read(512)
if not is_html(first_bytes):
yt_error = try_get(
self._parse_json(
self._webpage_read_content(e.cause, None, item_id, prefix=first_bytes) or '{}', item_id, fatal=False),
- lambda x: x['error']['message'], compat_str)
+ lambda x: x['error']['message'], str)
if yt_error:
self._report_alerts([('ERROR', yt_error)], fatal=False)
# 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
+ # Sometimes a 404 is also received. See: https://github.com/ytdl-org/youtube-dl/issues/28289
# 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):
+ if not isinstance(e.cause, urllib.error.HTTPError) or e.cause.code not in (403, 429):
last_error = error_to_compat_str(e.cause or e.msg)
if count < retries:
continue
(?:\#|$)""" % {
'invidious': '|'.join(YoutubeBaseInfoExtractor._INVIDIOUS_SITES),
}
+ _EMBED_REGEX = [r'''(?x)
+ (?:
+ <iframe[^>]+?src=|
+ data-video-url=|
+ <embed[^>]+?src=|
+ embedSWF\(?:\s*|
+ <object[^>]+data=|
+ new\s+SWFObject\(
+ )
+ (["\'])
+ (?P<url>(?:https?:)?//(?:www\.)?youtube(?:-nocookie)?\.com/
+ (?:embed|v|p)/[0-9A-Za-z_-]{11}.*?)
+ \1''']
_PLAYER_INFO_RE = (
r'/s/player/(?P<id>[a-zA-Z0-9_-]{8,})/player',
r'/(?P<id>[a-zA-Z0-9_-]{8,})/player(?:_ias\.vflset(?:/[a-zA-Z]{2,3}_[a-zA-Z]{2,3})?|-plasma-ias-(?:phone|tablet)-[a-z]{2}_[A-Z]{2}\.vflset)/base\.js$',
'age_limit': 0,
'start_time': 1,
'end_time': 9,
+ 'comment_count': int,
'channel_follower_count': int
}
},
'thumbnail': 'https://i.ytimg.com/vi/BaW_jenozKc/maxresdefault.jpg',
'live_status': 'not_live',
'age_limit': 0,
+ 'comment_count': int,
'channel_follower_count': int
},
'params': {
'categories': ['Entertainment'],
'duration': 106,
'channel_url': 'https://www.youtube.com/channel/UC1yoRdFoFJaCY-AGfD9W0wQ',
+ 'comment_count': int,
'channel_follower_count': int
},
},
'upload_date': '20150827',
'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',
+ 'description': 'md5:04bbbf3ccceb6795947572ca36f45904',
'uploader': 'Olympics',
'title': 'Hockey - Women - GER-AUS - London 2012 Olympic Games',
'like_count': int,
'like_count': int,
'live_status': 'not_live',
'availability': 'unlisted',
+ 'comment_count': int,
'channel_follower_count': int
},
},
'thumbnail': 'https://i.ytimg.com/vi_webp/M4gD1WSo5mA/maxresdefault.webp',
'live_status': 'not_live',
'playable_in_embed': True,
+ 'comment_count': int,
'channel_follower_count': int
},
'params': {
'view_count': int,
'live_status': 'not_live',
'channel_url': 'https://www.youtube.com/channel/UCH1dpzjCEiGAt8CXkryhkZg',
+ 'comment_count': int,
'channel_follower_count': int
},
'params': {
'view_count': int,
'duration': 522,
'channel': 'kudvenkat',
+ 'comment_count': int,
'channel_follower_count': int
},
'params': {
'availability': 'public',
'channel': 'Leon Nguyen',
'thumbnail': 'https://i.ytimg.com/vi_webp/2NUZ8W2llS4/maxresdefault.webp',
+ 'comment_count': int,
'channel_follower_count': int
}
}, {
'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',
'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',
- }
+ },
+ 'skip': 'stories get removed after some period of time',
}, {
'url': 'https://www.youtube.com/watch?v=tjjjtzRLHvA',
'info_dict': {
}
]
+ _WEBPAGE_TESTS = [
+ # YouTube <object> embed
+ {
+ 'url': 'http://www.improbable.com/2017/04/03/untrained-modern-youths-and-ancient-masters-in-selfie-portraits/',
+ 'md5': '873c81d308b979f0e23ee7e620b312a3',
+ 'info_dict': {
+ 'id': 'msN87y-iEx0',
+ 'ext': 'mp4',
