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     format_field,
  20     get_element_by_id,
  21     get_first,
  22     int_or_none,
  23     js_to_json,
  24     merge_dicts,
  25     parse_count,
  26     parse_qs,
  27     qualities,
  28     str_or_none,
  29     traverse_obj,
  30     try_get,
  31     url_or_none,
  32     urlencode_postdata,
  33     urljoin,
  34     variadic,
  35 )
  36
  37
  38 class FacebookIE(InfoExtractor):
  39     _VALID_URL = r'''(?x)
  40                 (?:
  41                     https?://
  42                         (?:[\w-]+\.)?(?:facebook\.com|facebookwkhpilnemxj7asaniu7vnjjbiltxjqhye3mhbshg7kx5tfyd\.onion)/
  43                         (?:[^#]*?\#!/)?
  44                         (?:
  45                             (?:
  46                                 video/video\.php|
  47                                 photo\.php|
  48                                 video\.php|
  49                                 video/embed|
  50                                 story\.php|
  51                                 watch(?:/live)?/?
  52                             )\?(?:.*?)(?:v|video_id|story_fbid)=|
  53                             [^/]+/videos/(?:[^/]+/)?|
  54                             [^/]+/posts/|
  55                             groups/[^/]+/(?:permalink|posts)/|
  56                             watchparty/
  57                         )|
  58                     facebook:
  59                 )
  60                 (?P<id>pfbid[A-Za-z0-9]+|\d+)
  61                 '''
  62     _EMBED_REGEX = [
  63         r'<iframe[^>]+?src=(["\'])(?P<url>https?://www\.facebook\.com/(?:video/embed|plugins/video\.php).+?)\1',
  64         # Facebook API embed https://developers.facebook.com/docs/plugins/embedded-video-player
  65         r'''(?x)<div[^>]+
  66                 class=(?P<q1>[\'"])[^\'"]*\bfb-(?:video|post)\b[^\'"]*(?P=q1)[^>]+
  67                 data-href=(?P<q2>[\'"])(?P<url>(?:https?:)?//(?:www\.)?facebook.com/.+?)(?P=q2)''',
  68     ]
  69     _LOGIN_URL = 'https://www.facebook.com/login.php?next=http%3A%2F%2Ffacebook.com%2Fhome.php&login_attempt=1'
  70     _CHECKPOINT_URL = 'https://www.facebook.com/checkpoint/?next=http%3A%2F%2Ffacebook.com%2Fhome.php&_fb_noscript=1'
  71     _NETRC_MACHINE = 'facebook'
  72     IE_NAME = 'facebook'
  73
  74     _VIDEO_PAGE_TEMPLATE = 'https://www.facebook.com/video/video.php?v=%s'
  75     _VIDEO_PAGE_TAHOE_TEMPLATE = 'https://www.facebook.com/video/tahoe/async/%s/?chain=true&isvideo=true&payloadtype=primary'
  76
  77     _TESTS = [{
  78         'url': 'https://www.facebook.com/radiokicksfm/videos/3676516585958356/',
  79         'info_dict': {
  80             'id': '3676516585958356',
  81             'ext': 'mp4',
  82             'title': 'dr Adam Przygoda',
  83             'description': 'md5:34675bda53336b1d16400265c2bb9b3b',
  84             'uploader': 'RADIO KICKS FM',
  85             'upload_date': '20230818',
  86             'timestamp': 1692346159,
  87             'thumbnail': r're:^https?://.*',
  88             'uploader_id': '100063551323670',
  89             'duration': 3132.184,
  90             'view_count': int,
  91             'concurrent_view_count': 0,
  92         },
  93     }, {
  94         'url': 'https://www.facebook.com/video.php?v=637842556329505&fref=nf',
  95         'md5': '6a40d33c0eccbb1af76cf0485a052659',
  96         'info_dict': {
  97             'id': '637842556329505',
  98             'ext': 'mp4',
  99             'title': 're:Did you know Kei Nishikori is the first Asian man to ever reach a Grand Slam',
 100             'uploader': 'Tennis on Facebook',
 101             'upload_date': '20140908',
 102             'timestamp': 1410199200,
 103         },
 104         'skip': 'Requires logging in',
 105     }, {
 106         # data.video
 107         'url': 'https://www.facebook.com/video.php?v=274175099429670',
 108         'info_dict': {
 109             'id': '274175099429670',
 110             'ext': 'mp4',
 111             'title': 'Asif',
 112             'description': '',
 113             'uploader': 'Asif Nawab Butt',
 114             'upload_date': '20140506',
 115             'timestamp': 1399398998,
 116             'thumbnail': r're:^https?://.*',
 117             'uploader_id': 'pfbid028wxorhX2ErLFJ578N6P3crHD3PHmXTCqCvfBpsnbSLmbokwSY75p5hWBjHGkG4zxl',
 118             'duration': 131.03,
 119             'concurrent_view_count': int,
 120         },
 121     }, {
 122         'note': 'Video with DASH manifest',
 123         'url': 'https://www.facebook.com/video.php?v=957955867617029',
 124         'md5': 'b2c28d528273b323abe5c6ab59f0f030',
 125         'info_dict': {
 126             'id': '957955867617029',
 127             'ext': 'mp4',
 128             'title': 'When you post epic content on instagram.com/433 8 million followers, this is ...',
 129             'uploader': 'Demy de Zeeuw',
 130             'upload_date': '20160110',
 131             'timestamp': 1452431627,
 132         },
 133         'skip': 'Requires logging in',
 134     }, {
