yt_dlp/extractor/facebook.py

   1 import json
   2 import re
   3 import urllib.parse
   4
   5 from .common import InfoExtractor
   6 from ..compat import (
   7     compat_etree_fromstring,
   8     compat_str,
   9     compat_urllib_parse_unquote,
  10 )
  11 from ..networking import Request
  12 from ..networking.exceptions import network_exceptions
  13 from ..utils import (
  14     ExtractorError,
  15     clean_html,
  16     determine_ext,
  17     error_to_compat_str,
  18     float_or_none,
  19     get_element_by_id,
  20     get_first,
  21     int_or_none,
  22     js_to_json,
  23     merge_dicts,
  24     parse_count,
  25     parse_qs,
  26     qualities,
  27     str_or_none,
  28     traverse_obj,
  29     try_get,
  30     url_or_none,
  31     urlencode_postdata,
  32     urljoin,
  33     variadic,
  34 )
  35
  36
  37 class FacebookIE(InfoExtractor):
  38     _VALID_URL = r'''(?x)
  39                 (?:
  40                     https?://
  41                         (?:[\w-]+\.)?(?:facebook\.com|facebookwkhpilnemxj7asaniu7vnjjbiltxjqhye3mhbshg7kx5tfyd\.onion)/
  42                         (?:[^#]*?\#!/)?
  43                         (?:
  44                             (?:
  45                                 video/video\.php|
  46                                 photo\.php|
  47                                 video\.php|
  48                                 video/embed|
  49                                 story\.php|
  50                                 watch(?:/live)?/?
  51                             )\?(?:.*?)(?:v|video_id|story_fbid)=|
  52                             [^/]+/videos/(?:[^/]+/)?|
  53                             [^/]+/posts/|
  54                             groups/[^/]+/permalink/|
  55                             watchparty/
  56                         )|
  57                     facebook:
  58                 )
  59                 (?P<id>[0-9]+)
  60                 '''
  61     _EMBED_REGEX = [
  62         r'<iframe[^>]+?src=(["\'])(?P<url>https?://www\.facebook\.com/(?:video/embed|plugins/video\.php).+?)\1',
  63         # Facebook API embed https://developers.facebook.com/docs/plugins/embedded-video-player
  64         r'''(?x)<div[^>]+
  65                 class=(?P<q1>[\'"])[^\'"]*\bfb-(?:video|post)\b[^\'"]*(?P=q1)[^>]+
  66                 data-href=(?P<q2>[\'"])(?P<url>(?:https?:)?//(?:www\.)?facebook.com/.+?)(?P=q2)''',
  67     ]
  68     _LOGIN_URL = 'https://www.facebook.com/login.php?next=http%3A%2F%2Ffacebook.com%2Fhome.php&login_attempt=1'
  69     _CHECKPOINT_URL = 'https://www.facebook.com/checkpoint/?next=http%3A%2F%2Ffacebook.com%2Fhome.php&_fb_noscript=1'
  70     _NETRC_MACHINE = 'facebook'
  71     IE_NAME = 'facebook'
  72
  73     _VIDEO_PAGE_TEMPLATE = 'https://www.facebook.com/video/video.php?v=%s'
  74     _VIDEO_PAGE_TAHOE_TEMPLATE = 'https://www.facebook.com/video/tahoe/async/%s/?chain=true&isvideo=true&payloadtype=primary'
  75
  76     _TESTS = [{
  77         'url': 'https://www.facebook.com/radiokicksfm/videos/3676516585958356/',
  78         'info_dict': {
  79             'id': '3676516585958356',
  80             'ext': 'mp4',
  81             'title': 'dr Adam Przygoda',
  82             'description': 'md5:34675bda53336b1d16400265c2bb9b3b',
  83             'uploader': 'RADIO KICKS FM',
  84             'upload_date': '20230818',
  85             'timestamp': 1692346159,
  86             'thumbnail': r're:^https?://.*',
  87             'uploader_id': '100063551323670',
  88             'duration': 3132.184,
  89             'view_count': int,
  90             'concurrent_view_count': 0,
  91         },
  92     }, {
  93         'url': 'https://www.facebook.com/video.php?v=637842556329505&fref=nf',
  94         'md5': '6a40d33c0eccbb1af76cf0485a052659',
  95         'info_dict': {
  96             'id': '637842556329505',
  97             'ext': 'mp4',
  98             'title': 're:Did you know Kei Nishikori is the first Asian man to ever reach a Grand Slam',
  99             'uploader': 'Tennis on Facebook',
 100             'upload_date': '20140908',
 101             'timestamp': 1410199200,
 102         },
 103         'skip': 'Requires logging in',
 104     }, {
 105         # data.video
 106         'url': 'https://www.facebook.com/video.php?v=274175099429670',
 107         'info_dict': {
 108             'id': '274175099429670',
 109             'ext': 'mp4',
 110             'title': 'Asif',
 111             'description': '',
 112             'uploader': 'Asif Nawab Butt',
 113             'upload_date': '20140506',
 114             'timestamp': 1399398998,
 115             'thumbnail': r're:^https?://.*',
 116             'uploader_id': 'pfbid028wxorhX2ErLFJ578N6P3crHD3PHmXTCqCvfBpsnbSLmbokwSY75p5hWBjHGkG4zxl',