+ 'title': 'Feynman: Mirrors FUN TO IMAGINE 6',
+ 'upload_date': '20080526',
+ 'description': 'md5:873c81d308b979f0e23ee7e620b312a3',
+ 'uploader': 'Christopher Sykes',
+ 'uploader_id': 'ChristopherJSykes',
+ 'age_limit': 0,
+ 'tags': ['feynman', 'mirror', 'science', 'physics', 'imagination', 'fun', 'cool', 'puzzle'],
+ 'channel_id': 'UCCeo--lls1vna5YJABWAcVA',
+ 'playable_in_embed': True,
+ 'thumbnail': 'https://i.ytimg.com/vi/msN87y-iEx0/hqdefault.jpg',
+ 'like_count': int,
+ 'comment_count': int,
+ 'channel': 'Christopher Sykes',
+ 'live_status': 'not_live',
+ 'channel_url': 'https://www.youtube.com/channel/UCCeo--lls1vna5YJABWAcVA',
+ 'availability': 'public',
+ 'duration': 195,
+ 'view_count': int,
+ 'categories': ['Science & Technology'],
+ 'channel_follower_count': int,
+ 'uploader_url': 'http://www.youtube.com/user/ChristopherJSykes',
+ },
+ 'params': {
+ 'skip_download': True,
+ }
+ },
+ ]
+
@classmethod
def suitable(cls, url):
from ..utils import parse_qs
microformats = traverse_obj(
prs, (..., 'microformat', 'playerMicroformatRenderer'),
expected_type=dict, default=[])
- _, is_live, _, formats = self._list_formats(video_id, microformats, video_details, prs, player_url)
+ _, is_live, _, formats, _ = self._list_formats(video_id, microformats, video_details, prs, player_url)
start_time = time.time()
def mpd_feed(format_id, delay):
# Obtain from MPD's maximum seq value
old_mpd_url = mpd_url
last_error = ctx.pop('last_error', None)
- expire_fast = immediate or last_error and isinstance(last_error, compat_HTTPError) and last_error.code == 403
+ expire_fast = immediate or last_error and isinstance(last_error, urllib.error.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:
def _extract_player_url(self, *ytcfgs, webpage=None):
player_url = traverse_obj(
ytcfgs, (..., 'PLAYER_JS_URL'), (..., 'WEB_PLAYER_CONTEXT_CONFIGS', ..., 'jsUrl'),
- get_all=False, expected_type=compat_str)
+ get_all=False, expected_type=str)
if not player_url:
return
return urljoin('https://www.youtube.com', player_url)
def _signature_cache_id(self, example_sig):
""" Return a string representation of a signature """
- return '.'.join(compat_str(len(part)) for part in example_sig.split('.'))
+ return '.'.join(str(len(part)) for part in example_sig.split('.'))
@classmethod
def _extract_player_info(cls, player_url):
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)
+ self.write_debug(f'Extracting signature function {func_id}')
+ cache_spec = self.cache.load('youtube-sigfuncs', func_id)
if cache_spec is not None:
return lambda s: ''.join(s[i] for i in cache_spec)
if code:
res = self._parse_sig_js(code)
- test_string = ''.join(map(compat_chr, range(len(example_sig))))
+ test_string = ''.join(map(chr, range(len(example_sig))))
cache_res = res(test_string)
cache_spec = [ord(c) for c in cache_res]
- self._downloader.cache.store('youtube-sigfuncs', func_id, cache_spec)
+ self.cache.store('youtube-sigfuncs', func_id, cache_spec)
return res
def _print_sig_code(self, func, example_sig):
else:
yield _genslice(start, i, step)
- test_string = ''.join(map(compat_chr, range(len(example_sig))))
+ test_string = ''.join(map(chr, range(len(example_sig))))
cache_res = func(test_string)
cache_spec = [ord(c) for c in cache_res]
expr_code = ' + '.join(gen_sig_code(cache_spec))
signature_id_tuple = '(%s)' % (
- ', '.join(compat_str(len(p)) for p in example_sig.split('.')))