 135         'url': 'https://www.facebook.com/maxlayn/posts/10153807558977570',
 136         'md5': '037b1fa7f3c2d02b7a0d7bc16031ecc6',
 137         'info_dict': {
 138             'id': '544765982287235',
 139             'ext': 'mp4',
 140             'title': '"What are you doing running in the snow?"',
 141             'uploader': 'FailArmy',
 142         },
 143         'skip': 'Video gone',
 144     }, {
 145         'url': 'https://m.facebook.com/story.php?story_fbid=1035862816472149&id=116132035111903',
 146         'md5': '1deb90b6ac27f7efcf6d747c8a27f5e3',
 147         'info_dict': {
 148             'id': '1035862816472149',
 149             'ext': 'mp4',
 150             'title': 'What the Flock Is Going On In New Zealand  Credit: ViralHog',
 151             'uploader': 'S. Saint',
 152         },
 153         'skip': 'Video gone',
 154     }, {
 155         'note': 'swf params escaped',
 156         'url': 'https://www.facebook.com/barackobama/posts/10153664894881749',
 157         'md5': '97ba073838964d12c70566e0085c2b91',
 158         'info_dict': {
 159             'id': '10153664894881749',
 160             'ext': 'mp4',
 161             'title': 'Average time to confirm recent Supreme Court nominees: 67 days Longest it\'s t...',
 162             'thumbnail': r're:^https?://.*',
 163             'timestamp': 1456259628,
 164             'upload_date': '20160223',
 165             'uploader': 'Barack Obama',
 166         },
 167         'skip': 'Gif on giphy.com gone',
 168     }, {
 169         # have 1080P, but only up to 720p in swf params
 170         # data.video.story.attachments[].media
 171         'url': 'https://www.facebook.com/cnn/videos/10155529876156509/',
 172         'md5': 'ca63897a90c9452efee5f8c40d080e25',
 173         'info_dict': {
 174             'id': '10155529876156509',
 175             'ext': 'mp4',
 176             'title': 'Holocaust survivor becomes US citizen',
 177             'description': 'She survived the holocaust — and years later, she’s getting her citizenship so she can vote for Hillary Clinton http://cnn.it/2eERh5f',
 178             'timestamp': 1477818095,
 179             'upload_date': '20161030',
 180             'uploader': 'CNN',
 181             'thumbnail': r're:^https?://.*',
 182             'view_count': int,
 183             'uploader_id': '100059479812265',
 184             'concurrent_view_count': int,
 185             'duration': 44.478,
 186         },
 187     }, {
 188         # bigPipe.onPageletArrive ... onPageletArrive pagelet_group_mall
 189         # data.node.comet_sections.content.story.attachments[].style_type_renderer.attachment.media
 190         'url': 'https://www.facebook.com/yaroslav.korpan/videos/1417995061575415/',
 191         'info_dict': {
 192             'id': '1417995061575415',
 193             'ext': 'mp4',
 194             'title': 'Довгоочікуване відео | By Yaroslav - Facebook',
 195             'description': 'Довгоочікуване відео',
 196             'timestamp': 1486648217,
 197             'upload_date': '20170209',
 198             'uploader': 'Yaroslav Korpan',
 199             'uploader_id': 'pfbid06AScABAWcW91qpiuGrLt99Ef9tvwHoXP6t8KeFYEqkSfreMtfa9nTveh8b2ZEVSWl',
 200             'concurrent_view_count': int,
 201             'thumbnail': r're:^https?://.*',
 202             'view_count': int,
 203             'duration': 11736.446,
 204         },
 205         'params': {
 206             'skip_download': True,
 207         },
 208     }, {
 209         # FIXME
 210         'url': 'https://www.facebook.com/LaGuiaDelVaron/posts/1072691702860471',
 211         'info_dict': {
 212             'id': '1072691702860471',
 213             'ext': 'mp4',
 214             'title': 'md5:ae2d22a93fbb12dad20dc393a869739d',
 215             'timestamp': 1477305000,
 216             'upload_date': '20161024',
 217             'uploader': 'La Guía Del Varón',
 218             'thumbnail': r're:^https?://.*',
 219         },
 220         'skip': 'Requires logging in',
 221     }, {
 222         # data.node.comet_sections.content.story.attachments[].style_type_renderer.attachment.media
 223         'url': 'https://www.facebook.com/groups/1024490957622648/permalink/1396382447100162/',
 224         'info_dict': {
 225             'id': '202882990186699',
 226             'ext': 'mp4',
 227             'title': 'birb (O v O") | Hello? Yes your uber ride is here',
 228             'description': 'Hello? Yes your uber ride is here * Jukin Media Verified * Find this video and others like it by visiting...',
 229             'timestamp': 1486035513,
 230             'upload_date': '20170202',
 231             'uploader': 'Elisabeth Ahtn',
 232             'uploader_id': '100013949973717',
 233         },
 234         'skip': 'Requires logging in',
 235     }, {
 236         # data.node.comet_sections.content.story.attachments[].throwbackStyles.attachment_target_renderer.attachment.target.attachments[].styles.attachment.media
 237         'url': 'https://www.facebook.com/groups/1645456212344334/posts/3737828833107051/',
 238         'info_dict': {