 117             'duration': 131.03,
 118             'concurrent_view_count': int,
 119         },
 120     }, {
 121         'note': 'Video with DASH manifest',
 122         'url': 'https://www.facebook.com/video.php?v=957955867617029',
 123         'md5': 'b2c28d528273b323abe5c6ab59f0f030',
 124         'info_dict': {
 125             'id': '957955867617029',
 126             'ext': 'mp4',
 127             'title': 'When you post epic content on instagram.com/433 8 million followers, this is ...',
 128             'uploader': 'Demy de Zeeuw',
 129             'upload_date': '20160110',
 130             'timestamp': 1452431627,
 131         },
 132         'skip': 'Requires logging in',
 133     }, {
 134         'url': 'https://www.facebook.com/maxlayn/posts/10153807558977570',
 135         'md5': '037b1fa7f3c2d02b7a0d7bc16031ecc6',
 136         'info_dict': {
 137             'id': '544765982287235',
 138             'ext': 'mp4',
 139             'title': '"What are you doing running in the snow?"',
 140             'uploader': 'FailArmy',
 141         },
 142         'skip': 'Video gone',
 143     }, {
 144         'url': 'https://m.facebook.com/story.php?story_fbid=1035862816472149&id=116132035111903',
 145         'md5': '1deb90b6ac27f7efcf6d747c8a27f5e3',
 146         'info_dict': {
 147             'id': '1035862816472149',
 148             'ext': 'mp4',
 149             'title': 'What the Flock Is Going On In New Zealand  Credit: ViralHog',
 150             'uploader': 'S. Saint',
 151         },
 152         'skip': 'Video gone',
 153     }, {
 154         'note': 'swf params escaped',
 155         'url': 'https://www.facebook.com/barackobama/posts/10153664894881749',
 156         'md5': '97ba073838964d12c70566e0085c2b91',
 157         'info_dict': {
 158             'id': '10153664894881749',
 159             'ext': 'mp4',
 160             'title': 'Average time to confirm recent Supreme Court nominees: 67 days Longest it\'s t...',
 161             'thumbnail': r're:^https?://.*',
 162             'timestamp': 1456259628,
 163             'upload_date': '20160223',
 164             'uploader': 'Barack Obama',
 165         },
 166         'skip': 'Gif on giphy.com gone',
 167     }, {
 168         # have 1080P, but only up to 720p in swf params
 169         # data.video.story.attachments[].media
 170         'url': 'https://www.facebook.com/cnn/videos/10155529876156509/',
 171         'md5': 'ca63897a90c9452efee5f8c40d080e25',
 172         'info_dict': {
 173             'id': '10155529876156509',
 174             'ext': 'mp4',
 175             'title': 'Holocaust survivor becomes US citizen',
 176             'description': 'She survived the holocaust — and years later, she’s getting her citizenship so she can vote for Hillary Clinton http://cnn.it/2eERh5f',
 177             'timestamp': 1477818095,
 178             'upload_date': '20161030',
 179             'uploader': 'CNN',
 180             'thumbnail': r're:^https?://.*',
 181             'view_count': int,
 182             'uploader_id': '100059479812265',
 183             'concurrent_view_count': int,
 184             'duration': 44.478,
 185         },
 186     }, {
 187         # bigPipe.onPageletArrive ... onPageletArrive pagelet_group_mall
 188         # data.node.comet_sections.content.story.attachments[].style_type_renderer.attachment.media
 189         'url': 'https://www.facebook.com/yaroslav.korpan/videos/1417995061575415/',
 190         'info_dict': {
 191             'id': '1417995061575415',
 192             'ext': 'mp4',
 193             'title': 'Довгоочікуване відео | By Yaroslav - Facebook',
 194             'description': 'Довгоочікуване відео',
 195             'timestamp': 1486648217,
 196             'upload_date': '20170209',
 197             'uploader': 'Yaroslav Korpan',
 198             'uploader_id': 'pfbid06AScABAWcW91qpiuGrLt99Ef9tvwHoXP6t8KeFYEqkSfreMtfa9nTveh8b2ZEVSWl',
 199             'concurrent_view_count': int,
 200             'thumbnail': r're:^https?://.*',
 201             'view_count': int,
 202             'duration': 11736.446,
 203         },
 204         'params': {
 205             'skip_download': True,
 206         },
 207     }, {
 208         # FIXME
 209         'url': 'https://www.facebook.com/LaGuiaDelVaron/posts/1072691702860471',
 210         'info_dict': {
 211             'id': '1072691702860471',
 212             'ext': 'mp4',
 213             'title': 'md5:ae2d22a93fbb12dad20dc393a869739d',
 214             'timestamp': 1477305000,
 215             'upload_date': '20161024',
 216             'uploader': 'La Guía Del Varón',
 217             'thumbnail': r're:^https?://.*',
 218         },
 219         'skip': 'Requires logging in',
 220     }, {