+ ', '.join(str(len(p)) for p in example_sig.split('.')))
code = ('if tuple(len(p) for p in s.split(\'.\')) == %s:\n'
' return %s\n') % (signature_id_tuple, expr_code)
self.to_screen('Extracted signature function:\n' + code)
def _extract_n_function(self, video_id, player_url):
player_id = self._extract_player_info(player_url)
- func_code = self._downloader.cache.load('youtube-nsig', player_id)
+ func_code = self.cache.load('youtube-nsig', player_id)
if func_code:
jsi = JSInterpreter(func_code)
funcname = self._extract_n_function_name(jscode)
jsi = JSInterpreter(jscode)
func_code = jsi.extract_function_code(funcname)
- self._downloader.cache.store('youtube-nsig', player_id, func_code)
+ self.cache.store('youtube-nsig', player_id, func_code)
if self.get_param('youtube_print_sig_code'):
self.to_screen(f'Extracted nsig function from {player_id}:\n{func_code[1]}\n')
return sts
def _mark_watched(self, video_id, player_responses):
- playback_url = get_first(
- player_responses, ('playbackTracking', 'videostatsPlaybackUrl', 'baseUrl'),
- expected_type=url_or_none)
- 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)
+ for is_full, key in enumerate(('videostatsPlaybackUrl', 'videostatsWatchtimeUrl')):
+ label = 'fully ' if is_full else ''
+ url = get_first(player_responses, ('playbackTracking', key, 'baseUrl'),
+ expected_type=url_or_none)
+ if not url:
+ self.report_warning(f'Unable to mark {label}watched')
+ return
+ parsed_url = urllib.parse.urlparse(url)
+ qs = urllib.parse.parse_qs(parsed_url.query)
+
+ # 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))
+
+ # # more consistent results setting it to right before the end
+ video_length = [str(float((qs.get('len') or ['1.5'])[0]) - 1)]
+
+ qs.update({
+ 'ver': ['2'],
+ 'cpn': [cpn],
+ 'cmt': video_length,
+ 'el': 'detailpage', # otherwise defaults to "shorts"
+ })
- # 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))
+ if is_full:
+ # these seem to mark watchtime "history" in the real world
+ # they're required, so send in a single value
+ qs.update({
+ 'st': video_length,
+ 'et': video_length,
+ })
- qs.update({
- 'ver': ['2'],
- 'cpn': [cpn],
- })
- playback_url = compat_urlparse.urlunparse(
- parsed_playback_url._replace(query=compat_urllib_parse_urlencode(qs, True)))
+ url = urllib.parse.urlunparse(
+ parsed_url._replace(query=urllib.parse.urlencode(qs, True)))
- self._download_webpage(
- playback_url, video_id, 'Marking watched',
- 'Unable to mark watched', fatal=False)
+ self._download_webpage(
+ url, video_id, f'Marking {label}watched',
+ 'Unable to mark watched', fatal=False)
- @staticmethod
- def _extract_urls(webpage):
- # Embedded YouTube player
- entries = [
- unescapeHTML(mobj.group('url'))
- for mobj in re.finditer(r'''(?x)
- (?:
- <iframe[^>]+?src=|
- data-video-url=|
- <embed[^>]+?src=|
- embedSWF\(?:\s*|
- <object[^>]+data=|
- new\s+SWFObject\(
- )
- (["\'])
- (?P<url>(?:https?:)?//(?:www\.)?youtube(?:-nocookie)?\.com/
- (?:embed|v|p)/[0-9A-Za-z_-]{11}.*?)