 239             'id': '1569199726448814',
 240             'ext': 'mp4',
 241             'title': 'Pence MUST GO!',
 242             'description': 'Vickie Gentry shared a memory.',
 243             'timestamp': 1511548260,
 244             'upload_date': '20171124',
 245             'uploader': 'Vickie Gentry',
 246             'uploader_id': 'pfbid0FuZhHCeWDAxWxEbr3yKPFaRstXvRxgsp9uCPG6GjD4J2AitB35NUAuJ4Q75KcjiDl',
 247             'thumbnail': r're:^https?://.*',
 248             'duration': 148.435,
 249         },
 250     }, {
 251         'url': 'https://www.facebook.com/attn/posts/pfbid0j1Czf2gGDVqeQ8KiMLFm3pWN8GxsQmeRrVhimWDzMuKQoR8r4b1knNsejELmUgyhl',
 252         'info_dict': {
 253             'id': '6968553779868435',
 254             'ext': 'mp4',
 255             'description': 'md5:2f2fcf93e97ac00244fe64521bbdb0cb',
 256             'uploader': 'ATTN:',
 257             'upload_date': '20231207',
 258             'title': 'ATTN:',
 259             'duration': 132.675,
 260             'uploader_id': '100064451419378',
 261             'view_count': int,
 262             'thumbnail': r're:^https?://.*',
 263             'timestamp': 1701975646,
 264         },
 265     }, {
 266         'url': 'https://www.facebook.com/story.php?story_fbid=pfbid0Fnzhm8UuzjBYpPMNFzaSpFE9UmLdU4fJN8qTANi1Dmtj5q7DNrL5NERXfsAzDEV7l&id=100073071055552',
 267         'only_matching': True,
 268     }, {
 269         'url': 'https://www.facebook.com/video.php?v=10204634152394104',
 270         'only_matching': True,
 271     }, {
 272         'url': 'https://www.facebook.com/amogood/videos/1618742068337349/?fref=nf',
 273         'only_matching': True,
 274     }, {
 275         # data.mediaset.currMedia.edges
 276         'url': 'https://www.facebook.com/ChristyClarkForBC/videos/vb.22819070941/10153870694020942/?type=2&theater',
 277         'only_matching': True,
 278     }, {
 279         # data.video.story.attachments[].media
 280         'url': 'facebook:544765982287235',
 281         'only_matching': True,
 282     }, {
 283         # data.node.comet_sections.content.story.attachments[].style_type_renderer.attachment.media
 284         'url': 'https://www.facebook.com/groups/164828000315060/permalink/764967300301124/',
 285         'only_matching': True,
 286     }, {
 287         # data.video.creation_story.attachments[].media
 288         'url': 'https://zh-hk.facebook.com/peoplespower/videos/1135894589806027/',
 289         'only_matching': True,
 290     }, {
 291         # data.video
 292         'url': 'https://www.facebookwkhpilnemxj7asaniu7vnjjbiltxjqhye3mhbshg7kx5tfyd.onion/video.php?v=274175099429670',
 293         'only_matching': True,
 294     }, {
 295         # no title
 296         'url': 'https://www.facebook.com/onlycleverentertainment/videos/1947995502095005/',
 297         'only_matching': True,
 298     }, {
 299         # data.video
 300         'url': 'https://www.facebook.com/WatchESLOne/videos/359649331226507/',
 301         'info_dict': {
 302             'id': '359649331226507',
 303             'ext': 'mp4',
 304             'title': 'Fnatic vs. EG - Group A - Opening Match - ESL One Birmingham Day 1',
 305             'description': '#ESLOne VoD - Birmingham Finals Day#1 Fnatic vs. @Evil Geniuses',
 306             'timestamp': 1527084179,
 307             'upload_date': '20180523',
 308             'uploader': 'ESL One Dota 2',
 309             'uploader_id': '100066514874195',
 310             'duration': 4524.212,
 311             'view_count': int,
 312             'thumbnail': r're:^https?://.*',
 313             'concurrent_view_count': int,
 314         },
 315         'params': {
 316             'skip_download': True,
 317         },
 318     }, {
 319         # data.node.comet_sections.content.story.attachments[].style_type_renderer.attachment.all_subattachments.nodes[].media
 320         'url': 'https://www.facebook.com/100033620354545/videos/106560053808006/',
 321         'info_dict': {
 322             'id': '106560053808006',
 323             'ext': 'mp4',
 324             'title': 'Josef',
 325             'thumbnail': r're:^https?://.*',
 326             'concurrent_view_count': int,
 327             'uploader_id': 'pfbid0cibUN6tV7DYgdbJdsUFN46wc4jKpVSPAvJQhFofGqBGmVn3V3JtAs2tfUwziw2hUl',
 328             'timestamp': 1549275572,
 329             'duration': 3.413,
 330             'uploader': 'Josef Novak',
 331             'description': '',
 332             'upload_date': '20190204',
 333         },
 334     }, {
 335         # data.video.story.attachments[].media
 336         'url': 'https://www.facebook.com/watch/?v=647537299265662',
 337         'only_matching': True,
 338     }, {
 339         # FIXME: https://github.com/yt-dlp/yt-dlp/issues/542
 340         # data.node.comet_sections.content.story.attachments[].style_type_renderer.attachment.all_subattachments.nodes[].media
 341         'url': 'https://www.facebook.com/PankajShahLondon/posts/10157667649866271',
 342         'info_dict': {
 343             'id': '10157667649866271',
 344         },
 345         'playlist_count': 3,