 221         # data.node.comet_sections.content.story.attachments[].style_type_renderer.attachment.media
 222         'url': 'https://www.facebook.com/groups/1024490957622648/permalink/1396382447100162/',
 223         'info_dict': {
 224             'id': '202882990186699',
 225             'ext': 'mp4',
 226             'title': 'birb (O v O") | Hello? Yes your uber ride is here',
 227             'description': 'Hello? Yes your uber ride is here * Jukin Media Verified * Find this video and others like it by visiting...',
 228             'timestamp': 1486035513,
 229             'upload_date': '20170202',
 230             'uploader': 'Elisabeth Ahtn',
 231             'uploader_id': '100013949973717',
 232         },
 233         'skip': 'Requires logging in',
 234     }, {
 235         'url': 'https://www.facebook.com/video.php?v=10204634152394104',
 236         'only_matching': True,
 237     }, {
 238         'url': 'https://www.facebook.com/amogood/videos/1618742068337349/?fref=nf',
 239         'only_matching': True,
 240     }, {
 241         # data.mediaset.currMedia.edges
 242         'url': 'https://www.facebook.com/ChristyClarkForBC/videos/vb.22819070941/10153870694020942/?type=2&theater',
 243         'only_matching': True,
 244     }, {
 245         # data.video.story.attachments[].media
 246         'url': 'facebook:544765982287235',
 247         'only_matching': True,
 248     }, {
 249         # data.node.comet_sections.content.story.attachments[].style_type_renderer.attachment.media
 250         'url': 'https://www.facebook.com/groups/164828000315060/permalink/764967300301124/',
 251         'only_matching': True,
 252     }, {
 253         # data.video.creation_story.attachments[].media
 254         'url': 'https://zh-hk.facebook.com/peoplespower/videos/1135894589806027/',
 255         'only_matching': True,
 256     }, {
 257         # data.video
 258         'url': 'https://www.facebookwkhpilnemxj7asaniu7vnjjbiltxjqhye3mhbshg7kx5tfyd.onion/video.php?v=274175099429670',
 259         'only_matching': True,
 260     }, {
 261         # no title
 262         'url': 'https://www.facebook.com/onlycleverentertainment/videos/1947995502095005/',
 263         'only_matching': True,
 264     }, {
 265         # data.video
 266         'url': 'https://www.facebook.com/WatchESLOne/videos/359649331226507/',
 267         'info_dict': {
 268             'id': '359649331226507',
 269             'ext': 'mp4',
 270             'title': 'Fnatic vs. EG - Group A - Opening Match - ESL One Birmingham Day 1',
 271             'description': '#ESLOne VoD - Birmingham Finals Day#1 Fnatic vs. @Evil Geniuses',
 272             'timestamp': 1527084179,
 273             'upload_date': '20180523',
 274             'uploader': 'ESL One Dota 2',
 275             'uploader_id': '100066514874195',
 276             'duration': 4524.212,
 277             'view_count': int,
 278             'thumbnail': r're:^https?://.*',
 279             'concurrent_view_count': int,
 280         },
 281         'params': {
 282             'skip_download': True,
 283         },
 284     }, {
 285         # data.node.comet_sections.content.story.attachments[].style_type_renderer.attachment.all_subattachments.nodes[].media
 286         'url': 'https://www.facebook.com/100033620354545/videos/106560053808006/',
 287         'info_dict': {
 288             'id': '106560053808006',
 289             'ext': 'mp4',
 290             'title': 'Josef',
 291             'thumbnail': r're:^https?://.*',
 292             'concurrent_view_count': int,
 293             'uploader_id': 'pfbid0cibUN6tV7DYgdbJdsUFN46wc4jKpVSPAvJQhFofGqBGmVn3V3JtAs2tfUwziw2hUl',
 294             'timestamp': 1549275572,
 295             'duration': 3.413,
 296             'uploader': 'Josef Novak',
 297             'description': '',
 298             'upload_date': '20190204',
 299         },
 300     }, {
 301         # data.video.story.attachments[].media
 302         'url': 'https://www.facebook.com/watch/?v=647537299265662',
 303         'only_matching': True,
 304     }, {
 305         # FIXME: https://github.com/yt-dlp/yt-dlp/issues/542
 306         # data.node.comet_sections.content.story.attachments[].style_type_renderer.attachment.all_subattachments.nodes[].media
 307         'url': 'https://www.facebook.com/PankajShahLondon/posts/10157667649866271',
 308         'info_dict': {
 309             'id': '10157667649866271',
 310         },
 311         'playlist_count': 3,
 312         'skip': 'Requires logging in',
 313     }, {
 314         # data.nodes[].comet_sections.content.story.attachments[].style_type_renderer.attachment.media
 315         'url': 'https://m.facebook.com/Alliance.Police.Department/posts/4048563708499330',
 316         'info_dict': {
 317             'id': '117576630041613',