- \1''', webpage)]
+ @classmethod
+ def _extract_from_webpage(cls, url, webpage):
+ # Invidious Instances
+ # https://github.com/yt-dlp/yt-dlp/issues/195
+ # https://github.com/iv-org/invidious/pull/1730
+ mobj = re.search(
+ r'<link rel="alternate" href="(?P<url>https://www\.youtube\.com/watch\?v=[0-9A-Za-z_-]{11})"',
+ webpage)
+ if mobj:
+ yield cls.url_result(mobj.group('url'), cls)
+ raise cls.StopExtraction()
+
+ yield from super()._extract_from_webpage(url, webpage)
# lazyYT YouTube embed
- entries.extend(list(map(
- unescapeHTML,
- re.findall(r'class="lazyYT" data-youtube-id="([^"]+)"', webpage))))
+ for id_ in re.findall(r'class="lazyYT" data-youtube-id="([^"]+)"', webpage):
+ yield cls.url_result(unescapeHTML(id_), cls, id_)
# Wordpress "YouTube Video Importer" plugin
- matches = re.findall(r'''(?x)<div[^>]+
- class=(?P<q1>[\'"])[^\'"]*\byvii_single_video_player\b[^\'"]*(?P=q1)[^>]+
- data-video_id=(?P<q2>[\'"])([^\'"]+)(?P=q2)''', webpage)
- entries.extend(m[-1] for m in matches)
-
- return entries
-
- @staticmethod
- def _extract_url(webpage):
- urls = YoutubeIE._extract_urls(webpage)
- return urls[0] if urls else None
+ for m in re.findall(r'''(?x)<div[^>]+
+ class=(?P<q1>[\'"])[^\'"]*\byvii_single_video_player\b[^\'"]*(?P=q1)[^>]+
+ data-video_id=(?P<q2>[\'"])([^\'"]+)(?P=q2)''', webpage):
+ yield cls.url_result(m[-1], cls, m[-1])
@classmethod
def extract_id(cls, url):
- mobj = re.match(cls._VALID_URL, url, re.VERBOSE)
- if mobj is None:
- raise ExtractorError('Invalid URL: %s' % url)
- return mobj.group('id')
+ video_id = cls.get_temp_id(url)
+ if not video_id:
+ raise ExtractorError(f'Invalid URL: {url}')
+ return video_id
def _extract_chapters_from_json(self, data, duration):
chapter_list = traverse_obj(
if not strict:
chapter_list.sort(key=lambda c: c['start_time'] or 0)
- chapters = [{'start_time': 0, 'title': '<Untitled>'}]
+ chapters = [{'start_time': 0}]
for idx, chapter in enumerate(chapter_list):
- if chapter['start_time'] is None or not chapter['title']:
+ if chapter['start_time'] is None:
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:]
+ return chapters[1:]
def _extract_comment(self, comment_renderer, parent=None):
comment_id = comment_renderer.get('commentId')
timestamp, time_text = self._extract_time_text(comment_renderer, 'publishedTimeText')
author = self._get_text(comment_renderer, 'authorText')
author_id = try_get(comment_renderer,
- lambda x: x['authorEndpoint']['browseEndpoint']['browseId'], compat_str)
+ lambda x: x['authorEndpoint']['browseEndpoint']['browseId'], str)
votes = parse_count(try_get(comment_renderer, (lambda x: x['voteCount']['simpleText'],
- lambda x: x['likeCount']), compat_str)) or 0
+ lambda x: x['likeCount']), str)) or 0
author_thumbnail = try_get(comment_renderer,
- lambda x: x['authorThumbnail']['thumbnails'][-1]['url'], compat_str)
+ lambda x: x['authorThumbnail']['thumbnails'][-1]['url'], str)
author_is_uploader = try_get(comment_renderer, lambda x: x['authorIsChannelOwner'], bool)
is_favorited = 'creatorHeart' in (try_get(
self.report_warning(last_error)
return prs, player_url
- def _extract_formats(self, streaming_data, video_id, player_url, is_live, duration):
+ def _extract_formats_and_subtitles(self, streaming_data, video_id, player_url, is_live, duration):
itags, stream_ids = {}, []
itag_qualities, res_qualities = {}, {}
q = qualities([
fmt_url = fmt.get('url')
if not fmt_url:
- sc = compat_parse_qs(fmt.get('signatureCipher'))
+ sc = urllib.parse.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 all((sc, fmt_url, player_url, encrypted_sig)):
if val in qdict), -1)
return True
+ subtitles = {}
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):
+ fmts, subs = self._extract_m3u8_formats_and_subtitles(hls_manifest_url, video_id, 'mp4', fatal=False, live=is_live)