 346         'skip': 'Requires logging in',
 347     }, {
 348         # data.nodes[].comet_sections.content.story.attachments[].style_type_renderer.attachment.media
 349         'url': 'https://m.facebook.com/Alliance.Police.Department/posts/4048563708499330',
 350         'info_dict': {
 351             'id': '117576630041613',
 352             'ext': 'mp4',
 353             # TODO: title can be extracted from video page
 354             'title': 'Facebook video #117576630041613',
 355             'uploader_id': '189393014416438',
 356             'upload_date': '20201123',
 357             'timestamp': 1606162592,
 358         },
 359         'skip': 'Requires logging in',
 360     }, {
 361         # node.comet_sections.content.story.attached_story.attachments.style_type_renderer.attachment.media
 362         'url': 'https://www.facebook.com/groups/ateistiskselskab/permalink/10154930137678856/',
 363         'info_dict': {
 364             'id': '211567722618337',
 365             'ext': 'mp4',
 366             'title': 'Facebook video #211567722618337',
 367             'uploader_id': '127875227654254',
 368             'upload_date': '20161122',
 369             'timestamp': 1479793574,
 370         },
 371         'skip': 'No video',
 372     }, {
 373         # data.video.creation_story.attachments[].media
 374         'url': 'https://www.facebook.com/watch/live/?v=1823658634322275',
 375         'only_matching': True,
 376     }, {
 377         'url': 'https://www.facebook.com/watchparty/211641140192478',
 378         'info_dict': {
 379             'id': '211641140192478',
 380         },
 381         'playlist_count': 1,
 382         'skip': 'Requires logging in',
 383     }]
 384     _SUPPORTED_PAGLETS_REGEX = r'(?:pagelet_group_mall|permalink_video_pagelet|hyperfeed_story_id_[0-9a-f]+)'
 385     _api_config = {
 386         'graphURI': '/api/graphql/'
 387     }
 388
 389     def _perform_login(self, username, password):
 390         login_page_req = Request(self._LOGIN_URL)
 391         self._set_cookie('facebook.com', 'locale', 'en_US')
 392         login_page = self._download_webpage(login_page_req, None,
 393                                             note='Downloading login page',
 394                                             errnote='Unable to download login page')
 395         lsd = self._search_regex(
 396             r'<input type="hidden" name="lsd" value="([^"]*)"',
 397             login_page, 'lsd')
 398         lgnrnd = self._search_regex(r'name="lgnrnd" value="([^"]*?)"', login_page, 'lgnrnd')
 399
 400         login_form = {
 401             'email': username,
 402             'pass': password,
 403             'lsd': lsd,
 404             'lgnrnd': lgnrnd,
 405             'next': 'http://facebook.com/home.php',
 406             'default_persistent': '0',
 407             'legacy_return': '1',
 408             'timezone': '-60',
 409             'trynum': '1',
 410         }
 411         request = Request(self._LOGIN_URL, urlencode_postdata(login_form))
 412         request.headers['Content-Type'] = 'application/x-www-form-urlencoded'
 413         try:
 414             login_results = self._download_webpage(request, None,
 415                                                    note='Logging in', errnote='unable to fetch login page')
 416             if re.search(r'<form(.*)name="login"(.*)</form>', login_results) is not None:
 417                 error = self._html_search_regex(
 418                     r'(?s)<div[^>]+class=(["\']).*?login_error_box.*?\1[^>]*><div[^>]*>.*?</div><div[^>]*>(?P<error>.+?)</div>',
 419                     login_results, 'login error', default=None, group='error')
 420                 if error:
 421                     raise ExtractorError('Unable to login: %s' % error, expected=True)
 422                 self.report_warning('unable to log in: bad username/password, or exceeded login rate limit (~3/min). Check credentials or wait.')
 423                 return
 424
 425             fb_dtsg = self._search_regex(
 426                 r'name="fb_dtsg" value="(.+?)"', login_results, 'fb_dtsg', default=None)
 427             h = self._search_regex(
 428                 r'name="h"\s+(?:\w+="[^"]+"\s+)*?value="([^"]+)"', login_results, 'h', default=None)
 429
 430             if not fb_dtsg or not h:
 431                 return
 432
 433             check_form = {
 434                 'fb_dtsg': fb_dtsg,
 435                 'h': h,
 436                 'name_action_selected': 'dont_save',
 437             }
 438             check_req = Request(self._CHECKPOINT_URL, urlencode_postdata(check_form))
 439             check_req.headers['Content-Type'] = 'application/x-www-form-urlencoded'
 440             check_response = self._download_webpage(check_req, None,
 441                                                     note='Confirming login')
 442             if re.search(r'id="checkpointSubmitButton"', check_response) is not None:
 443                 self.report_warning('Unable to confirm login, you have to login in your browser and authorize the login.')