 318             'ext': 'mp4',
 319             # TODO: title can be extracted from video page
 320             'title': 'Facebook video #117576630041613',
 321             'uploader_id': '189393014416438',
 322             'upload_date': '20201123',
 323             'timestamp': 1606162592,
 324         },
 325         'skip': 'Requires logging in',
 326     }, {
 327         # node.comet_sections.content.story.attached_story.attachments.style_type_renderer.attachment.media
 328         'url': 'https://www.facebook.com/groups/ateistiskselskab/permalink/10154930137678856/',
 329         'info_dict': {
 330             'id': '211567722618337',
 331             'ext': 'mp4',
 332             'title': 'Facebook video #211567722618337',
 333             'uploader_id': '127875227654254',
 334             'upload_date': '20161122',
 335             'timestamp': 1479793574,
 336         },
 337         'skip': 'No video',
 338     }, {
 339         # data.video.creation_story.attachments[].media
 340         'url': 'https://www.facebook.com/watch/live/?v=1823658634322275',
 341         'only_matching': True,
 342     }, {
 343         'url': 'https://www.facebook.com/watchparty/211641140192478',
 344         'info_dict': {
 345             'id': '211641140192478',
 346         },
 347         'playlist_count': 1,
 348         'skip': 'Requires logging in',
 349     }]
 350     _SUPPORTED_PAGLETS_REGEX = r'(?:pagelet_group_mall|permalink_video_pagelet|hyperfeed_story_id_[0-9a-f]+)'
 351     _api_config = {
 352         'graphURI': '/api/graphql/'
 353     }
 354
 355     def _perform_login(self, username, password):
 356         login_page_req = Request(self._LOGIN_URL)
 357         self._set_cookie('facebook.com', 'locale', 'en_US')
 358         login_page = self._download_webpage(login_page_req, None,
 359                                             note='Downloading login page',
 360                                             errnote='Unable to download login page')
 361         lsd = self._search_regex(
 362             r'<input type="hidden" name="lsd" value="([^"]*)"',
 363             login_page, 'lsd')
 364         lgnrnd = self._search_regex(r'name="lgnrnd" value="([^"]*?)"', login_page, 'lgnrnd')
 365
 366         login_form = {
 367             'email': username,
 368             'pass': password,
 369             'lsd': lsd,
 370             'lgnrnd': lgnrnd,
 371             'next': 'http://facebook.com/home.php',
 372             'default_persistent': '0',
 373             'legacy_return': '1',
 374             'timezone': '-60',
 375             'trynum': '1',
 376         }
 377         request = Request(self._LOGIN_URL, urlencode_postdata(login_form))
 378         request.headers['Content-Type'] = 'application/x-www-form-urlencoded'
 379         try:
 380             login_results = self._download_webpage(request, None,
 381                                                    note='Logging in', errnote='unable to fetch login page')
 382             if re.search(r'<form(.*)name="login"(.*)</form>', login_results) is not None:
 383                 error = self._html_search_regex(
 384                     r'(?s)<div[^>]+class=(["\']).*?login_error_box.*?\1[^>]*><div[^>]*>.*?</div><div[^>]*>(?P<error>.+?)</div>',
 385                     login_results, 'login error', default=None, group='error')
 386                 if error:
 387                     raise ExtractorError('Unable to login: %s' % error, expected=True)
 388                 self.report_warning('unable to log in: bad username/password, or exceeded login rate limit (~3/min). Check credentials or wait.')
 389                 return
 390
 391             fb_dtsg = self._search_regex(
 392                 r'name="fb_dtsg" value="(.+?)"', login_results, 'fb_dtsg', default=None)
 393             h = self._search_regex(
 394                 r'name="h"\s+(?:\w+="[^"]+"\s+)*?value="([^"]+)"', login_results, 'h', default=None)
 395
 396             if not fb_dtsg or not h:
 397                 return
 398
 399             check_form = {
 400                 'fb_dtsg': fb_dtsg,
 401                 'h': h,
 402                 'name_action_selected': 'dont_save',
 403             }
 404             check_req = Request(self._CHECKPOINT_URL, urlencode_postdata(check_form))
 405             check_req.headers['Content-Type'] = 'application/x-www-form-urlencoded'
 406             check_response = self._download_webpage(check_req, None,
 407                                                     note='Confirming login')
 408             if re.search(r'id="checkpointSubmitButton"', check_response) is not None:
 409                 self.report_warning('Unable to confirm login, you have to login in your browser and authorize the login.')