+ subtitles = self._merge_subtitles(subs, subtitles)
+ for f in fmts:
if process_manifest_format(f, 'hls', self._search_regex(
r'/itag/(\d+)', f['url'], 'itag', default=None)):
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):
+ formats, subs = self._extract_mpd_formats_and_subtitles(dash_manifest_url, video_id, fatal=False)
+ subtitles = self._merge_subtitles(subs, subtitles) # Prioritize HLS subs over DASH
+ for f in formats:
if process_manifest_format(f, 'dash', f['format_id']):
f['filesize'] = int_or_none(self._search_regex(
r'/clen/(\d+)', f.get('fragment_base_url') or f['url'], 'file size', default=None))
f['is_from_start'] = True
yield f
+ yield subtitles
def _extract_storyboard(self, player_responses, duration):
spec = get_first(
'url': url,
'width': width,
'height': height,
+ 'fps': frame_count / duration,
+ 'rows': rows,
+ 'columns': cols,
'fragments': [{
'url': url.replace('$M', str(j)),
'duration': min(fragment_duration, duration - (j * fragment_duration)),
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, duration))
+ *formats, subtitles = self._extract_formats_and_subtitles(streaming_data, video_id, player_url, is_live, duration)
- return live_broadcast_details, is_live, streaming_data, formats
+ return live_broadcast_details, is_live, streaming_data, formats, subtitles
def _real_extract(self, url):
url, smuggled_data = unsmuggle_url(url, {})
# 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))
+ feed_data = urllib.parse.parse_qs(
+ urllib.parse.unquote_plus(feed))
def feed_entry(name):
return try_get(
- feed_data, lambda x: x[name][0], compat_str)
+ feed_data, lambda x: x[name][0], str)
feed_id = feed_entry('id')
if not feed_id:
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)
+ live_broadcast_details, is_live, streaming_data, formats, automatic_captions = \
+ self._list_formats(video_id, microformats, video_details, player_responses, player_url)
if not formats:
if not self.get_param('allow_unplayable_formats') and traverse_obj(streaming_data, (..., 'licenseInfos')):
# See: https://github.com/yt-dlp/yt-dlp/issues/340
# List of possible thumbnails - Ref: <https://stackoverflow.com/a/20542029>
thumbnail_names = [
- # While the *1,*2,*3 thumbnails are just below their correspnding "*default" variants
+ # While the *1,*2,*3 thumbnails are just below their corresponding "*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'
formats.extend(self._extract_storyboard(player_responses, duration))
- # Source is given priority since formats that throttle are given lower source_preference
- # When throttling issue is fully fixed, remove this
+ # source_preference is lower for throttled/potentially damaged formats
self._sort_formats(formats, ('quality', 'res', 'fps', 'hdr:12', 'source', 'codec:vp9.2', 'lang', 'proto'))
info = {
'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': format_field(channel_id, template='https://www.youtube.com/channel/%s'),
+ 'channel_url': format_field(channel_id, None, 'https://www.youtube.com/channel/%s'),
'duration': duration,
'view_count': int_or_none(
get_first((video_details, microformats), (..., 'viewCount'))
'release_timestamp': live_start_time,
}
+ if get_first(video_details, 'isPostLiveDvr'):
+ self.write_debug('Video is in Post-Live Manifestless mode')
+ info['live_status'] = 'post_live'
+ 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')
+
+ subtitles = {}
pctr = traverse_obj(player_responses, (..., 'captions', 'playerCaptionsTracklistRenderer'), expected_type=dict)
if pctr:
def get_lang_code(track):
'name': sub_name,
})
- subtitles, automatic_captions = {}, {}
+ # NB: Constructing the full subtitle dictionary is slow
+ get_translated_subs = 'translated_subs' not in self._configuration_arg('skip') and (