 444         except network_exceptions as err:
 445             self.report_warning('unable to log in: %s' % error_to_compat_str(err))
 446             return
 447
 448     def _extract_from_url(self, url, video_id):
 449         webpage = self._download_webpage(
 450             url.replace('://m.facebook.com/', '://www.facebook.com/'), video_id)
 451
 452         def extract_metadata(webpage):
 453             post_data = [self._parse_json(j, video_id, fatal=False) for j in re.findall(
 454                 r'data-sjs>({.*?ScheduledServerJS.*?})</script>', webpage)]
 455             post = traverse_obj(post_data, (
 456                 ..., 'require', ..., ..., ..., '__bbox', 'require', ..., ..., ..., '__bbox', 'result', 'data'), expected_type=dict) or []
 457
 458             automatic_captions, subtitles = {}, {}
 459             subs_data = traverse_obj(post, (..., 'video', ..., 'attachments', ..., lambda k, v: (
 460                 k == 'media' and str(v['id']) == video_id and v['__typename'] == 'Video')))
 461             is_video_broadcast = get_first(subs_data, 'is_video_broadcast', expected_type=bool)
 462             captions = get_first(subs_data, 'video_available_captions_locales', 'captions_url')
 463             if url_or_none(captions):  # if subs_data only had a 'captions_url'
 464                 locale = self._html_search_meta(['og:locale', 'twitter:locale'], webpage, 'locale', default='en_US')
 465                 subtitles[locale] = [{'url': captions}]
 466             # or else subs_data had 'video_available_captions_locales', a list of dicts
 467             for caption in traverse_obj(captions, (
 468                 {lambda x: sorted(x, key=lambda c: c['locale'])}, lambda _, v: v['captions_url'])
 469             ):
 470                 lang = caption.get('localized_language') or ''
 471                 subs = {
 472                     'url': caption['captions_url'],
 473                     'name': format_field(caption, 'localized_country', f'{lang} (%s)', default=lang),
 474                 }
 475                 if caption.get('localized_creation_method') or is_video_broadcast:
 476                     automatic_captions.setdefault(caption['locale'], []).append(subs)
 477                 else:
 478                     subtitles.setdefault(caption['locale'], []).append(subs)
 479
 480             media = traverse_obj(post, (..., 'attachments', ..., lambda k, v: (
 481                 k == 'media' and str(v['id']) == video_id and v['__typename'] == 'Video')), expected_type=dict)
 482             title = get_first(media, ('title', 'text'))
 483             description = get_first(media, ('creation_story', 'comet_sections', 'message', 'story', 'message', 'text'))
 484             uploader_data = (
 485                 get_first(media, ('owner', {dict}))
 486                 or get_first(post, (..., 'video', lambda k, v: k == 'owner' and v['name']))
 487                 or get_first(post, ('node', 'actors', ..., {dict})) or {})
 488
 489             page_title = title or self._html_search_regex((
 490                 r'<h2\s+[^>]*class="uiHeaderTitle"[^>]*>(?P<content>[^<]*)</h2>',
 491                 r'(?s)<span class="fbPhotosPhotoCaption".*?id="fbPhotoPageCaption"><span class="hasCaption">(?P<content>.*?)</span>',
 492                 self._meta_regex('og:title'), self._meta_regex('twitter:title'), r'<title>(?P<content>.+?)</title>'
 493             ), webpage, 'title', default=None, group='content')
 494             description = description or self._html_search_meta(
 495                 ['description', 'og:description', 'twitter:description'],
 496                 webpage, 'description', default=None)
 497             uploader = uploader_data.get('name') or (
 498                 clean_html(get_element_by_id('fbPhotoPageAuthorName', webpage))
 499                 or self._search_regex(
 500                     (r'ownerName\s*:\s*"([^"]+)"', *self._og_regexes('title')), webpage, 'uploader', fatal=False))
 501
 502             timestamp = int_or_none(self._search_regex(
 503                 r'<abbr[^>]+data-utime=["\'](\d+)', webpage,
 504                 'timestamp', default=None))
 505             thumbnail = self._html_search_meta(
 506                 ['og:image', 'twitter:image'], webpage, 'thumbnail', default=None)
 507             # some webpages contain unretrievable thumbnail urls
 508             # like https://lookaside.fbsbx.com/lookaside/crawler/media/?media_id=10155168902769113&get_thumbnail=1