 410         except network_exceptions as err:
 411             self.report_warning('unable to log in: %s' % error_to_compat_str(err))
 412             return
 413
 414     def _extract_from_url(self, url, video_id):
 415         webpage = self._download_webpage(
 416             url.replace('://m.facebook.com/', '://www.facebook.com/'), video_id)
 417
 418         def extract_metadata(webpage):
 419             post_data = [self._parse_json(j, video_id, fatal=False) for j in re.findall(
 420                 r'data-sjs>({.*?ScheduledServerJS.*?})</script>', webpage)]
 421             post = traverse_obj(post_data, (
 422                 ..., 'require', ..., ..., ..., '__bbox', 'require', ..., ..., ..., '__bbox', 'result', 'data'), expected_type=dict) or []
 423             media = traverse_obj(post, (..., 'attachments', ..., lambda k, v: (
 424                 k == 'media' and str(v['id']) == video_id and v['__typename'] == 'Video')), expected_type=dict)
 425             title = get_first(media, ('title', 'text'))
 426             description = get_first(media, ('creation_story', 'comet_sections', 'message', 'story', 'message', 'text'))
 427             uploader_data = (
 428                 get_first(media, ('owner', {dict}))
 429                 or get_first(post, (..., 'video', lambda k, v: k == 'owner' and v['name']))
 430                 or get_first(post, ('node', 'actors', ..., {dict})) or {})
 431
 432             page_title = title or self._html_search_regex((
 433                 r'<h2\s+[^>]*class="uiHeaderTitle"[^>]*>(?P<content>[^<]*)</h2>',
 434                 r'(?s)<span class="fbPhotosPhotoCaption".*?id="fbPhotoPageCaption"><span class="hasCaption">(?P<content>.*?)</span>',
 435                 self._meta_regex('og:title'), self._meta_regex('twitter:title'), r'<title>(?P<content>.+?)</title>'
 436             ), webpage, 'title', default=None, group='content')
 437             description = description or self._html_search_meta(
 438                 ['description', 'og:description', 'twitter:description'],
 439                 webpage, 'description', default=None)
 440             uploader = uploader_data.get('name') or (
 441                 clean_html(get_element_by_id('fbPhotoPageAuthorName', webpage))
 442                 or self._search_regex(
 443                     (r'ownerName\s*:\s*"([^"]+)"', *self._og_regexes('title')), webpage, 'uploader', fatal=False))
 444
 445             timestamp = int_or_none(self._search_regex(
 446                 r'<abbr[^>]+data-utime=["\'](\d+)', webpage,
 447                 'timestamp', default=None))
 448             thumbnail = self._html_search_meta(
 449                 ['og:image', 'twitter:image'], webpage, 'thumbnail', default=None)
 450             # some webpages contain unretrievable thumbnail urls
 451             # like https://lookaside.fbsbx.com/lookaside/crawler/media/?media_id=10155168902769113&get_thumbnail=1
 452             # in https://www.facebook.com/yaroslav.korpan/videos/1417995061575415/
 453             if thumbnail and not re.search(r'\.(?:jpg|png)', thumbnail):
 454                 thumbnail = None
 455             info_dict = {
 456                 'description': description,
 457                 'uploader': uploader,
 458                 'uploader_id': uploader_data.get('id'),
 459                 'timestamp': timestamp,
 460                 'thumbnail': thumbnail,
 461                 'view_count': parse_count(self._search_regex(
 462                     (r'\bviewCount\s*:\s*["\']([\d,.]+)', r'video_view_count["\']\s*:\s*(\d+)',),
 463                     webpage, 'view count', default=None)),
 464                 'concurrent_view_count': get_first(post, (
 465                     ('video', (..., ..., 'attachments', ..., 'media')), 'liveViewerCount', {int_or_none})),
 466             }
 467
 468             info_json_ld = self._search_json_ld(webpage, video_id, default={})
 469             info_json_ld['title'] = (re.sub(r'\s*\|\s*Facebook$', '', title or info_json_ld.get('title') or page_title or '')
 470                                      or (description or '').replace('\n', ' ') or f'Facebook video #{video_id}')
 471             return merge_dicts(info_json_ld, info_dict)
 472
 473         video_data = None
 474
 475         def extract_video_data(instances):
 476             video_data = []
 477             for item in instances:
 478                 if try_get(item, lambda x: x[1][0]) == 'VideoConfig':
 479                     video_item = item[2][0]
 480                     if video_item.get('video_id'):
 481                         video_data.append(video_item['videoData'])
 482             return video_data
 483
 484         server_js_data = self._parse_json(self._search_regex(
 485             [r'handleServerJS\(({.+})(?:\);|,")', r'\bs\.handle\(({.+?})\);'],
 486             webpage, 'server js data', default='{}'), video_id, fatal=False)
 487
 488         if server_js_data:
 489             video_data = extract_video_data(server_js_data.get('instances', []))
 490
 491         def extract_from_jsmods_instances(js_data):
 492             if js_data:
 493                 return extract_video_data(try_get(
 494                     js_data, lambda x: x['jsmods']['instances'], list) or [])
 495
 496         def extract_dash_manifest(video, formats):
 497             dash_manifest = video.get('dash_manifest')
 498             if dash_manifest:
 499                 formats.extend(self._parse_mpd_formats(
 500                     compat_etree_fromstring(urllib.parse.unquote_plus(dash_manifest)),
 501                     mpd_url=video.get('dash_manifest_url')))
 502
 503         def process_formats(info):
 504             # Downloads with browser's User-Agent are rate limited. Working around
 505             # with non-browser User-Agent.
