X-Git-Url: https://jfr.im/git/yt-dlp.git/blobdiff_plain/ff2751ac9cc7d4150797d3207da9b566396bc796..0ba692acc8feffd46b6e1085fb4a2849b685945c:/yt_dlp/extractor/youtube.py diff --git a/yt_dlp/extractor/youtube.py b/yt_dlp/extractor/youtube.py index 145b89a6e..ae1c1bca5 100644 --- a/yt_dlp/extractor/youtube.py +++ b/yt_dlp/extractor/youtube.py @@ -2,7 +2,10 @@ from __future__ import unicode_literals +import base64 import calendar +import copy +import datetime import hashlib import itertools import json @@ -26,6 +29,7 @@ from ..jsinterp import JSInterpreter from ..utils import ( bool_or_none, + bytes_to_intlist, clean_html, dict_get, datetime_from_str, @@ -34,8 +38,10 @@ format_field, float_or_none, int_or_none, + intlist_to_bytes, mimetype2ext, parse_codecs, + parse_count, parse_duration, qualities, remove_start, @@ -49,7 +55,8 @@ update_url_query, url_or_none, urlencode_postdata, - urljoin + urljoin, + variadic ) @@ -67,8 +74,8 @@ class YoutubeBaseInfoExtractor(InfoExtractor): _TFA_URL = 'https://accounts.google.com/_/signin/challenge?hl=en&TL={0}' _RESERVED_NAMES = ( - r'channel|c|user|browse|playlist|watch|w|v|embed|e|watch_popup|' - r'movies|results|shared|hashtag|trending|feed|feeds|oembed|' + r'channel|c|user|browse|playlist|watch|w|v|embed|e|watch_popup|shorts|' + r'movies|results|shared|hashtag|trending|feed|feeds|oembed|get_video_info|' r'storefront|oops|index|account|reporthistory|t/terms|about|upload|signin|logout') _NETRC_MACHINE = 'youtube' @@ -85,7 +92,21 @@ def _login(self): If _LOGIN_REQUIRED is set and no authentication was provided, an error is raised. """ + + def warn(message): + self.report_warning(message) + + # username+password login is broken + if self._LOGIN_REQUIRED and self.get_param('cookiefile') is None: + self.raise_login_required( + 'Login details are needed to download this content', method='cookies') username, password = self._get_login_info() + if username: + warn('Logging in using username and password is broken. %s' % self._LOGIN_HINTS['cookies']) + return + + # Everything below this is broken! + r''' # No authentication to be performed if username is None: if self._LOGIN_REQUIRED and self.get_param('cookiefile') is None: @@ -126,9 +147,6 @@ def req(url, f_req, note, errnote): 'Google-Accounts-XSRF': 1, }) - def warn(message): - self.report_warning(message) - lookup_req = [ username, None, [], None, 'US', None, None, 2, False, True, @@ -261,6 +279,7 @@ def warn(message): return False return True + ''' def _initialize_consent(self): cookies = self._get_cookies('https://www.youtube.com/') @@ -284,39 +303,185 @@ def _real_initialize(self): if not self._login(): return - _YT_WEB_CLIENT_VERSION = '2.20210407.08.00' - _YT_INNERTUBE_API_KEY = 'AIzaSyAO_FJ2SlqU8Q4STEHLGCilw_Y9_11qcW8' _YT_INITIAL_DATA_RE = r'(?:window\s*\[\s*["\']ytInitialData["\']\s*\]|ytInitialData)\s*=\s*({.+?})\s*;' _YT_INITIAL_PLAYER_RESPONSE_RE = r'ytInitialPlayerResponse\s*=\s*({.+?})\s*;' _YT_INITIAL_BOUNDARY_RE = r'(?:var\s+meta|= 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 - return sync_ids[0] - # ytcfg includes channel_syncid if on secondary channel - return data.get('DELEGATED_SESSION_ID') + 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) + 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("||") + 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 + return sync_ids[0] def _extract_ytcfg(self, video_id, webpage): if not webpage: @@ -358,70 +526,219 @@ def _extract_ytcfg(self, video_id, webpage): r'ytcfg\.set\s*\(\s*({.+?})\s*\)\s*;', webpage, 'ytcfg', default='{}'), video_id, fatal=False) or {} - def __extract_client_version(self, ytcfg): - return try_get(ytcfg, lambda x: x['INNERTUBE_CLIENT_VERSION'], compat_str) or self._YT_WEB_CLIENT_VERSION - - def _extract_context(self, ytcfg=None): - context = try_get(ytcfg, lambda x: x['INNERTUBE_CONTEXT'], dict) - if context: - return context - - # Recreate the client context (required) - client_version = self.__extract_client_version(ytcfg) - client_name = try_get(ytcfg, lambda x: x['INNERTUBE_CLIENT_NAME'], compat_str) or 'WEB' - context = { - 'client': { - 'clientName': client_name, - 'clientVersion': client_version, - } - } - visitor_data = try_get(ytcfg, lambda x: x['VISITOR_DATA'], compat_str) - if visitor_data: - context['client']['visitorData'] = visitor_data - return context - - def _generate_api_headers(self, ytcfg=None, identity_token=None, account_syncid=None, visitor_data=None): + 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)) headers = { - 'X-YouTube-Client-Name': '1', - 'X-YouTube-Client-Version': self.__extract_client_version(ytcfg), + '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), + 'Origin': origin } + if not visitor_data and ytcfg: + visitor_data = try_get( + self._extract_context(ytcfg, client), lambda x: x['client']['visitorData'], compat_str) if identity_token: - headers['x-youtube-identity-token'] = identity_token + headers['X-Youtube-Identity-Token'] = identity_token if account_syncid: headers['X-Goog-PageId'] = account_syncid - headers['X-Goog-AuthUser'] = 0 + if session_index is None and ytcfg: + session_index = self._extract_session_index(ytcfg) + if account_syncid or session_index is not None: + headers['X-Goog-AuthUser'] = session_index if session_index is not None else 0 if visitor_data: - headers['x-goog-visitor-id'] = visitor_data - auth = self._generate_sapisidhash_header() + headers['X-Goog-Visitor-Id'] = visitor_data + auth = self._generate_sapisidhash_header(origin) if auth is not None: headers['Authorization'] = auth - headers['X-Origin'] = 'https://www.youtube.com' + headers['X-Origin'] = origin return headers + @staticmethod + def _build_api_continuation_query(continuation, ctp=None): + query = { + 'continuation': continuation + } + # TODO: Inconsistency with clickTrackingParams. + # Currently we have a fixed ctp contained within context (from ytcfg) + # and a ctp in root query for continuation. + if ctp: + query['clickTracking'] = {'clickTrackingParams': ctp} + return query + + @classmethod + def _extract_next_continuation_data(cls, renderer): + next_continuation = try_get( + renderer, (lambda x: x['continuations'][0]['nextContinuationData'], + lambda x: x['continuation']['reloadContinuationData']), dict) + if not next_continuation: + return + continuation = next_continuation.get('continuation') + if not continuation: + return + ctp = next_continuation.get('clickTrackingParams') + return cls._build_api_continuation_query(continuation, ctp) + + @classmethod + 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) + if not continuation: + return + ctp = continuation_ep.get('clickTrackingParams') + return cls._build_api_continuation_query(continuation, ctp) + + @classmethod + def _extract_continuation(cls, renderer): + next_continuation = cls._extract_next_continuation_data(renderer) + if next_continuation: + return next_continuation + + contents = [] + for key in ('contents', 'items'): + contents.extend(try_get(renderer, lambda x: x[key], list) or []) + + for content in contents: + if not isinstance(content, dict): + continue + continuation_ep = try_get( + content, (lambda x: x['continuationItemRenderer']['continuationEndpoint'], + lambda x: x['continuationItemRenderer']['button']['buttonRenderer']['command']), + dict) + continuation = cls._extract_continuation_ep_data(continuation_ep) + if continuation: + return continuation + + @classmethod + def _extract_alerts(cls, data): + for alert_dict in try_get(data, lambda x: x['alerts'], list) or []: + if not isinstance(alert_dict, dict): + continue + for alert in alert_dict.values(): + alert_type = alert.get('type') + if not alert_type: + continue + message = cls._get_text(alert.get('text')) + if message: + yield alert_type, message + + def _report_alerts(self, alerts, expected=True): + errors = [] + warnings = [] + for alert_type, alert_message in alerts: + if alert_type.lower() == 'error': + errors.append([alert_type, alert_message]) + else: + warnings.append([alert_type, alert_message]) + + for alert_type, alert_message in (warnings + errors[:-1]): + self.report_warning('YouTube said: %s - %s' % (alert_type, alert_message)) + if errors: + raise ExtractorError('YouTube said: %s' % errors[-1][1], expected=expected) + + def _extract_and_report_alerts(self, data, *args, **kwargs): + return self._report_alerts(self._extract_alerts(data), *args, **kwargs) + + 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) + if label: + badges.add(label.lower()) + 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 _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, + default_client='WEB'): + response = None + last_error = None + count = -1 + retries = self.get_param('extractor_retries', 3) + if check_get_keys is None: + check_get_keys = [] + while count < retries: + count += 1 + if last_error: + self.report_warning('%s. Retrying ...' % last_error) + try: + response = self._call_api( + ep=ep, fatal=True, headers=headers, + video_id=item_id, query=query, + context=self._extract_context(ytcfg, default_client), + api_key=self._extract_api_key(ytcfg, default_client), + 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): + # 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 + if fatal: + raise + else: + self.report_warning(error_to_compat_str(e)) + return + + else: + # Youtube may send alerts if there was an issue with the continuation page + try: + self._extract_and_report_alerts(response, expected=False) + except ExtractorError as e: + if fatal: + raise + self.report_warning(error_to_compat_str(e)) + return + if not check_get_keys or dict_get(response, check_get_keys): + break + # Youtube sometimes sends incomplete data + # See: https://github.com/ytdl-org/youtube-dl/issues/28194 + last_error = 'Incomplete data received' + if count >= retries: + if fatal: + raise ExtractorError(last_error) + else: + self.report_warning(last_error) + return + return response + @staticmethod def is_music_url(url): return re.match(r'https?://music\.youtube\.com/', url) is not None def _extract_video(self, renderer): video_id = renderer.get('videoId') - title = try_get( - renderer, - (lambda x: x['title']['runs'][0]['text'], - lambda x: x['title']['simpleText']), compat_str) - description = try_get( - renderer, lambda x: x['descriptionSnippet']['runs'][0]['text'], - compat_str) - duration = parse_duration(try_get( - renderer, lambda x: x['lengthText']['simpleText'], compat_str)) - view_count_text = try_get( - renderer, lambda x: x['viewCountText']['simpleText'], compat_str) or '' + 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 '' view_count = str_to_int(self._search_regex( r'^([\d,]+)', re.sub(r'\s', '', view_count_text), 'view count', default=None)) - uploader = try_get( - renderer, - (lambda x: x['ownerText']['runs'][0]['text'], - lambda x: x['shortBylineText']['runs'][0]['text']), compat_str) + + uploader = self._get_text(renderer, (lambda x: x['ownerText'], lambda x: x['shortBylineText'])) + return { '_type': 'url', 'ie_key': YoutubeIE.ie_key(), @@ -444,20 +761,15 @@ class YoutubeIE(YoutubeBaseInfoExtractor): # Invidious instances taken from https://github.com/iv-org/documentation/blob/master/Invidious-Instances.md r'(?:www\.)?invidious\.pussthecat\.org', r'(?:www\.)?invidious\.zee\.li', - r'(?:(?:www|au)\.)?ytprivate\.com', - r'(?:www\.)?invidious\.namazso\.eu', r'(?:www\.)?invidious\.ethibox\.fr', - r'(?:www\.)?w6ijuptxiku4xpnnaetxvnkc5vqcdu7mgns2u77qefoixi63vbvnpnqd\.onion', - r'(?:www\.)?kbjggqkzv65ivcqj6bumvp337z6264huv5kpkwuv6gu5yjiskvan7fad\.onion', r'(?:www\.)?invidious\.3o7z6yfxhbw7n3za4rss6l434kmv55cgw2vuziwuigpwegswvwzqipyd\.onion', - r'(?:www\.)?grwp24hodrefzvjjuccrkw3mjq4tzhaaq32amf33dzpmuxe7ilepcmad\.onion', # youtube-dl invidious instances list r'(?:(?:www|no)\.)?invidiou\.sh', r'(?:(?:www|fi)\.)?invidious\.snopyta\.org', r'(?:www\.)?invidious\.kabi\.tk', r'(?:www\.)?invidious\.mastodon\.host', r'(?:www\.)?invidious\.zapashcanon\.fr', - r'(?:www\.)?invidious\.kavin\.rocks', + r'(?:www\.)?(?:invidious(?:-us)?|piped)\.kavin\.rocks', r'(?:www\.)?invidious\.tinfoil-hat\.net', r'(?:www\.)?invidious\.himiko\.cloud', r'(?:www\.)?invidious\.reallyancient\.tech', @@ -484,6 +796,14 @@ class YoutubeIE(YoutubeBaseInfoExtractor): r'(?:www\.)?invidious\.toot\.koeln', r'(?:www\.)?invidious\.fdn\.fr', r'(?:www\.)?watch\.nettohikari\.com', + r'(?:www\.)?invidious\.namazso\.eu', + r'(?:www\.)?invidious\.silkky\.cloud', + r'(?:www\.)?invidious\.exonip\.de', + r'(?:www\.)?invidious\.riverside\.rocks', + r'(?:www\.)?invidious\.blamefran\.net', + r'(?:www\.)?invidious\.moomoo\.de', + r'(?:www\.)?ytb\.trom\.tf', + r'(?:www\.)?yt\.cyberhost\.uk', r'(?:www\.)?kgg2m7yk5aybusll\.onion', r'(?:www\.)?qklhadlycap4cnod\.onion', r'(?:www\.)?axqzx4s6s54s32yentfqojs3x5i7faxza6xo3ehd4bzzsg2ii4fv2iid\.onion', @@ -492,6 +812,10 @@ class YoutubeIE(YoutubeBaseInfoExtractor): r'(?:www\.)?invidious\.l4qlywnpwqsluw65ts7md3khrivpirse744un3x7mlskqauz5pyuzgqd\.onion', r'(?:www\.)?owxfohz4kjyv25fvlqilyxast7inivgiktls3th44jhk3ej3i7ya\.b32\.i2p', r'(?:www\.)?4l2dgddgsrkf2ous66i6seeyi6etzfgrue332grh2n7madpwopotugyd\.onion', + r'(?:www\.)?w6ijuptxiku4xpnnaetxvnkc5vqcdu7mgns2u77qefoixi63vbvnpnqd\.onion', + r'(?:www\.)?kbjggqkzv65ivcqj6bumvp337z6264huv5kpkwuv6gu5yjiskvan7fad\.onion', + r'(?:www\.)?grwp24hodrefzvjjuccrkw3mjq4tzhaaq32amf33dzpmuxe7ilepcmad\.onion', + r'(?:www\.)?hpniueoejy4opn7bc4ftgazyqjoeqwlvh2uiku2xqku6zpoa4bf5ruid\.onion', ) _VALID_URL = r"""(?x)^ ( @@ -640,6 +964,11 @@ class YoutubeIE(YoutubeBaseInfoExtractor): } _SUBTITLE_FORMATS = ('json3', 'srv1', 'srv2', 'srv3', 'ttml', 'vtt') + _AGE_GATE_REASONS = ( + 'Sign in to confirm your age', + 'This video may be inappropriate for some users.', + 'Sorry, this content is age-restricted.') + _GEO_BYPASS = False IE_NAME = 'youtube' @@ -1319,7 +1648,32 @@ class YoutubeIE(YoutubeBaseInfoExtractor): # multiple subtitles with same lang_code 'url': 'https://www.youtube.com/watch?v=wsQiKKfKxug', 'only_matching': True, + }, { + # Force use android client fallback + 'url': 'https://www.youtube.com/watch?v=YOelRv7fMxY', + 'info_dict': { + 'id': 'YOelRv7fMxY', + 'title': 'Digging a Secret Tunnel from my Workshop', + 'ext': '3gp', + 'upload_date': '20210624', + 'channel_id': 'UCp68_FLety0O-n9QU6phsgw', + 'uploader': 'colinfurze', + 'channel_url': r're:https?://(?:www\.)?youtube\.com/channel/UCp68_FLety0O-n9QU6phsgw', + 'description': 'md5:ecb672623246d98c6c562eed6ae798c3' + }, + 'params': { + 'format': '17', # 3gp format available on android + 'extractor_args': {'youtube': {'player_client': ['android']}}, + }, }, + { + # Skip download of additional client configs (remix client config in this case) + 'url': 'https://music.youtube.com/watch?v=MgNrAu2pzNs', + 'only_matching': True, + 'params': { + 'extractor_args': {'youtube': {'player_skip': ['configs']}}, + }, + } ] @classmethod @@ -1337,6 +1691,19 @@ def __init__(self, *args, **kwargs): self._code_cache = {} self._player_cache = {} + 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: + player_url = self._search_regex( + r'"(?:PLAYER_JS_URL|jsUrl)"\s*:\s*"([^"]+)"', + webpage, 'player URL', fatal=False) + if player_url.startswith('//'): + player_url = 'https:' + player_url + elif not re.match(r'https?://', player_url): + player_url = compat_urlparse.urljoin( + 'https://www.youtube.com', player_url) + return 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('.')) @@ -1351,6 +1718,15 @@ def _extract_player_info(cls, player_url): raise ExtractorError('Cannot identify player %r' % player_url) return id_m.group('id') + def _load_player(self, video_id, player_url, fatal=True) -> bool: + player_id = self._extract_player_info(player_url) + if player_id not in self._code_cache: + self._code_cache[player_id] = self._download_webpage( + player_url, video_id, fatal=fatal, + note='Downloading player ' + player_id, + errnote='Download of %s failed' % player_url) + return player_id in self._code_cache + def _extract_signature_function(self, video_id, player_url, example_sig): player_id = self._extract_player_info(player_url) @@ -1363,20 +1739,16 @@ def _extract_signature_function(self, video_id, player_url, example_sig): if cache_spec is not None: return lambda s: ''.join(s[i] for i in cache_spec) - if player_id not in self._code_cache: - self._code_cache[player_id] = self._download_webpage( - player_url, video_id, - note='Downloading player ' + player_id, - errnote='Download of %s failed' % player_url) - code = self._code_cache[player_id] - res = self._parse_sig_js(code) + if self._load_player(video_id, player_url): + code = self._code_cache[player_id] + res = self._parse_sig_js(code) - test_string = ''.join(map(compat_chr, range(len(example_sig)))) - cache_res = res(test_string) - cache_spec = [ord(c) for c in cache_res] + test_string = ''.join(map(compat_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) - return res + self._downloader.cache.store('youtube-sigfuncs', func_id, cache_spec) + return res def _print_sig_code(self, func, example_sig): def gen_sig_code(idxs): @@ -1447,11 +1819,6 @@ def _decrypt_signature(self, s, video_id, player_url): if player_url is None: raise ExtractorError('Cannot decrypt signature without player_url') - if player_url.startswith('//'): - player_url = 'https:' + player_url - elif not re.match(r'https?://', player_url): - player_url = compat_urlparse.urljoin( - 'https://www.youtube.com', player_url) try: player_id = (player_url, self._signature_cache_id(s)) if player_id not in self._player_cache: @@ -1468,6 +1835,31 @@ def _decrypt_signature(self, s, video_id, player_url): raise ExtractorError( 'Signature extraction failed: ' + tb, cause=e) + def _extract_signature_timestamp(self, video_id, player_url, ytcfg=None, fatal=False): + """ + Extract signatureTimestamp (sts) + Required to tell API what sig/player version is in use. + """ + sts = None + if isinstance(ytcfg, dict): + sts = int_or_none(ytcfg.get('STS')) + + if not sts: + # Attempt to extract from player + if player_url is None: + error_msg = 'Cannot extract signature timestamp without player_url.' + if fatal: + raise ExtractorError(error_msg) + self.report_warning(error_msg) + return + if self._load_player(video_id, player_url, fatal=fatal): + player_id = self._extract_player_info(player_url) + code = self._code_cache[player_id] + sts = int_or_none(self._search_regex( + r'(?:signatureTimestamp|sts)\s*:\s*(?P[0-9]{5})', code, + '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, @@ -1593,46 +1985,37 @@ def parse_time_text(time_text): if len(time_text_split) >= 3: return datetime_from_str('now-%s%s' % (time_text_split[0], time_text_split[1]), precision='auto') - @staticmethod - def _join_text_entries(runs): - text = None - for run in runs: - if not isinstance(run, dict): - continue - sub_text = try_get(run, lambda x: x['text'], compat_str) - if sub_text: - if not text: - text = sub_text - continue - text += sub_text - return text - def _extract_comment(self, comment_renderer, parent=None): comment_id = comment_renderer.get('commentId') if not comment_id: return - comment_text_runs = try_get(comment_renderer, lambda x: x['contentText']['runs']) or [] - text = self._join_text_entries(comment_text_runs) or '' - comment_time_text = try_get(comment_renderer, lambda x: x['publishedTimeText']['runs']) or [] - time_text = self._join_text_entries(comment_time_text) - timestamp = calendar.timegm(self.parse_time_text(time_text).timetuple()) - author = try_get(comment_renderer, lambda x: x['authorText']['simpleText'], compat_str) + + text = self._get_text(comment_renderer.get('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_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_id = try_get(comment_renderer, lambda x: x['authorEndpoint']['browseEndpoint']['browseId'], compat_str) - votes = str_to_int(try_get(comment_renderer, (lambda x: x['voteCount']['simpleText'], - lambda x: x['likeCount']), compat_str)) or 0 + + votes = parse_count(try_get(comment_renderer, (lambda x: x['voteCount']['simpleText'], + lambda x: x['likeCount']), compat_str)) or 0 author_thumbnail = try_get(comment_renderer, lambda x: x['authorThumbnail']['thumbnails'][-1]['url'], compat_str) author_is_uploader = try_get(comment_renderer, lambda x: x['authorIsChannelOwner'], bool) - is_liked = try_get(comment_renderer, lambda x: x['isLiked'], bool) + is_favorited = 'creatorHeart' in (try_get( + comment_renderer, lambda x: x['actionButtons']['commentActionButtonsRenderer'], dict) or {}) return { 'id': comment_id, 'text': text, 'timestamp': timestamp, 'time_text': time_text, 'like_count': votes, - 'is_favorited': is_liked, + 'is_favorited': is_favorited, 'author': author, 'author_id': author_id, 'author_thumbnail': author_thumbnail, @@ -1641,10 +2024,42 @@ def _extract_comment(self, comment_renderer, parent=None): } def _comment_entries(self, root_continuation_data, identity_token, account_syncid, - ytcfg, session_token_list, parent=None, comment_counts=None): + ytcfg, video_id, parent=None, comment_counts=None): - def extract_thread(parent_renderer): - contents = try_get(parent_renderer, lambda x: x['contents'], list) or [] + def extract_header(contents): + _total_comments = 0 + _continuation = None + 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)) + + if expected_comment_count: + comment_counts[1] = expected_comment_count + self.to_screen('Downloading ~%d comments' % expected_comment_count) + _total_comments = comment_counts[1] + sort_mode_str = self._configuration_arg('comment_sort', [''])[0] + comment_sort_index = int(sort_mode_str != 'top') # 1 = new, 0 = top + + sort_menu_item = try_get( + comments_header_renderer, + lambda x: x['sortMenu']['sortFilterSubMenuRenderer']['subMenuItems'][comment_sort_index], dict) or {} + sort_continuation_ep = sort_menu_item.get('serviceEndpoint') or {} + + _continuation = self._extract_continuation_ep_data(sort_continuation_ep) or self._extract_continuation(sort_menu_item) + if not _continuation: + continue + + sort_text = sort_menu_item.get('title') + if isinstance(sort_text, compat_str): + sort_text = sort_text.lower() + else: + sort_text = 'top comments' if comment_sort_index == 0 else 'newest first' + self.to_screen('Sorting comments by %s' % sort_text) + break + return _total_comments, _continuation + + def extract_thread(contents): if not parent: comment_counts[2] = 0 for content in contents: @@ -1668,107 +2083,48 @@ def extract_thread(parent_renderer): comment_counts[2] += 1 comment_entries_iter = self._comment_entries( comment_replies_renderer, identity_token, account_syncid, ytcfg, - parent=comment.get('id'), session_token_list=session_token_list, - comment_counts=comment_counts) + video_id, parent=comment.get('id'), comment_counts=comment_counts) for reply_comment in comment_entries_iter: yield reply_comment + # YouTube comments have a max depth of 2 + max_depth = int_or_none(self._configuration_arg('max_comment_depth', [''])[0]) or float('inf') + if max_depth == 1 and parent: + return if not comment_counts: # comment so far, est. total comments, current comment thread # comment_counts = [0, 0, 0] - # TODO: Generalize the download code with TabIE - context = self._extract_context(ytcfg) - visitor_data = try_get(context, lambda x: x['client']['visitorData'], compat_str) - continuation = YoutubeTabIE._extract_continuation(root_continuation_data) # TODO - first_continuation = False - if parent is None: - first_continuation = True + continuation = self._extract_continuation(root_continuation_data) + if continuation and len(continuation['continuation']) < 27: + self.write_debug('Detected old API continuation token. Generating new API compatible token.') + continuation_token = self._generate_comment_continuation(video_id) + continuation = self._build_api_continuation_query(continuation_token, None) + + visitor_data = None + is_first_continuation = parent is None for page_num in itertools.count(0): if not continuation: break headers = self._generate_api_headers(ytcfg, identity_token, account_syncid, visitor_data) - retries = self.get_param('extractor_retries', 3) - count = -1 - last_error = None - - while count < retries: - count += 1 - if last_error: - self.report_warning('%s. Retrying ...' % last_error) - try: - query = { - 'ctoken': continuation['ctoken'], - 'pbj': 1, - 'type': 'next', - } - if parent: - query['action_get_comment_replies'] = 1 - else: - query['action_get_comments'] = 1 - - comment_prog_str = '(%d/%d)' % (comment_counts[0], comment_counts[1]) - if page_num == 0: - if first_continuation: - note_prefix = 'Downloading initial comment continuation page' - else: - note_prefix = ' Downloading comment reply thread %d %s' % (comment_counts[2], comment_prog_str) - else: - note_prefix = '%sDownloading comment%s page %d %s' % ( - ' ' if parent else '', - ' replies' if parent else '', - page_num, - comment_prog_str) - - browse = self._download_json( - 'https://www.youtube.com/comment_service_ajax', None, - '%s %s' % (note_prefix, '(retry #%d)' % count if count else ''), - headers=headers, query=query, - data=urlencode_postdata({ - 'session_token': session_token_list[0] - })) - except ExtractorError as e: - if isinstance(e.cause, compat_HTTPError) and e.cause.code in (500, 503, 404, 413): - if e.cause.code == 413: - self.report_warning('Assumed end of comments (received HTTP Error 413)') - return - # 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 e.cause.code == 404: - last_error = last_error + ' (this API is probably deprecated)' - if count < retries: - continue - raise - else: - session_token = try_get(browse, lambda x: x['xsrf_token'], compat_str) - if session_token: - session_token_list[0] = session_token - - response = try_get(browse, - (lambda x: x['response'], - lambda x: x[1]['response'])) or {} - - if response.get('continuationContents'): - break - - # YouTube sometimes gives reload: now json if something went wrong (e.g. bad auth) - if browse.get('reload'): - raise ExtractorError('Invalid or missing params in continuation request', expected=False) - - # TODO: not tested, merged from old extractor - err_msg = browse.get('externalErrorMessage') - if err_msg: - raise ExtractorError('YouTube said: %s' % err_msg, expected=False) - - # Youtube sometimes sends incomplete data - # See: https://github.com/ytdl-org/youtube-dl/issues/28194 - last_error = 'Incomplete data received' - if count >= retries: - raise ExtractorError(last_error) + comment_prog_str = '(%d/%d)' % (comment_counts[0], comment_counts[1]) + if page_num == 0: + if is_first_continuation: + note_prefix = 'Downloading comment section API JSON' + else: + note_prefix = ' Downloading comment API JSON reply thread %d %s' % ( + comment_counts[2], comment_prog_str) + else: + note_prefix = '%sDownloading comment%s API JSON page %d %s' % ( + ' ' if parent else '', ' replies' if parent else '', + page_num, comment_prog_str) + response = self._extract_response( + item_id=None, query=continuation, + ep='next', ytcfg=ytcfg, headers=headers, note=note_prefix, + check_get_keys=('onResponseReceivedEndpoints', 'continuationContents')) if not response: break visitor_data = try_get( @@ -1776,86 +2132,146 @@ def extract_thread(parent_renderer): lambda x: x['responseContext']['webResponseContextExtensionData']['ytConfigData']['visitorData'], compat_str) or visitor_data - known_continuation_renderers = { - 'itemSectionContinuation': extract_thread, - 'commentRepliesContinuation': extract_thread - } - - # extract next root continuation from the results - continuation_contents = try_get( - response, lambda x: x['continuationContents'], dict) or {} - - for key, value in continuation_contents.items(): - if key not in known_continuation_renderers: - continue - continuation_renderer = value + continuation_contents = dict_get(response, ('onResponseReceivedEndpoints', 'continuationContents')) - if first_continuation: - first_continuation = False - expected_comment_count = try_get( - continuation_renderer, - (lambda x: x['header']['commentsHeaderRenderer']['countText']['runs'][0]['text'], - lambda x: x['header']['commentsHeaderRenderer']['commentsCount']['runs'][0]['text']), - compat_str) - - if expected_comment_count: - comment_counts[1] = str_to_int(expected_comment_count) - self.to_screen('Downloading ~%d comments' % str_to_int(expected_comment_count)) - yield comment_counts[1] - - # TODO: cli arg. - # 1/True for newest, 0/False for popular (default) - comment_sort_index = int(True) - sort_continuation_renderer = try_get( - continuation_renderer, - lambda x: x['header']['commentsHeaderRenderer']['sortMenu']['sortFilterSubMenuRenderer']['subMenuItems'] - [comment_sort_index]['continuation']['reloadContinuationData'], dict) - # If this fails, the initial continuation page - # starts off with popular anyways. - if sort_continuation_renderer: - continuation = YoutubeTabIE._build_continuation_query( - continuation=sort_continuation_renderer.get('continuation'), - ctp=sort_continuation_renderer.get('clickTrackingParams')) - self.to_screen('Sorting comments by %s' % ('popular' if comment_sort_index == 0 else 'newest')) + continuation = None + if isinstance(continuation_contents, list): + for continuation_section in continuation_contents: + if not isinstance(continuation_section, dict): + continue + continuation_items = try_get( + continuation_section, + (lambda x: x['reloadContinuationItemsCommand']['continuationItems'], + lambda x: x['appendContinuationItemsAction']['continuationItems']), + list) or [] + if is_first_continuation: + total_comments, continuation = extract_header(continuation_items) + if total_comments: + yield total_comments + is_first_continuation = False + if continuation: + break + continue + count = 0 + for count, entry in enumerate(extract_thread(continuation_items)): + yield entry + continuation = self._extract_continuation({'contents': continuation_items}) + if continuation: + # Sometimes YouTube provides a continuation without any comments + # In most cases we end up just downloading these with very little comments to come. + if count == 0: + if not parent: + self.report_warning('No comments received - assuming end of comments') + continuation = None break - for entry in known_continuation_renderers[key](continuation_renderer): - yield entry + # Deprecated response structure + elif isinstance(continuation_contents, dict): + known_continuation_renderers = ('itemSectionContinuation', 'commentRepliesContinuation') + for key, continuation_renderer in continuation_contents.items(): + if key not in known_continuation_renderers: + continue + if not isinstance(continuation_renderer, dict): + continue + if is_first_continuation: + header_continuation_items = [continuation_renderer.get('header') or {}] + total_comments, continuation = extract_header(header_continuation_items) + if total_comments: + yield total_comments + is_first_continuation = False + if continuation: + break - continuation = YoutubeTabIE._extract_continuation(continuation_renderer) # TODO - break + # Sometimes YouTube provides a continuation without any comments + # In most cases we end up just downloading these with very little comments to come. + count = 0 + for count, entry in enumerate(extract_thread(continuation_renderer.get('contents') or {})): + yield entry + continuation = self._extract_continuation(continuation_renderer) + if count == 0: + if not parent: + self.report_warning('No comments received - assuming end of comments') + continuation = None + break - def _extract_comments(self, ytcfg, video_id, contents, webpage, xsrf_token): + @staticmethod + def _generate_comment_continuation(video_id): + """ + Generates initial comment section continuation token from given video id + """ + b64_vid_id = base64.b64encode(bytes(video_id.encode('utf-8'))) + parts = ('Eg0SCw==', b64_vid_id, 'GAYyJyIRIgs=', b64_vid_id, 'MAB4AjAAQhBjb21tZW50cy1zZWN0aW9u') + new_continuation_intlist = list(itertools.chain.from_iterable( + [bytes_to_intlist(base64.b64decode(part)) for part in parts])) + return base64.b64encode(intlist_to_bytes(new_continuation_intlist)).decode('utf-8') + + def _extract_comments(self, ytcfg, video_id, contents, webpage): """Entry for comment extraction""" + def _real_comment_extract(contents): + if isinstance(contents, list): + for entry in contents: + for key, renderer in entry.items(): + if key not in known_entry_comment_renderers: + continue + yield from self._comment_entries( + renderer, video_id=video_id, ytcfg=ytcfg, + identity_token=self._extract_identity_token(webpage, item_id=video_id), + account_syncid=self._extract_account_syncid(ytcfg)) + break comments = [] - known_entry_comment_renderers = ( - 'itemSectionRenderer', - ) + known_entry_comment_renderers = ('itemSectionRenderer',) estimated_total = 0 - for entry in contents: - for key, renderer in entry.items(): - if key not in known_entry_comment_renderers: - continue + max_comments = int_or_none(self._configuration_arg('max_comments', [''])[0]) or float('inf') - comment_iter = self._comment_entries( - renderer, - identity_token=self._extract_identity_token(webpage, item_id=video_id), - account_syncid=self._extract_account_syncid(ytcfg), - ytcfg=ytcfg, - session_token_list=[xsrf_token]) - - for comment in comment_iter: - if isinstance(comment, int): - estimated_total = comment - continue - comments.append(comment) - break + try: + for comment in _real_comment_extract(contents): + if len(comments) >= max_comments: + break + if isinstance(comment, int): + estimated_total = comment + continue + comments.append(comment) + except KeyboardInterrupt: + self.to_screen('Interrupted by user') self.to_screen('Downloaded %d/%d comments' % (len(comments), estimated_total)) return { 'comments': comments, 'comment_count': len(comments), } + @staticmethod + def _generate_player_context(sts=None): + context = { + 'html5Preference': 'HTML5_PREF_WANTS', + } + if sts is not None: + context['signatureTimestamp'] = sts + 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' + } + 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) @@ -1867,61 +2283,134 @@ def _real_extract(self, url): webpage = self._download_webpage( webpage_url + '&bpctr=9999999999&has_verified=1', video_id, fatal=False) - def get_text(x): - if not x: - return - text = x.get('simpleText') - if text and isinstance(text, compat_str): - return text - runs = x.get('runs') - if not isinstance(runs, list): - return - return ''.join([r['text'] for r in runs if isinstance(r.get('text'), compat_str)]) - - ytm_streaming_data = {} - if is_music_url: - # we are forcing to use parse_json because 141 only appeared in get_video_info. - # el, c, cver, cplayer field required for 141(aac 256kbps) codec - # maybe paramter of youtube music player? - ytm_player_response = self._parse_json(try_get(compat_parse_qs( - self._download_webpage( - base_url + 'get_video_info', video_id, - 'Fetching youtube music info webpage', - 'unable to download youtube music info webpage', query={ - 'video_id': video_id, - 'eurl': 'https://youtube.googleapis.com/v/' + video_id, - 'el': 'detailpage', - 'c': 'WEB_REMIX', - 'cver': '0.1', - 'cplayer': 'UNIPLAYER' - }, fatal=False)), - lambda x: x['player_response'][0], - compat_str) or '{}', video_id) - ytm_streaming_data = ytm_player_response.get('streamingData') 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 if webpage: player_response = self._extract_yt_initial_variable( webpage, self._YT_INITIAL_PLAYER_RESPONSE_RE, video_id, 'initial player response') - ytcfg = self._extract_ytcfg(video_id, webpage) - if not player_response: - player_response = self._call_api( - 'player', {'videoId': video_id}, video_id, api_key=self._extract_api_key(ytcfg)) + 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') == 'Sign in to confirm your age': - pr = self._parse_json(try_get(compat_parse_qs( - self._download_webpage( - base_url + 'get_video_info', video_id, - 'Refetching age-gated info webpage', - 'unable to download video info webpage', query={ - 'video_id': video_id, - 'eurl': 'https://youtube.googleapis.com/v/' + video_id, - }, fatal=False)), - lambda x: x['player_response'][0], - compat_str) or '{}', video_id) + 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 {} + if pr: player_response = pr @@ -1943,7 +2432,7 @@ def get_text(x): lambda x: x['microformat']['playerMicroformatRenderer'], dict) or {} video_title = video_details.get('title') \ - or get_text(microformat.get('title')) \ + or self._get_text(microformat.get('title')) \ or search_meta(['og:title', 'twitter:title', 'title']) video_description = video_details.get('shortDescription') @@ -1993,8 +2482,12 @@ def feed_entry(name): formats, itags, stream_ids = [], [], [] itag_qualities = {} - player_url = None - q = qualities(['tiny', 'small', 'medium', 'large', 'hd720', 'hd1080', 'hd1440', 'hd2160', 'hd2880', 'highres']) + 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 + 'small', 'medium', 'large', 'hd720', 'hd1080', 'hd1440', 'hd2160', 'hd2880', 'highres' + ]) streaming_data = player_response.get('streamingData') or {} streaming_formats = streaming_data.get('formats') or [] @@ -2013,6 +2506,8 @@ def feed_entry(name): continue quality = fmt.get('quality') + if quality == 'tiny' or not quality: + quality = fmt.get('audioQuality', '').lower() or quality if itag and quality: itag_qualities[itag] = quality # FORMAT_STREAM_TYPE_OTF(otf=1) requires downloading the init fragment @@ -2028,12 +2523,6 @@ def feed_entry(name): encrypted_sig = try_get(sc, lambda x: x['s'][0]) if not (sc and fmt_url and encrypted_sig): continue - if not player_url: - if not webpage: - continue - player_url = self._search_regex( - r'"(?:PLAYER_JS_URL|jsUrl)"\s*:\s*"([^"]+)"', - webpage, 'player URL', fatal=False) if not player_url: continue signature = self._decrypt_signature(sc['s'][0], video_id, player_url) @@ -2059,13 +2548,15 @@ def feed_entry(name): 'width': fmt.get('width'), 'language': audio_track.get('id', '').split('.')[0], } - mimetype = fmt.get('mimeType') - if mimetype: - mobj = re.match( - r'((?:[^/]+)/(?:[^;]+))(?:;\s*codecs="([^"]+)")?', mimetype) - if mobj: - dct['ext'] = mimetype2ext(mobj.group(1)) - dct.update(parse_codecs(mobj.group(2))) + mime_mobj = re.match( + r'((?:[^/]+)/(?:[^;]+))(?:;\s*codecs="([^"]+)")?', fmt.get('mimeType') or '') + 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: @@ -2081,8 +2572,12 @@ def feed_entry(name): dct['container'] = dct['ext'] + '_dash' formats.append(dct) + skip_manifests = self._configuration_arg('skip') + get_dash = '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): - hls_manifest_url = sd.get('hlsManifestUrl') + 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): @@ -2090,28 +2585,23 @@ def feed_entry(name): r'/itag/(\d+)', f['url'], 'itag', default=None) if itag: f['format_id'] = itag - formats.append(f) - - if self.get_param('youtube_include_dash_manifest', True): - for sd in (streaming_data, ytm_streaming_data): - dash_manifest_url = sd.get('dashManifestUrl') - if dash_manifest_url: - 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: - # Not actually usefull since the sorting is already done with "quality,res,fps,codec" - # but kept to maintain feature parity (and code similarity) with youtube-dl - # Remove if this causes any issues with sorting in future - f['quality'] = q(itag_qualities[itag]) - 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) + formats.append(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): + itag = f['format_id'] + if itag in itags: + continue + if itag in itag_qualities: + f['quality'] = q(itag_qualities[itag]) + 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) if not formats: if not self.get_param('allow_unplayable_formats') and streaming_data.get('licenseInfos'): @@ -2121,10 +2611,10 @@ def feed_entry(name): playability_status, lambda x: x['errorScreen']['playerErrorMessageRenderer'], dict) or {} - reason = get_text(pemr.get('reason')) or playability_status.get('reason') + reason = self._get_text(pemr.get('reason')) or playability_status.get('reason') subreason = pemr.get('subreason') if subreason: - subreason = clean_html(get_text(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') if not countries: @@ -2155,7 +2645,21 @@ def feed_entry(name): 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, @@ -2172,33 +2676,25 @@ def feed_entry(name): 'url': thumbnail_url, 'height': int_or_none(thumbnail.get('height')), 'width': int_or_none(thumbnail.get('width')), - 'preference': 1 if 'maxresdefault' in thumbnail_url else -1 }) thumbnail_url = search_meta(['og:image', 'twitter:image']) if thumbnail_url: thumbnails.append({ 'url': thumbnail_url, - 'preference': 1 if 'maxresdefault' in thumbnail_url else -1 }) - # All videos have a maxresdefault thumbnail, but sometimes it does not appear in the webpage - # See: https://github.com/ytdl-org/youtube-dl/issues/29049 - thumbnails.append({ - 'url': 'https://i.ytimg.com/vi/%s/maxresdefault.jpg' % video_id, - 'preference': 1, - }) + # 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 + 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')) + for thumb in thumbnails: + i = next((i for i, t in enumerate(thumbnail_types) if f'/{video_id}/{t}' in thumb['url']), 20) + thumb['preference'] = (0 if '.webp' in thumb['url'] else -1) - (2 * i) self._remove_duplicate_formats(thumbnails) - 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') - owner_profile_url = microformat.get('ownerProfileUrl') - info = { 'id': video_id, 'title': self._live_title(video_title) if is_live else video_title, @@ -2260,7 +2756,7 @@ def process_language(container, base_url, lang_code, sub_name, query): continue process_language( subtitles, base_url, lang_code, - try_get(caption_track, lambda x: x.get('name').get('simpleText')), + try_get(caption_track, lambda x: x['name']['simpleText']), {}) continue automatic_captions = {} @@ -2270,7 +2766,7 @@ def process_language(container, base_url, lang_code, sub_name, query): continue process_language( automatic_captions, base_url, translation_language_code, - try_get(translation_language, lambda x: x['languageName']['simpleText']), + self._get_text(translation_language.get('languageName'), max_runs=1), {'tlang': translation_language_code}) info['automatic_captions'] = automatic_captions info['subtitles'] = subtitles @@ -2308,21 +2804,22 @@ def process_language(container, base_url, lang_code, sub_name, query): webpage, self._YT_INITIAL_DATA_RE, video_id, 'yt initial data') if not initial_data: - initial_data = self._call_api( - 'next', {'videoId': video_id}, video_id, fatal=False, api_key=self._extract_api_key(ytcfg)) + initial_data = self._extract_response( + item_id=video_id, ep='next', fatal=False, + ytcfg=ytcfg, headers=headers, query={'videoId': video_id}, + note='Downloading initial data API JSON') - if not is_live: - try: - # This will error if there is no livechat - initial_data['contents']['twoColumnWatchNextResults']['conversationBar']['liveChatRenderer']['continuations'][0]['reloadContinuationData']['continuation'] - info['subtitles']['live_chat'] = [{ - 'url': 'https://www.youtube.com/watch?v=%s' % video_id, # url is needed to set cookies - 'video_id': video_id, - 'ext': 'json', - 'protocol': 'youtube_live_chat_replay', - }] - except (KeyError, IndexError, TypeError): - pass + try: + # This will error if there is no livechat + initial_data['contents']['twoColumnWatchNextResults']['conversationBar']['liveChatRenderer']['continuations'][0]['reloadContinuationData']['continuation'] + info['subtitles']['live_chat'] = [{ + 'url': 'https://www.youtube.com/watch?v=%s' % video_id, # url is needed to set cookies + 'video_id': video_id, + 'ext': 'json', + 'protocol': 'youtube_live_chat' if is_live or is_upcoming else 'youtube_live_chat_replay', + }] + except (KeyError, IndexError, TypeError): + pass if initial_data: chapters = self._extract_chapters_from_json( @@ -2337,7 +2834,7 @@ def process_language(container, base_url, lang_code, sub_name, query): def chapter_time(mmlir): return parse_duration( - get_text(mmlir.get('timeDescription'))) + self._get_text(mmlir.get('timeDescription'))) chapters = [] for next_num, content in enumerate(contents, start=1): @@ -2351,7 +2848,7 @@ def chapter_time(mmlir): chapters.append({ 'start_time': start_time, 'end_time': end_time, - 'title': get_text(mmlir.get('title')), + 'title': self._get_text(mmlir.get('title')), }) if chapters: break @@ -2367,7 +2864,7 @@ def chapter_time(mmlir): if vpir: stl = vpir.get('superTitleLink') if stl: - stl = get_text(stl) + stl = self._get_text(stl) if try_get( vpir, lambda x: x['superTitleIcon']['iconType']) == 'LOCATION_PIN': @@ -2407,7 +2904,7 @@ def chapter_time(mmlir): }) vsir = content.get('videoSecondaryInfoRenderer') if vsir: - info['channel'] = get_text(try_get( + info['channel'] = self._get_text(try_get( vsir, lambda x: x['owner']['videoOwnerRenderer']['title'], dict)) @@ -2425,8 +2922,8 @@ def chapter_time(mmlir): mrr_title = mrr.get('title') if not mrr_title: continue - mrr_title = get_text(mrr['title']) - mrr_contents_text = 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: @@ -2458,21 +2955,20 @@ def chapter_time(mmlir): if initial_data and is_private is not None: is_membersonly = False is_premium = False - contents = try_get(initial_data, lambda x: x['contents']['twoColumnWatchNextResults']['results']['results']['contents'], list) - for content in contents or []: - badges = try_get(content, lambda x: x['videoPrimaryInfoRenderer']['badges'], list) - for badge in badges or []: - label = try_get(badge, lambda x: x['metadataBadgeRenderer']['label']) or '' - if label.lower() == 'members only': - is_membersonly = True - break - elif label.lower() == 'premium': - is_premium = True - break - if is_membersonly or is_premium: - break + contents = try_get(initial_data, lambda x: x['contents']['twoColumnWatchNextResults']['results']['results']['contents'], list) or [] + badge_labels = set() + for content in contents: + if not isinstance(content, dict): + continue + badge_labels.update(self._extract_badges(content.get('videoPrimaryInfoRenderer'))) + for badge_label in badge_labels: + if badge_label.lower() == 'members only': + is_membersonly = True + elif badge_label.lower() == 'premium': + is_premium = True + elif badge_label.lower() == 'unlisted': + is_unlisted = True - # TODO: Add this for playlists info['availability'] = self._availability( is_private=is_private, needs_premium=is_premium, @@ -2513,7 +3009,7 @@ def chapter_time(mmlir): data=urlencode_postdata({xsrf_field_name: xsrf_token})) if get_comments: - info['__post_extractor'] = lambda: self._extract_comments(ytcfg, video_id, contents, webpage, xsrf_token) + info['__post_extractor'] = lambda: self._extract_comments(ytcfg, video_id, contents, webpage) self.mark_watched(video_id, player_response) @@ -2840,6 +3336,10 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor): }, { 'url': 'https://www.youtube.com/c/CommanderVideoHq/live', 'only_matching': True, + }, { + 'note': 'A channel that is not live. Should raise error', + 'url': 'https://www.youtube.com/user/numberphile/live', + 'only_matching': True, }, { 'url': 'https://www.youtube.com/feed/trending', 'only_matching': True, @@ -2942,6 +3442,17 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor): 'title': 'Album - Royalty Free Music Library V2 (50 Songs)', }, 'playlist_count': 50, + }, { + 'note': 'unlisted single video playlist', + 'url': 'https://www.youtube.com/playlist?list=PLwL24UFy54GrB3s2KMMfjZscDi1x5Dajf', + 'info_dict': { + 'uploader_id': 'UC9zHu_mHU96r19o-wV5Qs1Q', + 'uploader': 'colethedj', + 'id': 'PLwL24UFy54GrB3s2KMMfjZscDi1x5Dajf', + 'title': 'yt-dlp unlisted playlist test', + 'availability': 'unlisted' + }, + 'playlist_count': 1, }] @classmethod @@ -2983,9 +3494,8 @@ def _grid_entries(self, grid_renderer): renderer = self._extract_basic_item_renderer(item) if not isinstance(renderer, dict): continue - title = try_get( - renderer, (lambda x: x['title']['runs'][0]['text'], - lambda x: x['title']['simpleText']), compat_str) + title = self._get_text(renderer.get('title')) + # playlist playlist_id = renderer.get('playlistId') if playlist_id: @@ -3002,8 +3512,6 @@ def _grid_entries(self, grid_renderer): # channel channel_id = renderer.get('channelId') if channel_id: - title = try_get( - renderer, lambda x: x['title']['simpleText'], compat_str) yield self.url_result( 'https://www.youtube.com/channel/%s' % channel_id, ie=YoutubeTabIE.ie_key(), video_title=title) @@ -3046,8 +3554,7 @@ def _shelf_entries(self, shelf_renderer, skip_channels=False): # will not work if skip_channels and '/channels?' in shelf_url: return - title = try_get( - shelf_renderer, lambda x: x['title']['runs'][0]['text'], compat_str) + title = self._get_text(shelf_renderer, lambda x: x['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): @@ -3134,52 +3641,6 @@ def _rich_grid_entries(self, contents): if entry: yield entry ''' - - @staticmethod - def _build_continuation_query(continuation, ctp=None): - query = { - 'ctoken': continuation, - 'continuation': continuation, - } - if ctp: - query['itct'] = ctp - return query - - @staticmethod - def _extract_next_continuation_data(renderer): - next_continuation = try_get( - renderer, lambda x: x['continuations'][0]['nextContinuationData'], dict) - if not next_continuation: - return - continuation = next_continuation.get('continuation') - if not continuation: - return - ctp = next_continuation.get('clickTrackingParams') - return YoutubeTabIE._build_continuation_query(continuation, ctp) - - @classmethod - def _extract_continuation(cls, renderer): - next_continuation = cls._extract_next_continuation_data(renderer) - if next_continuation: - return next_continuation - contents = [] - for key in ('contents', 'items'): - contents.extend(try_get(renderer, lambda x: x[key], list) or []) - for content in contents: - if not isinstance(content, dict): - continue - continuation_ep = try_get( - content, lambda x: x['continuationItemRenderer']['continuationEndpoint'], - dict) - if not continuation_ep: - continue - continuation = try_get( - continuation_ep, lambda x: x['continuationCommand']['token'], compat_str) - if not continuation: - continue - ctp = continuation_ep.get('clickTrackingParams') - return YoutubeTabIE._build_continuation_query(continuation, ctp) - def _entries(self, tab, item_id, identity_token, account_syncid, ytcfg): def extract_entries(parent_renderer): # this needs to called again for continuation to work with feeds @@ -3232,20 +3693,15 @@ def extract_entries(parent_renderer): # this needs to called again for continua for entry in extract_entries(parent_renderer): yield entry continuation = continuation_list[0] - context = self._extract_context(ytcfg) - visitor_data = try_get(context, lambda x: x['client']['visitorData'], compat_str) + visitor_data = None for page_num in itertools.count(1): if not continuation: break - query = { - 'continuation': continuation['continuation'], - 'clickTracking': {'clickTrackingParams': continuation['itct']} - } 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=query, headers=headers, ytcfg=ytcfg, + query=continuation, headers=headers, ytcfg=ytcfg, check_get_keys=('continuationContents', 