 509             # in https://www.facebook.com/yaroslav.korpan/videos/1417995061575415/
 510             if thumbnail and not re.search(r'\.(?:jpg|png)', thumbnail):
 511                 thumbnail = None
 512             info_dict = {
 513                 'description': description,
 514                 'uploader': uploader,
 515                 'uploader_id': uploader_data.get('id'),
 516                 'timestamp': timestamp,
 517                 'thumbnail': thumbnail,
 518                 'view_count': parse_count(self._search_regex(
 519                     (r'\bviewCount\s*:\s*["\']([\d,.]+)', r'video_view_count["\']\s*:\s*(\d+)',),
 520                     webpage, 'view count', default=None)),
 521                 'concurrent_view_count': get_first(post, (
 522                     ('video', (..., ..., 'attachments', ..., 'media')), 'liveViewerCount', {int_or_none})),
 523                 'automatic_captions': automatic_captions,
 524                 'subtitles': subtitles,
 525             }
 526
 527             info_json_ld = self._search_json_ld(webpage, video_id, default={})
 528             info_json_ld['title'] = (re.sub(r'\s*\|\s*Facebook$', '', title or info_json_ld.get('title') or page_title or '')
 529                                      or (description or '').replace('\n', ' ') or f'Facebook video #{video_id}')
 530             return merge_dicts(info_json_ld, info_dict)
 531
 532         video_data = None
 533
 534         def extract_video_data(instances):
 535             video_data = []
 536             for item in instances:
 537                 if try_get(item, lambda x: x[1][0]) == 'VideoConfig':
 538                     video_item = item[2][0]
 539                     if video_item.get('video_id'):
 540                         video_data.append(video_item['videoData'])
 541             return video_data
 542
 543         server_js_data = self._parse_json(self._search_regex(
 544             [r'handleServerJS\(({.+})(?:\);|,")', r'\bs\.handle\(({.+?})\);'],
 545             webpage, 'server js data', default='{}'), video_id, fatal=False)
 546
 547         if server_js_data:
 548             video_data = extract_video_data(server_js_data.get('instances', []))
 549
 550         def extract_from_jsmods_instances(js_data):
 551             if js_data:
 552                 return extract_video_data(try_get(
 553                     js_data, lambda x: x['jsmods']['instances'], list) or [])
 554
 555         def extract_dash_manifest(video, formats):
 556             dash_manifest = video.get('dash_manifest')
 557             if dash_manifest:
 558                 formats.extend(self._parse_mpd_formats(
 559                     compat_etree_fromstring(urllib.parse.unquote_plus(dash_manifest)),
 560                     mpd_url=video.get('dash_manifest_url')))
 561
 562         def process_formats(info):
 563             # Downloads with browser's User-Agent are rate limited. Working around
 564             # with non-browser User-Agent.
 565             for f in info['formats']:
 566                 f.setdefault('http_headers', {})['User-Agent'] = 'facebookexternalhit/1.1'
 567
 568         def extract_relay_data(_filter):
 569             return self._parse_json(self._search_regex(
 570                 r'data-sjs>({.*?%s.*?})</script>' % _filter,
 571                 webpage, 'replay data', default='{}'), video_id, fatal=False) or {}
 572
 573         def extract_relay_prefetched_data(_filter):
 574             return traverse_obj(extract_relay_data(_filter), (
 575                 'require', (None, (..., ..., ..., '__bbox', 'require')),
 576                 lambda _, v: 'RelayPrefetchedStreamCache' in v, ..., ...,
 577                 '__bbox', 'result', 'data', {dict}), get_all=False) or {}
 578
 579         if not video_data:
 580             server_js_data = self._parse_json(self._search_regex([
 581                 r'bigPipe\.onPageletArrive\(({.+?})\)\s*;\s*}\s*\)\s*,\s*["\']onPageletArrive\s+' + self._SUPPORTED_PAGLETS_REGEX,
 582                 r'bigPipe\.onPageletArrive\(({.*?id\s*:\s*"%s".*?})\);' % self._SUPPORTED_PAGLETS_REGEX
 583             ], webpage, 'js data', default='{}'), video_id, js_to_json, False)
 584             video_data = extract_from_jsmods_instances(server_js_data)
 585
 586         if not video_data:
 587             data = extract_relay_prefetched_data(
 588                 r'"(?:dash_manifest|playable_url(?:_quality_hd)?)')