 506             for f in info['formats']:
 507                 f.setdefault('http_headers', {})['User-Agent'] = 'facebookexternalhit/1.1'
 508
 509         def extract_relay_data(_filter):
 510             return self._parse_json(self._search_regex(
 511                 r'data-sjs>({.*?%s.*?})</script>' % _filter,
 512                 webpage, 'replay data', default='{}'), video_id, fatal=False) or {}
 513
 514         def extract_relay_prefetched_data(_filter):
 515             return traverse_obj(extract_relay_data(_filter), (
 516                 'require', (None, (..., ..., ..., '__bbox', 'require')),
 517                 lambda _, v: 'RelayPrefetchedStreamCache' in v, ..., ...,
 518                 '__bbox', 'result', 'data', {dict}), get_all=False) or {}
 519
 520         if not video_data:
 521             server_js_data = self._parse_json(self._search_regex([
 522                 r'bigPipe\.onPageletArrive\(({.+?})\)\s*;\s*}\s*\)\s*,\s*["\']onPageletArrive\s+' + self._SUPPORTED_PAGLETS_REGEX,
 523                 r'bigPipe\.onPageletArrive\(({.*?id\s*:\s*"%s".*?})\);' % self._SUPPORTED_PAGLETS_REGEX
 524             ], webpage, 'js data', default='{}'), video_id, js_to_json, False)
 525             video_data = extract_from_jsmods_instances(server_js_data)
 526
 527         if not video_data:
 528             data = extract_relay_prefetched_data(
 529                 r'"(?:dash_manifest|playable_url(?:_quality_hd)?)')
 530             if data:
 531                 entries = []
 532
 533                 def parse_graphql_video(video):
 534                     v_id = video.get('videoId') or video.get('id') or video_id
 535                     reel_info = traverse_obj(
 536                         video, ('creation_story', 'short_form_video_context', 'playback_video', {dict}))
 537                     if reel_info:
 538                         video = video['creation_story']
 539                         video['owner'] = traverse_obj(video, ('short_form_video_context', 'video_owner'))
 540                         video.update(reel_info)
 541                     formats = []
 542                     q = qualities(['sd', 'hd'])
 543                     for key, format_id in (('playable_url', 'sd'), ('playable_url_quality_hd', 'hd'),
 544                                            ('playable_url_dash', ''), ('browser_native_hd_url', 'hd'),
 545                                            ('browser_native_sd_url', 'sd')):
 546                         playable_url = video.get(key)
 547                         if not playable_url:
 548                             continue
 549                         if determine_ext(playable_url) == 'mpd':
 550                             formats.extend(self._extract_mpd_formats(playable_url, video_id))
 551                         else:
 552                             formats.append({
 553                                 'format_id': format_id,
 554                                 # sd, hd formats w/o resolution info should be deprioritized below DASH
 555                                 'quality': q(format_id) - 3,
 556                                 'url': playable_url,
 557                             })
 558                     extract_dash_manifest(video, formats)
 559                     info = {
 560                         'id': v_id,
 561                         'formats': formats,
 562                         'thumbnail': traverse_obj(
 563                             video, ('thumbnailImage', 'uri'), ('preferred_thumbnail', 'image', 'uri')),
 564                         'uploader_id': traverse_obj(video, ('owner', 'id', {str_or_none})),
 565                         'timestamp': traverse_obj(video, 'publish_time', 'creation_time', expected_type=int_or_none),
 566                         'duration': (float_or_none(video.get('playable_duration_in_ms'), 1000)
 567                                      or float_or_none(video.get('length_in_second'))),
 568                     }
 569                     process_formats(info)
 570                     description = try_get(video, lambda x: x['savable_description']['text'])
 571                     title = video.get('name')
 572                     if title:
 573                         info.update({
 574                             'title': title,
 575                             'description': description,
 576                         })
 577                     else:
 578                         info['title'] = description or 'Facebook video #%s' % v_id
 579                     entries.append(info)
 580
 581                 def parse_attachment(attachment, key='media'):
 582                     media = attachment.get(key) or {}
 583                     if media.get('__typename') == 'Video':
 584                         return parse_graphql_video(media)
 585
 586                 nodes = variadic(traverse_obj(data, 'nodes', 'node') or [])
 587                 attachments = traverse_obj(nodes, (
 588                     ..., 'comet_sections', 'content', 'story', (None, 'attached_story'), 'attachments',
 589                     ..., ('styles', 'style_type_renderer'), 'attachment'), expected_type=dict) or []
 590                 for attachment in attachments:
 591                     ns = try_get(attachment, lambda x: x['all_subattachments']['nodes'], list) or []
 592                     for n in ns:
 593                         parse_attachment(n)
 594                     parse_attachment(attachment)
 595
 596                 edges = try_get(data, lambda x: x['mediaset']['currMedia']['edges'], list) or []
 597                 for edge in edges:
 598                     parse_attachment(edge, key='node')
 599
 600                 video = data.get('video') or {}
 601                 if video:
 602                     attachments = try_get(video, [
 603                         lambda x: x['story']['attachments'],
 604                         lambda x: x['creation_story']['attachments']
 605                     ], list) or []
 606                     for attachment in attachments:
 607                         parse_attachment(attachment)
 608                     if not entries:
 609                         parse_graphql_video(video)
 610
 611                 if len(entries) > 1:
 612                     return self.playlist_result(entries, video_id)
 613
 614                 video_info = entries[0]
 615                 webpage_info = extract_metadata(webpage)
 616                 # honor precise duration in video info
 617                 if video_info.get('duration'):
 618                     webpage_info['duration'] = video_info['duration']
 619                 return merge_dicts(webpage_info, video_info)
 620
 621         if not video_data:
 622             m_msg = re.search(r'class="[^"]*uiInterstitialContent[^"]*"><div>(.*?)</div>', webpage)
 623             if m_msg is not None:
 624                 raise ExtractorError(
 625                     'The video is not available, Facebook said: "%s"' % m_msg.group(1),
 626                     expected=True)
 627             elif any(p in webpage for p in (
 628                     '>You must log in to continue',
 629                     'id="login_form"',
 630                     'id="loginbutton"')):
 631                 self.raise_login_required()
 632
 633         if not video_data and '/watchparty/' in url:
 634             post_data = {
 635                 'doc_id': 3731964053542869,
 636                 'variables': json.dumps({
 637                     'livingRoomID': video_id,
 638                 }),
 639             }
 640
 641             prefetched_data = extract_relay_prefetched_data(r'"login_data"\s*:\s*{')
 642             if prefetched_data:
 643                 lsd = try_get(prefetched_data, lambda x: x['login_data']['lsd'], dict)
 644                 if lsd:
 645                     post_data[lsd['name']] = lsd['value']
 646
 647             relay_data = extract_relay_data(r'\[\s*"RelayAPIConfigDefaults"\s*,')
 648             for define in (relay_data.get('define') or []):
 649                 if define[0] == 'RelayAPIConfigDefaults':
 650                     self._api_config = define[2]
 651
 652             living_room = self._download_json(
 653                 urljoin(url, self._api_config['graphURI']), video_id,
 654                 data=urlencode_postdata(post_data))['data']['living_room']
 655
 656             entries = []
 657             for edge in (try_get(living_room, lambda x: x['recap']['watched_content']['edges']) or []):
 658                 video = try_get(edge, lambda x: x['node']['video']) or {}
 659                 v_id = video.get('id')
 660                 if not v_id:
 661                     continue
 662                 v_id = compat_str(v_id)
 663                 entries.append(self.url_result(
 664                     self._VIDEO_PAGE_TEMPLATE % v_id,
 665                     self.ie_key(), v_id, video.get('name')))
 666
 667             return self.playlist_result(entries, video_id)
 668
 669         if not video_data:
 670             # Video info not in first request, do a secondary request using
 671             # tahoe player specific URL
 672             tahoe_data = self._download_webpage(
 673                 self._VIDEO_PAGE_TAHOE_TEMPLATE % video_id, video_id,
 674                 data=urlencode_postdata({
 675                     '__a': 1,
 676                     '__pc': self._search_regex(
 677                         r'pkg_cohort["\']\s*:\s*["\'](.+?)["\']', webpage,
 678                         'pkg cohort', default='PHASED:DEFAULT'),
 679                     '__rev': self._search_regex(
 680                         r'client_revision["\']\s*:\s*(\d+),', webpage,
 681                         'client revision', default='3944515'),
 682                     'fb_dtsg': self._search_regex(
 683                         r'"DTSGInitialData"\s*,\s*\[\]\s*,\s*{\s*"token"\s*:\s*"([^"]+)"',
 684                         webpage, 'dtsg token', default=''),
 685                 }),
 686                 headers={
 687                     'Content-Type': 'application/x-www-form-urlencoded',
 688                 })
 689             tahoe_js_data = self._parse_json(
 690                 self._search_regex(
 691                     r'for\s+\(\s*;\s*;\s*\)\s*;(.+)', tahoe_data,
 692                     'tahoe js data', default='{}'),
 693                 video_id, fatal=False)
 694             video_data = extract_from_jsmods_instances(tahoe_js_data)
 695
 696         if not video_data:
 697             raise ExtractorError('Cannot parse data')
 698
 699         if len(video_data) > 1:
 700             entries = []
 701             for v in video_data:
 702                 video_url = v[0].get('video_url')
 703                 if not video_url:
 704                     continue
 705                 entries.append(self.url_result(urljoin(
 706                     url, video_url), self.ie_key(), v[0].get('video_id')))
 707             return self.playlist_result(entries, video_id)
 708         video_data = video_data[0]
 709
 710         formats = []
 711         subtitles = {}
 712         for f in video_data:
 713             format_id = f['stream_type']
 714             if f and isinstance(f, dict):
 715                 f = [f]
 716             if not f or not isinstance(f, list):
 717                 continue
 718             for quality in ('sd', 'hd'):
 719                 for src_type in ('src', 'src_no_ratelimit'):
 720                     src = f[0].get('%s_%s' % (quality, src_type))
 721                     if src:
 722                         # sd, hd formats w/o resolution info should be deprioritized below DASH
 723                         # TODO: investigate if progressive or src formats still exist
 724                         preference = -10 if format_id == 'progressive' else -3
 725                         if quality == 'hd':
 726                             preference += 1
 727                         formats.append({
 728                             'format_id': '%s_%s_%s' % (format_id, quality, src_type),
 729                             'url': src,
 730                             'quality': preference,
 731                             'height': 720 if quality == 'hd' else None
 732                         })
 733             extract_dash_manifest(f[0], formats)
 734             subtitles_src = f[0].get('subtitles_src')
 735             if subtitles_src:
 736                 subtitles.setdefault('en', []).append({'url': subtitles_src})
 737
 738         info_dict = {
 739             'id': video_id,
 740             'formats': formats,
 741             'subtitles': subtitles,
 742         }
 743         process_formats(info_dict)
 744         info_dict.update(extract_metadata(webpage))
 745
 746         return info_dict
 747
 748     def _real_extract(self, url):
 749         video_id = self._match_id(url)
 750
 751         real_url = self._VIDEO_PAGE_TEMPLATE % video_id if url.startswith('facebook:') else url
 752         return self._extract_from_url(real_url, video_id)
 753
 754
 755 class FacebookPluginsVideoIE(InfoExtractor):
 756     _VALID_URL = r'https?://(?:[\w-]+\.)?facebook\.com/plugins/video\.php\?.*?\bhref=(?P<id>https.+)'
 757
 758     _TESTS = [{
 759         'url': 'https://www.facebook.com/plugins/video.php?href=https%3A%2F%2Fwww.facebook.com%2Fgov.sg%2Fvideos%2F10154383743583686%2F&show_text=0&width=560',
 760         'md5': '5954e92cdfe51fe5782ae9bda7058a07',
 761         'info_dict': {
 762             'id': '10154383743583686',
 763             'ext': 'mp4',
 764             # TODO: Fix title, uploader
 765             'title': 'What to do during the haze?',
 766             'uploader': 'Gov.sg',
 767             'upload_date': '20160826',
 768             'timestamp': 1472184808,
 769         },
 770         'add_ie': [FacebookIE.ie_key()],
 771     }, {
 772         'url': 'https://www.facebook.com/plugins/video.php?href=https%3A%2F%2Fwww.facebook.com%2Fvideo.php%3Fv%3D10204634152394104',
 773         'only_matching': True,
 774     }, {
 775         'url': 'https://www.facebook.com/plugins/video.php?href=https://www.facebook.com/gov.sg/videos/10154383743583686/&show_text=0&width=560',
 776         'only_matching': True,
 777     }]
 778
 779     def _real_extract(self, url):
 780         return self.url_result(
 781             compat_urllib_parse_unquote(self._match_id(url)),
 782             FacebookIE.ie_key())
 783
 784
 785 class FacebookRedirectURLIE(InfoExtractor):
 786     IE_DESC = False  # Do not list
 787     _VALID_URL = r'https?://(?:[\w-]+\.)?facebook\.com/flx/warn[/?]'
 788     _TESTS = [{
 789         'url': 'https://www.facebook.com/flx/warn/?h=TAQHsoToz&u=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DpO8h3EaFRdo&s=1',
 790         'info_dict': {
 791             'id': 'pO8h3EaFRdo',
 792             'ext': 'mp4',
 793             'title': 'Tripeo Boiler Room x Dekmantel Festival DJ Set',
 794             'description': 'md5:2d713ccbb45b686a1888397b2c77ca6b',
 795             'channel_id': 'UCGBpxWJr9FNOcFYA5GkKrMg',
 796             'playable_in_embed': True,
 797             'categories': ['Music'],
 798             'channel': 'Boiler Room',
 799             'uploader_id': 'brtvofficial',
 800             'uploader': 'Boiler Room',
 801             'tags': 'count:11',
 802             'duration': 3332,
 803             'live_status': 'not_live',
 804             'thumbnail': 'https://i.ytimg.com/vi/pO8h3EaFRdo/maxresdefault.jpg',
 805             'channel_url': 'https://www.youtube.com/channel/UCGBpxWJr9FNOcFYA5GkKrMg',
 806             'availability': 'public',
 807             'uploader_url': 'http://www.youtube.com/user/brtvofficial',
 808             'upload_date': '20150917',
 809             'age_limit': 0,
 810             'view_count': int,
 811             'like_count': int,
 812         },
 813         'add_ie': ['Youtube'],
 814         'params': {'skip_download': 'Youtube'},
 815     }]
 816
 817     def _real_extract(self, url):
 818         redirect_url = url_or_none(parse_qs(url).get('u', [None])[-1])
 819         if not redirect_url:
 820             raise ExtractorError('Invalid facebook redirect URL', expected=True)
 821         return self.url_result(redirect_url)
 822
 823
 824 class FacebookReelIE(InfoExtractor):
 825     _VALID_URL = r'https?://(?:[\w-]+\.)?facebook\.com/reel/(?P<id>\d+)'
 826     IE_NAME = 'facebook:reel'
 827
 828     _TESTS = [{
 829         'url': 'https://www.facebook.com/reel/1195289147628387',
 830         'md5': 'f13dd37f2633595982db5ed8765474d3',
 831         'info_dict': {
 832             'id': '1195289147628387',
 833             'ext': 'mp4',
 834             'title': 'md5:b05800b5b1ad56c0ca78bd3807b6a61e',
 835             'description': 'md5:22f03309b216ac84720183961441d8db',
 836             'uploader': 'md5:723e6cb3091241160f20b3c5dc282af1',
 837             'uploader_id': '100040874179269',
 838             'duration': 9.579,
 839             'timestamp': 1637502609,
 840             'upload_date': '20211121',
 841             'thumbnail': r're:^https?://.*',
 842         }
 843     }]
 844
 845     def _real_extract(self, url):
 846         video_id = self._match_id(url)
 847         return self.url_result(
 848             f'https://m.facebook.com/watch/?v={video_id}&_rdr', FacebookIE, video_id)