'onResponseReceivedActions', 'onResponseReceivedEndpoints')) if not response: @@ -3309,27 +3765,19 @@ def _extract_selected_tab(tabs): else: raise ExtractorError('Unable to find selected tab') - @staticmethod - def _extract_uploader(data): + @classmethod + def _extract_uploader(cls, data): uploader = {} - sidebar_renderer = try_get( - data, lambda x: x['sidebar']['playlistSidebarRenderer']['items'], list) - if sidebar_renderer: - for item in sidebar_renderer: - if not isinstance(item, dict): - continue - renderer = item.get('playlistSidebarSecondaryInfoRenderer') - if not isinstance(renderer, dict): - continue - owner = try_get( - renderer, lambda x: x['videoOwner']['videoOwnerRenderer']['title']['runs'][0], dict) - if owner: - uploader['uploader'] = owner.get('text') - uploader['uploader_id'] = try_get( - owner, lambda x: x['navigationEndpoint']['browseEndpoint']['browseId'], compat_str) - uploader['uploader_url'] = urljoin( - 'https://www.youtube.com/', - try_get(owner, lambda x: x['navigationEndpoint']['browseEndpoint']['canonicalBaseUrl'], compat_str)) + renderer = cls._extract_sidebar_info_renderer(data, 'playlistSidebarSecondaryInfoRenderer') or {} + owner = try_get( + renderer, lambda x: x['videoOwner']['videoOwnerRenderer']['title']['runs'][0], dict) + if owner: + uploader['uploader'] = owner.get('text') + uploader['uploader_id'] = try_get( + owner, lambda x: x['navigationEndpoint']['browseEndpoint']['browseId'], compat_str) + uploader['uploader_url'] = urljoin( + 'https://www.youtube.com/', + try_get(owner, lambda x: x['navigationEndpoint']['browseEndpoint']['canonicalBaseUrl'], compat_str)) return {k: v for k, v in uploader.items() if v is not None} def _extract_from_tabs(self, item_id, webpage, data, tabs): @@ -3355,8 +3803,8 @@ def _extract_from_tabs(self, item_id, webpage, data, tabs): thumbnails_list = ( try_get(renderer, lambda x: x['avatar']['thumbnails'], list) or try_get( - data, - lambda x: x['sidebar']['playlistSidebarRenderer']['items'][0]['playlistSidebarPrimaryInfoRenderer']['thumbnailRenderer']['playlistVideoThumbnailRenderer']['thumbnail']['thumbnails'], + self._extract_sidebar_info_renderer(data, 'playlistSidebarPrimaryInfoRenderer'), + lambda x: x['thumbnailRenderer']['playlistVideoThumbnailRenderer']['thumbnail']['thumbnails'], list) or []) @@ -3380,7 +3828,6 @@ def _extract_from_tabs(self, item_id, webpage, data, tabs): or playlist_id) title += format_field(selected_tab, 'title', ' - %s') title += format_field(selected_tab, 'expandedText', ' - %s') - metadata = { 'playlist_id': playlist_id, 'playlist_title': title, @@ -3391,27 +3838,29 @@ def _extract_from_tabs(self, item_id, webpage, data, tabs): 'thumbnails': thumbnails, 'tags': tags, } + availability = self._extract_availability(data) + if availability: + metadata['availability'] = availability if not channel_id: metadata.update(self._extract_uploader(data)) metadata.update({ 'channel': metadata['uploader'], 'channel_id': metadata['uploader_id'], 'channel_url': metadata['uploader_url']}) + ytcfg = self._extract_ytcfg(item_id, webpage) return self.playlist_result( self._entries( selected_tab, playlist_id, self._extract_identity_token(webpage, item_id), - self._extract_account_syncid(data), - self._extract_ytcfg(item_id, webpage)), + self._extract_account_syncid(ytcfg, data), ytcfg), **metadata) 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, account_syncid=self._extract_account_syncid(data), - identity_token=self._extract_identity_token(webpage, item_id=playlist_id), - visitor_data=try_get(self._extract_context(ytcfg), lambda x: x['client']['visitorData'], compat_str)) + 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): videos = list(self._playlist_entries(playlist)) if not videos: @@ -3436,9 +3885,7 @@ def _extract_mix_playlist(self, playlist, playlist_id, data, webpage): } response = self._extract_response( item_id='%s page %d' % (playlist_id, page_num), - query=query, - ep='next', - headers=headers, + query=query, ep='next', headers=headers, ytcfg=ytcfg, check_get_keys='contents' ) playlist = try_get( @@ -3462,131 +3909,86 @@ def _extract_from_playlist(self, item_id, url, data, playlist, webpage): self._extract_mix_playlist(playlist, playlist_id, data, webpage), playlist_id=playlist_id, playlist_title=title) - @staticmethod - def _extract_alerts(data): - for alert_dict in try_get(data, lambda x: x['alerts'], list) or []: - if not isinstance(alert_dict, dict): + def _extract_availability(self, data): + """ + Gets the availability of a given playlist/tab. + Note: Unless YouTube tells us explicitly, we do not assume it is public + @param data: response + """ + is_private = is_unlisted = None + renderer = self._extract_sidebar_info_renderer(data, 'playlistSidebarPrimaryInfoRenderer') or {} + badge_labels = self._extract_badges(renderer) + + # Personal playlists, when authenticated, have a dropdown visibility selector instead of a badge + privacy_dropdown_entries = try_get( + renderer, lambda x: x['privacyForm']['dropdownFormFieldRenderer']['dropdown']['dropdownRenderer']['entries'], list) or [] + for renderer_dict in privacy_dropdown_entries: + is_selected = try_get( + renderer_dict, lambda x: x['privacyDropdownItemRenderer']['isSelected'], bool) or False + if not is_selected: continue - for alert in alert_dict.values(): - alert_type = alert.get('type') - if not alert_type: - continue - message = try_get(alert, lambda x: x['text']['simpleText'], compat_str) or '' - if message: - yield alert_type, message - for run in try_get(alert, lambda x: x['text']['runs'], list) or []: - message += try_get(run, lambda x: x['text'], compat_str) - if message: - yield alert_type, message - - def _report_alerts(self, alerts, expected=True): - errors = [] - warnings = [] - for alert_type, alert_message in alerts: - if alert_type.lower() == 'error': - errors.append([alert_type, alert_message]) - else: - warnings.append([alert_type, alert_message]) + label = self._get_text( + try_get(renderer_dict, lambda x: x['privacyDropdownItemRenderer']['label'], dict) or []) + if label: + badge_labels.add(label.lower()) + break - for alert_type, alert_message in (warnings + errors[:-1]): - self.report_warning('YouTube said: %s - %s' % (alert_type, alert_message)) - if errors: - raise ExtractorError('YouTube said: %s' % errors[-1][1], expected=expected) + for badge_label in badge_labels: + if badge_label == 'unlisted': + is_unlisted = True + elif badge_label == 'private': + is_private = True + elif badge_label == 'public': + is_unlisted = is_private = False + return self._availability(is_private, False, False, False, is_unlisted) - def _extract_and_report_alerts(self, data, *args, **kwargs): - return self._report_alerts(self._extract_alerts(data), *args, **kwargs) + @staticmethod + def _extract_sidebar_info_renderer(data, info_renderer, expected_type=dict): + sidebar_renderer = try_get( + data, lambda x: x['sidebar']['playlistSidebarRenderer']['items'], list) or [] + for item in sidebar_renderer: + renderer = try_get(item, lambda x: x[info_renderer], expected_type) + if renderer: + return renderer def _reload_with_unavailable_videos(self, item_id, data, webpage): """ Get playlist with unavailable videos if the 'show unavailable videos' button exists. """ - sidebar_renderer = try_get( - data, lambda x: x['sidebar']['playlistSidebarRenderer']['items'], list) - if not sidebar_renderer: - return browse_id = params = None - for item in sidebar_renderer: - if not isinstance(item, dict): + renderer = self._extract_sidebar_info_renderer(data, 'playlistSidebarPrimaryInfoRenderer') + if not renderer: + return + menu_renderer = try_get( + renderer, lambda x: x['menu']['menuRenderer']['items'], list) or [] + for menu_item in menu_renderer: + if not isinstance(menu_item, dict): continue - renderer = item.get('playlistSidebarPrimaryInfoRenderer') - menu_renderer = try_get( - renderer, lambda x: x['menu']['menuRenderer']['items'], list) or [] - for menu_item in menu_renderer: - if not isinstance(menu_item, dict): - continue - nav_item_renderer = menu_item.get('menuNavigationItemRenderer') - text = try_get( - nav_item_renderer, lambda x: x['text']['simpleText'], compat_str) - if not text or text.lower() != 'show unavailable videos': - continue - browse_endpoint = try_get( - nav_item_renderer, lambda x: x['navigationEndpoint']['browseEndpoint'], dict) or {} - browse_id = browse_endpoint.get('browseId') - params = browse_endpoint.get('params') - break - - ytcfg = self._extract_ytcfg(item_id, webpage) - headers = self._generate_api_headers( - ytcfg, account_syncid=self._extract_account_syncid(ytcfg), - identity_token=self._extract_identity_token(webpage, item_id=item_id), - visitor_data=try_get( - self._extract_context(ytcfg), lambda x: x['client']['visitorData'], compat_str)) - query = { - 'params': params or 'wgYCCAA=', - 'browseId': browse_id or 'VL%s' % item_id - } - return self._extract_response( - item_id=item_id, headers=headers, query=query, - check_get_keys='contents', fatal=False, - note='Downloading API JSON with unavailable videos') - - def _extract_response(self, item_id, query, note='Downloading API JSON', headers=None, - ytcfg=None, check_get_keys=None, ep='browse', fatal=True): - response = None - last_error = None - count = -1 - retries = self.get_param('extractor_retries', 3) - if check_get_keys is None: - check_get_keys = [] - while count < retries: - count += 1 - if last_error: - self.report_warning('%s. Retrying ...' % last_error) - try: - response = self._call_api( - ep=ep, fatal=True, headers=headers, - video_id=item_id, query=query, - context=self._extract_context(ytcfg), - api_key=self._extract_api_key(ytcfg), - 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): - # 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 - if fatal: - raise - else: - self.report_warning(error_to_compat_str(e)) - return + nav_item_renderer = menu_item.get('menuNavigationItemRenderer') + text = try_get( + nav_item_renderer, lambda x: x['text']['simpleText'], compat_str) + if not text or text.lower() != 'show unavailable videos': + continue + browse_endpoint = try_get( + nav_item_renderer, lambda x: x['navigationEndpoint']['browseEndpoint'], dict) or {} + browse_id = browse_endpoint.get('browseId') + params = browse_endpoint.get('params') + break - else: - # Youtube may send alerts if there was an issue with the continuation page - self._extract_and_report_alerts(response, expected=False) - if not check_get_keys or dict_get(response, check_get_keys): - break - # Youtube sometimes sends incomplete data - # See: https://github.com/ytdl-org/youtube-dl/issues/28194 - last_error = 'Incomplete data received' - if count >= retries: - if fatal: - raise ExtractorError(last_error) - else: - self.report_warning(last_error) - return - return response + 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( + self._extract_context(ytcfg), lambda x: x['client']['visitorData'], compat_str)) + query = { + 'params': params or 'wgYCCAA=', + 'browseId': browse_id or 'VL%s' % item_id + } + return self._extract_response( + item_id=item_id, headers=headers, query=query, + check_get_keys='contents', fatal=False, ytcfg=ytcfg, + note='Downloading API JSON with unavailable videos') def _extract_webpage(self, url, item_id): retries = self.get_param('extractor_retries', 3) @@ -3636,7 +4038,7 @@ def __real_extract(self, url, smuggled_data): def get_mobj(url): mobj = self._url_re.match(url).groupdict() - mobj.update((k, '') for k,v in mobj.items() if v is None) + mobj.update((k, '') for k, v in mobj.items() if v is None) return mobj mobj = get_mobj(url) @@ -3696,23 +4098,26 @@ def get_mobj(url): if tabs: selected_tab = self._extract_selected_tab(tabs) tab_name = selected_tab.get('title', '') - if (mobj['tab'] == '/videos' and tab_name.lower() != mobj['tab'][1:] - and 'no-youtube-channel-redirect' not in compat_opts): - if not mobj['not_channel'] and item_id[:2] == 'UC': - # Topic channels don't have /videos. Use the equivalent playlist instead - self.report_warning('The URL does not have a %s tab. Trying to redirect to playlist UU%s instead' % (mobj['tab'][1:], item_id[2:])) - pl_id = 'UU%s' % item_id[2:] - pl_url = 'https://www.youtube.com/playlist?list=%s%s' % (pl_id, mobj['post']) - try: - pl_webpage, pl_data = self._extract_webpage(pl_url, pl_id) - for alert_type, alert_message in self._extract_alerts(pl_data): - if alert_type == 'error': - raise ExtractorError('Youtube said: %s' % alert_message) - item_id, url, webpage, data = pl_id, pl_url, pl_webpage, pl_data - except ExtractorError: - self.report_warning('The playlist gave error. Falling back to channel URL') - else: - self.report_warning('The URL does not have a %s tab. %s is being downloaded instead' % (mobj['tab'][1:], tab_name)) + if 'no-youtube-channel-redirect' not in compat_opts: + if mobj['tab'] == '/live': + # Live tab should have redirected to the video + raise ExtractorError('The channel is not currently live', expected=True) + if mobj['tab'] == '/videos' and tab_name.lower() != mobj['tab'][1:]: + if not mobj['not_channel'] and item_id[:2] == 'UC': + # Topic channels don't have /videos. Use the equivalent playlist instead + self.report_warning('The URL does not have a %s tab. Trying to redirect to playlist UU%s instead' % (mobj['tab'][1:], item_id[2:])) + pl_id = 'UU%s' % item_id[2:] + pl_url = 'https://www.youtube.com/playlist?list=%s%s' % (pl_id, mobj['post']) + try: + pl_webpage, pl_data = self._extract_webpage(pl_url, pl_id) + for alert_type, alert_message in self._extract_alerts(pl_data): + if alert_type == 'error': + raise ExtractorError('Youtube said: %s' % alert_message) + item_id, url, webpage, data = pl_id, pl_url, pl_webpage, pl_data + except ExtractorError: + self.report_warning('The playlist gave error. Falling back to channel URL') + else: + self.report_warning('The URL does not have a %s tab. %s is being downloaded instead' % (mobj['tab'][1:], tab_name)) self.write_debug('Final URL: %s' % url) @@ -3720,7 +4125,6 @@ def get_mobj(url): if 'no-youtube-unavailable-videos' not in compat_opts: data = self._reload_with_unavailable_videos(item_id, data, webpage) or data self._extract_and_report_alerts(data) - tabs = try_get( data, lambda x: x['contents']['twoColumnBrowseResultsRenderer']['tabs'], list) if tabs: @@ -3735,7 +4139,8 @@ def get_mobj(url): data, lambda x: x['currentVideoEndpoint']['watchEndpoint']['videoId'], compat_str) or video_id if video_id: - self.report_warning('Unable to recognize playlist. Downloading just video %s' % video_id) + if mobj['tab'] != '/live': # live tab is expected to redirect to video + self.report_warning('Unable to recognize playlist. Downloading just video %s' % video_id) return self.url_result(video_id, ie=YoutubeIE.ie_key(), video_id=video_id) raise ExtractorError('Unable to recognize tab page') @@ -3816,7 +4221,7 @@ def suitable(cls, url): def _real_extract(self, url): playlist_id = self._match_id(url) - is_music_url = self.is_music_url(url) + is_music_url = YoutubeBaseInfoExtractor.is_music_url(url) url = update_url_query( 'https://www.youtube.com/playlist', parse_qs(url) or {'list': playlist_id}) @@ -3914,7 +4319,9 @@ def _entries(self, query, n): if self._SEARCH_PARAMS: data['params'] = self._SEARCH_PARAMS total = 0 + continuation = {} for page_num in itertools.count(1): + data.update(continuation) search = self._extract_response( item_id='query "%s" page %s' % (query, page_num), ep='search', query=data, check_get_keys=('contents', 'onResponseReceivedCommands') @@ -3932,13 +4339,10 @@ def _entries(self, query, n): # Youtube sometimes adds promoted content to searches, # changing the index location of videos and token. # So we search through all entries till we find them. - continuation_token = None + continuation = None for slr_content in slr_contents: - if continuation_token is None: - continuation_token = try_get( - slr_content, - lambda x: x['continuationItemRenderer']['continuationEndpoint']['continuationCommand']['token'], - compat_str) + if not continuation: + continuation = self._extract_continuation({'contents': [slr_content]}) isr_contents = try_get( slr_content, @@ -3961,9 +4365,8 @@ def _entries(self, query, n): if total == n: return - if not continuation_token: + if not continuation: break - data['continuation'] = continuation_token def _get_n_results(self, query, n): """Get a specified number of results for a query""" @@ -4016,9 +4419,6 @@ class YoutubeFeedsInfoExtractor(YoutubeTabIE): def IE_NAME(self): return 'youtube:%s' % self._FEED_NAME - def _real_initialize(self): - self._login() - def _real_extract(self, url): return self.url_result( 'https://www.youtube.com/feed/%s' % self._FEED_NAME, @@ -4043,6 +4443,7 @@ class YoutubeRecommendedIE(YoutubeFeedsInfoExtractor): IE_DESC = 'YouTube.com recommended videos, ":ytrec" for short (requires authentication)' _VALID_URL = r'https?://(?:www\.)?youtube\.com/?(?:[?#]|$)|:ytrec(?:ommended)?' _FEED_NAME = 'recommended' + _LOGIN_REQUIRED = False _TESTS = [{ 'url': ':ytrec', 'only_matching': True,