 589             if data:
 590                 entries = []
 591
 592                 def parse_graphql_video(video):
 593                     v_id = video.get('videoId') or video.get('id') or video_id
 594                     reel_info = traverse_obj(
 595                         video, ('creation_story', 'short_form_video_context', 'playback_video', {dict}))
 596                     if reel_info:
 597                         video = video['creation_story']
 598                         video['owner'] = traverse_obj(video, ('short_form_video_context', 'video_owner'))
 599                         video.update(reel_info)
 600                     formats = []
 601                     q = qualities(['sd', 'hd'])
 602                     for key, format_id in (('playable_url', 'sd'), ('playable_url_quality_hd', 'hd'),
 603                                            ('playable_url_dash', ''), ('browser_native_hd_url', 'hd'),
 604                                            ('browser_native_sd_url', 'sd')):
 605                         playable_url = video.get(key)
 606                         if not playable_url:
 607                             continue
 608                         if determine_ext(playable_url) == 'mpd':
 609                             formats.extend(self._extract_mpd_formats(playable_url, video_id))
 610                         else:
 611                             formats.append({
 612                                 'format_id': format_id,
 613                                 # sd, hd formats w/o resolution info should be deprioritized below DASH
 614                                 'quality': q(format_id) - 3,
 615                                 'url': playable_url,
 616                             })
 617                     extract_dash_manifest(video, formats)
 618                     info = {
 619                         'id': v_id,
 620                         'formats': formats,
 621                         'thumbnail': traverse_obj(
 622                             video, ('thumbnailImage', 'uri'), ('preferred_thumbnail', 'image', 'uri')),
 623                         'uploader_id': traverse_obj(video, ('owner', 'id', {str_or_none})),
 624                         'timestamp': traverse_obj(video, 'publish_time', 'creation_time', expected_type=int_or_none),
 625                         'duration': (float_or_none(video.get('playable_duration_in_ms'), 1000)
 626                                      or float_or_none(video.get('length_in_second'))),
 627                     }
 628                     process_formats(info)
 629                     description = try_get(video, lambda x: x['savable_description']['text'])
 630                     title = video.get('name')
 631                     if title:
 632                         info.update({
 633                             'title': title,
 634                             'description': description,
 635                         })
 636                     else:
 637                         info['title'] = description or 'Facebook video #%s' % v_id
 638                     entries.append(info)
 639
 640                 def parse_attachment(attachment, key='media'):
 641                     media = attachment.get(key) or {}
 642                     if media.get('__typename') == 'Video':
 643                         return parse_graphql_video(media)
 644
 645                 nodes = variadic(traverse_obj(data, 'nodes', 'node') or [])
 646                 attachments = traverse_obj(nodes, (
 647                     ..., 'comet_sections', 'content', 'story', (None, 'attached_story'), 'attachments',
 648                     ..., ('styles', 'style_type_renderer', ('throwbackStyles', 'attachment_target_renderer')),
 649                     'attachment', {dict}))
 650                 for attachment in attachments:
 651                     ns = traverse_obj(attachment, ('all_subattachments', 'nodes', ..., {dict}),
 652                                       ('target', 'attachments', ..., 'styles', 'attachment', {dict}))
 653                     for n in ns:
 654                         parse_attachment(n)
 655                     parse_attachment(attachment)
 656
 657                 edges = try_get(data, lambda x: x['mediaset']['currMedia']['edges'], list) or []
 658                 for edge in edges:
 659                     parse_attachment(edge, key='node')
 660
 661                 video = data.get('video') or {}
 662                 if video:
 663                     attachments = try_get(video, [
 664                         lambda x: x['story']['attachments'],
 665                         lambda x: x['creation_story']['attachments']
 666                     ], list) or []
 667                     for attachment in attachments:
 668                         parse_attachment(attachment)
 669                     if not entries:
 670                         parse_graphql_video(video)
 671
 672                 if len(entries) > 1:
 673                     return self.playlist_result(entries, video_id)
 674
 675                 video_info = entries[0] if entries else {'id': video_id}
 676                 webpage_info = extract_metadata(webpage)
 677                 # honor precise duration in video info
 678                 if video_info.get('duration'):
 679                     webpage_info['duration'] = video_info['duration']
 680                 return merge_dicts(webpage_info, video_info)
 681
 682         if not video_data:
 683             m_msg = re.search(r'class="[^"]*uiInterstitialContent[^"]*"><div>(.*?)</div>', webpage)
 684             if m_msg is not None:
 685                 raise ExtractorError(
 686                     'The video is not available, Facebook said: "%s"' % m_msg.group(1),
 687                     expected=True)
 688             elif any(p in webpage for p in (
 689                     '>You must log in to continue',
 690                     'id="login_form"',
 691                     'id="loginbutton"')):
 692                 self.raise_login_required()
 693
 694         if not video_data and '/watchparty/' in url:
 695             post_data = {
 696                 'doc_id': 3731964053542869,
 697                 'variables': json.dumps({
 698                     'livingRoomID': video_id,
 699                 }),
 700             }
 701
 702             prefetched_data = extract_relay_prefetched_data(r'"login_data"\s*:\s*{')
 703             if prefetched_data:
 704                 lsd = try_get(prefetched_data, lambda x: x['login_data']['lsd'], dict)
 705                 if lsd:
 706                     post_data[lsd['name']] = lsd['value']
 707
 708             relay_data = extract_relay_data(r'\[\s*"RelayAPIConfigDefaults"\s*,')
 709             for define in (relay_data.get('define') or []):