+ self.get_param('writeautomaticsub', False) or self.get_param('listsubtitles'))
for lang_code, caption_track in captions.items():
base_url = caption_track.get('baseUrl')
orig_lang = parse_qs(base_url).get('lang', [None])[-1]
continue
orig_trans_code = trans_code
if caption_track.get('kind') != 'asr':
- if 'translated_subs' in self._configuration_arg('skip'):
+ if not get_translated_subs:
continue
trans_code += f'-{lang_code}'
- trans_name += format_field(lang_name, template=' from %s')
+ trans_name += format_field(lang_name, None, ' from %s')
# 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}':
# 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
- parsed_url = compat_urllib_parse_urlparse(url)
+ info['automatic_captions'] = automatic_captions
+ info['subtitles'] = subtitles
+
+ parsed_url = urllib.parse.urlparse(url)
for component in [parsed_url.fragment, parsed_url.query]:
- query = compat_parse_qs(component)
+ query = urllib.parse.parse_qs(component)
for k, v in query.items():
for d_k, s_ks in [('start', ('start', 't')), ('end', ('end',))]:
d_k += '_time'
headers=self.generate_api_headers(ytcfg=master_ytcfg),
note='Downloading initial data API JSON')
+ info['comment_count'] = traverse_obj(initial_data, (
+ 'contents', 'twoColumnWatchNextResults', 'results', 'results', 'contents', ..., 'itemSectionRenderer',
+ 'contents', ..., 'commentsEntryPointHeaderRenderer', 'commentCount', 'simpleText'
+ ), (
+ 'engagementPanels', lambda _, v: v['engagementPanelSectionListRenderer']['panelIdentifier'] == 'comment-item-section',
+ 'engagementPanelSectionListRenderer', 'header', 'engagementPanelTitleHeaderRenderer', 'contextualInfo', 'runs', ..., 'text'
+ ), expected_type=int_or_none, get_all=False)
+
try: # This will error if there is no livechat
initial_data['contents']['twoColumnWatchNextResults']['conversationBar']['liveChatRenderer']['continuations'][0]['reloadContinuationData']['continuation']
except (KeyError, IndexError, TypeError):
# generic endpoint URL support
ep_url = urljoin('https://www.youtube.com/', try_get(
renderer, lambda x: x['navigationEndpoint']['commandMetadata']['webCommandMetadata']['url'],
- compat_str))
+ str))
if ep_url:
for ie in (YoutubeTabIE, YoutubePlaylistIE, YoutubeIE):
if ie.suitable(ep_url):
def _shelf_entries(self, shelf_renderer, skip_channels=False):
ep = try_get(
shelf_renderer, lambda x: x['endpoint']['commandMetadata']['webCommandMetadata']['url'],
- compat_str)
+ str)
shelf_url = urljoin('https://www.youtube.com', ep)
if shelf_url:
# Skipping links to another channels, note that checking for
yield entry
# playlist attachment
playlist_id = try_get(
- post_renderer, lambda x: x['backstageAttachment']['playlistRenderer']['playlistId'], compat_str)
+ post_renderer, lambda x: x['backstageAttachment']['playlistRenderer']['playlistId'], str)
if playlist_id:
yield self.url_result(
'https://www.youtube.com/playlist?list=%s' % playlist_id,
if not isinstance(run, dict):
continue
ep_url = try_get(
- run, lambda x: x['navigationEndpoint']['urlEndpoint']['url'], compat_str)
+ run, lambda x: x['navigationEndpoint']['urlEndpoint']['url'], str)
if not ep_url:
continue
if not YoutubeIE.suitable(ep_url):
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)
+ owner, lambda x: x['navigationEndpoint']['browseEndpoint']['browseId'], str)
uploader['uploader_url'] = urljoin(
'https://www.youtube.com/',
- try_get(owner, lambda x: x['navigationEndpoint']['browseEndpoint']['canonicalBaseUrl'], compat_str))
+ try_get(owner, lambda x: x['navigationEndpoint']['browseEndpoint']['canonicalBaseUrl'], str))
return {k: v for k, v in uploader.items() if v is not None}
def _extract_from_tabs(self, item_id, ytcfg, data, tabs):
def _extract_from_playlist(self, item_id, url, data, playlist, ytcfg):
title = playlist.get('title') or try_get(
- data, lambda x: x['titleText']['simpleText'], compat_str)
+ data, lambda x: x['titleText']['simpleText'], str)
playlist_id = playlist.get('playlistId') or item_id
# Delegating everything except mix playlists to regular tab-based playlist URL
playlist_url = urljoin(url, try_get(
playlist, lambda x: x['endpoint']['commandMetadata']['webCommandMetadata']['url'],
- compat_str))
+ str))
# Some playlists are unviewable but YouTube still provides a link to the (broken) playlist page [1]
# [1] MLCT, RLTDwFCb4jeqaKWnciAYM-ZVHg
continue
nav_item_renderer = menu_item.get('menuNavigationItemRenderer')
text = try_get(
- nav_item_renderer, lambda x: x['text']['simpleText'], compat_str)
+ nav_item_renderer, lambda x: x['text']['simpleText'], str)
if not text or text.lower() != 'show unavailable videos':
continue
browse_endpoint = try_get(
data = self.extract_yt_initial_data(item_id, webpage or '', fatal=fatal) or {}
except ExtractorError as e:
if isinstance(e.cause, network_exceptions):
- if not isinstance(e.cause, compat_HTTPError) or e.cause.code not in (403, 429):
+ if not isinstance(e.cause, urllib.error.HTTPError) or e.cause.code not in (403, 429):
last_error = error_to_compat_str(e.cause or e.msg)
if count < retries:
continue
}, {
'url': 'https://www.youtube.com/channel/UCoMdktPbSTixAyNGwb-UYkQ/live',
'info_dict': {
- 'id': 'GgL890LIznQ', # This will keep changing
+ 'id': 'Wq15eF5vCbI', # This will keep changing
'ext': 'mp4',
'title': str,
'uploader': 'Sky News',
'uploader': 'NoCopyrightSounds',
'description': 'Providing you with copyright free / safe music for gaming, live streaming, studying and more!',
'uploader_id': 'UC_aEa8K-EOJ3D6gOs7HcyNg',
- 'title': 'NCS Releases',
+ 'title': 'NCS : All Releases 💿',
'uploader_url': 'https://www.youtube.com/c/NoCopyrightSounds',
'channel_url': 'https://www.youtube.com/c/NoCopyrightSounds',
'modified_date': r're:\d{8}',
'title': 'yt-dlp unlisted playlist test',
'availability': 'unlisted',
'tags': [],
- 'modified_date': '20211208',
+ 'modified_date': '20220418',
'channel': 'colethedj',
'view_count': int,
'description': '',
'channel': 'pukkandan',
'description': 'Test for collaborative playlist',
'title': 'yt-dlp test - collaborative playlist',
+ 'view_count': int,
'uploader_url': 'https://www.youtube.com/channel/UCKcqXmCcyqnhgpA5P0oHH_Q',
},
'playlist_mincount': 2
@YoutubeTabBaseInfoExtractor.passthrough_smuggled_data
def _real_extract(self, url, smuggled_data):
item_id = self._match_id(url)
- url = compat_urlparse.urlunparse(
- compat_urlparse.urlparse(url)._replace(netloc='www.youtube.com'))
+ url = urllib.parse.urlunparse(
+ urllib.parse.urlparse(url)._replace(netloc='www.youtube.com'))
compat_opts = self.get_param('compat_opts', [])
def get_mobj(url):
mdata = self._extract_tab_endpoint(
f'https://music.youtube.com/channel/{item_id}', item_id, default_client='web_music')
murl = traverse_obj(mdata, ('microformat', 'microformatDataRenderer', 'urlCanonical'),
- get_all=False, expected_type=compat_str)
+ get_all=False, expected_type=str)
if not murl:
raise ExtractorError('Failed to resolve album to playlist')
return self.url_result(murl, ie=YoutubeTabIE.ie_key())
selected_tab_name = 'featured'
requested_tab_name = mobj['tab'][1:]
if 'no-youtube-channel-redirect' not in compat_opts:
- if requested_tab_name == 'live':
- # Live tab should have redirected to the video
- raise ExtractorError('The channel is not currently live', expected=True)
+ if requested_tab_name == 'live': # Live tab should have redirected to the video
+ raise UserNotLive(video_id=mobj['id'])
if requested_tab_name not in ('', selected_tab_name):
redirect_warning = f'The channel does not have a {requested_tab_name} tab'
if not original_tab_name:
'expected_warnings': [r'[Uu]navailable videos (are|will be) hidden'],
}, {
'url': 'http://www.youtube.com/embed/_xDOZElKyNU?list=PLsyOSbh5bs16vubvKePAQ1x3PhKavfBIl',
- 'playlist_mincount': 654,
+ 'playlist_mincount': 455,
'info_dict': {
'title': '2018 Chinese New Singles (11/6 updated)',
'id': 'PLsyOSbh5bs16vubvKePAQ1x3PhKavfBIl',
'channel_url': 'https://www.youtube.com/channel/UCEfMCQ9bs3tjvjy1s451zaw',
'availability': 'public',
'duration': 59,
+ 'comment_count': int,
+ 'channel_follower_count': int
},
'params': {
'noplaylist': True,
'info_dict': {
'id': '#cats',
'title': '#cats',
- 'entries': [{
- 'url': r're:https://(www\.)?youtube\.com/hashtag/cats',
- 'title': '#cats',
- }],
+ # The test suite does not have support for nested playlists
+ # 'entries': [{
+ # 'url': r're:https://(www\.)?youtube\.com/hashtag/cats',
+ # 'title': '#cats',
+ # }],
},
}, {
'url': 'https://www.youtube.com/results?q=test&sp=EgQIBBgB',
if params:
section = next((k for k, v in self._SECTIONS.items() if v == params), params)
else:
- section = compat_urllib_parse_unquote_plus((url.split('#') + [''])[1]).lower()
+ section = urllib.parse.unquote_plus((url.split('#') + [''])[1]).lower()
params = self._SECTIONS.get(section)
if not params:
section = None
expected=True)
-class YoutubeClipIE(InfoExtractor):
+class YoutubeClipIE(YoutubeTabBaseInfoExtractor):
IE_NAME = 'youtube:clip'
- IE_DESC = False # Do not list
- _VALID_URL = r'https?://(?:www\.)?youtube\.com/clip/'
+ _VALID_URL = r'https?://(?:www\.)?youtube\.com/clip/(?P<id>[^/?#]+)'
+ _TESTS = [{
+ # FIXME: Other metadata should be extracted from the clip, not from the base video
+ 'url': 'https://www.youtube.com/clip/UgytZKpehg-hEMBSn3F4AaABCQ',
+ 'info_dict': {
+ 'id': 'UgytZKpehg-hEMBSn3F4AaABCQ',
+ 'ext': 'mp4',
+ 'section_start': 29.0,
+ 'section_end': 39.7,
+ 'duration': 10.7,
+ 'age_limit': 0,
+ 'availability': 'public',
+ 'categories': ['Gaming'],
+ 'channel': 'Scott The Woz',
+ 'channel_id': 'UC4rqhyiTs7XyuODcECvuiiQ',
+ 'channel_url': 'https://www.youtube.com/channel/UC4rqhyiTs7XyuODcECvuiiQ',
+ 'description': 'md5:7a4517a17ea9b4bd98996399d8bb36e7',
+ 'like_count': int,
+ 'playable_in_embed': True,
+ 'tags': 'count:17',
+ 'thumbnail': 'https://i.ytimg.com/vi_webp/ScPX26pdQik/maxresdefault.webp',
+ 'title': 'Mobile Games on Console - Scott The Woz',
+ 'upload_date': '20210920',
+ 'uploader': 'Scott The Woz',
+ 'uploader_id': 'scottthewoz',
+ 'uploader_url': 'http://www.youtube.com/user/scottthewoz',
+ 'view_count': int,
+ 'live_status': 'not_live',
+ 'channel_follower_count': int
+ }
+ }]
def _real_extract(self, url):
- self.report_warning('YouTube clips are not currently supported. The entire video will be downloaded instead')
- return self.url_result(url, 'Generic')
+ clip_id = self._match_id(url)
+ _, data = self._extract_webpage(url, clip_id)
+
+ video_id = traverse_obj(data, ('currentVideoEndpoint', 'watchEndpoint', 'videoId'))
+ if not video_id:
+ raise ExtractorError('Unable to find video ID')
+
+ clip_data = traverse_obj(data, (
+ 'engagementPanels', ..., 'engagementPanelSectionListRenderer', 'content', 'clipSectionRenderer',
+ 'contents', ..., 'clipAttributionRenderer', 'onScrubExit', 'commandExecutorCommand', 'commands', ...,
+ 'openPopupAction', 'popup', 'notificationActionRenderer', 'actionButton', 'buttonRenderer', 'command',
+ 'commandExecutorCommand', 'commands', ..., 'loopCommand'), get_all=False)
+
+ return {
+ '_type': 'url_transparent',
+ 'url': f'https://www.youtube.com/watch?v={video_id}',
+ 'ie_key': YoutubeIE.ie_key(),
+ 'id': clip_id,
+ 'section_start': int(clip_data['startTimeMs']) / 1000,
+ 'section_end': int(clip_data['endTimeMs']) / 1000,
+ }
class YoutubeTruncatedIDIE(InfoExtractor):
projects / yt-dlp.git / blobdiff