 710                 if define[0] == 'RelayAPIConfigDefaults':
 711                     self._api_config = define[2]
 712
 713             living_room = self._download_json(
 714                 urljoin(url, self._api_config['graphURI']), video_id,
 715                 data=urlencode_postdata(post_data))['data']['living_room']
 716
 717             entries = []
 718             for edge in (try_get(living_room, lambda x: x['recap']['watched_content']['edges']) or []):
 719                 video = try_get(edge, lambda x: x['node']['video']) or {}
 720                 v_id = video.get('id')
 721                 if not v_id:
 722                     continue
 723                 v_id = compat_str(v_id)
 724                 entries.append(self.url_result(
 725                     self._VIDEO_PAGE_TEMPLATE % v_id,
 726                     self.ie_key(), v_id, video.get('name')))
 727
 728             return self.playlist_result(entries, video_id)
 729
 730         if not video_data:
 731             # Video info not in first request, do a secondary request using
 732             # tahoe player specific URL
 733             tahoe_data = self._download_webpage(
 734                 self._VIDEO_PAGE_TAHOE_TEMPLATE % video_id, video_id,
 735                 data=urlencode_postdata({
 736                     '__a': 1,
 737                     '__pc': self._search_regex(
 738                         r'pkg_cohort["\']\s*:\s*["\'](.+?)["\']', webpage,
 739                         'pkg cohort', default='PHASED:DEFAULT'),
 740                     '__rev': self._search_regex(
 741                         r'client_revision["\']\s*:\s*(\d+),', webpage,
 742                         'client revision', default='3944515'),
 743                     'fb_dtsg': self._search_regex(
 744                         r'"DTSGInitialData"\s*,\s*\[\]\s*,\s*{\s*"token"\s*:\s*"([^"]+)"',
 745                         webpage, 'dtsg token', default=''),
 746                 }),
 747                 headers={
 748                     'Content-Type': 'application/x-www-form-urlencoded',
 749                 })
 750             tahoe_js_data = self._parse_json(
 751                 self._search_regex(
 752                     r'for\s+\(\s*;\s*;\s*\)\s*;(.+)', tahoe_data,
 753                     'tahoe js data', default='{}'),
 754                 video_id, fatal=False)
 755             video_data = extract_from_jsmods_instances(tahoe_js_data)
 756
 757         if not video_data:
 758             raise ExtractorError('Cannot parse data')
 759
 760         if len(video_data) > 1:
 761             entries = []
 762             for v in video_data:
 763                 video_url = v[0].get('video_url')
 764                 if not video_url:
 765                     continue
 766                 entries.append(self.url_result(urljoin(
 767                     url, video_url), self.ie_key(), v[0].get('video_id')))
 768             return self.playlist_result(entries, video_id)
 769         video_data = video_data[0]
 770
 771         formats = []
 772         subtitles = {}
 773         for f in video_data:
 774             format_id = f['stream_type']
 775             if f and isinstance(f, dict):
 776                 f = [f]
 777             if not f or not isinstance(f, list):
 778                 continue
 779             for quality in ('sd', 'hd'):
 780                 for src_type in ('src', 'src_no_ratelimit'):
 781                     src = f[0].get('%s_%s' % (quality, src_type))
 782                     if src:
 783                         # sd, hd formats w/o resolution info should be deprioritized below DASH
 784                         # TODO: investigate if progressive or src formats still exist
 785                         preference = -10 if format_id == 'progressive' else -3
 786                         if quality == 'hd':
 787                             preference += 1
 788                         formats.append({
 789                             'format_id': '%s_%s_%s' % (format_id, quality, src_type),
 790                             'url': src,
 791                             'quality': preference,
 792                             'height': 720 if quality == 'hd' else None
 793                         })
 794             extract_dash_manifest(f[0], formats)
 795             subtitles_src = f[0].get('subtitles_src')
 796             if subtitles_src:
 797                 subtitles.setdefault('en', []).append({'url': subtitles_src})
 798
 799         info_dict = {
 800             'id': video_id,
 801             'formats': formats,
 802             'subtitles': subtitles,
 803         }
 804         process_formats(info_dict)
 805         info_dict.update(extract_metadata(webpage))
 806
 807         return info_dict
 808
 809     def _real_extract(self, url):
 810         video_id = self._match_id(url)
 811
 812         real_url = self._VIDEO_PAGE_TEMPLATE % video_id if url.startswith('facebook:') else url
 813         return self._extract_from_url(real_url, video_id)
 814
 815
 816 class FacebookPluginsVideoIE(InfoExtractor):
 817     _VALID_URL = r'https?://(?:[\w-]+\.)?facebook\.com/plugins/video\.php\?.*?\bhref=(?P<id>https.+)'
 818
 819     _TESTS = [{
 820         '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',
 821         'md5': '5954e92cdfe51fe5782ae9bda7058a07',
 822         'info_dict': {
 823             'id': '10154383743583686',
 824             'ext': 'mp4',
 825             # TODO: Fix title, uploader
 826             'title': 'What to do during the haze?',
 827             'uploader': 'Gov.sg',
 828             'upload_date': '20160826',
 829             'timestamp': 1472184808,
 830         },
 831         'add_ie': [FacebookIE.ie_key()],
 832     }, {
 833         'url': 'https://www.facebook.com/plugins/video.php?href=https%3A%2F%2Fwww.facebook.com%2Fvideo.php%3Fv%3D10204634152394104',
 834         'only_matching': True,
 835     }, {
 836         'url': 'https://www.facebook.com/plugins/video.php?href=https://www.facebook.com/gov.sg/videos/10154383743583686/&show_text=0&width=560',
 837         'only_matching': True,
 838     }]
 839
 840     def _real_extract(self, url):
 841         return self.url_result(
 842             compat_urllib_parse_unquote(self._match_id(url)),
 843             FacebookIE.ie_key())
 844
 845
 846 class FacebookRedirectURLIE(InfoExtractor):
 847     IE_DESC = False  # Do not list
 848     _VALID_URL = r'https?://(?:[\w-]+\.)?facebook\.com/flx/warn[/?]'
 849     _TESTS = [{
 850         'url': 'https://www.facebook.com/flx/warn/?h=TAQHsoToz&u=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DpO8h3EaFRdo&s=1',
 851         'info_dict': {
 852             'id': 'pO8h3EaFRdo',
 853             'ext': 'mp4',
 854             'title': 'Tripeo Boiler Room x Dekmantel Festival DJ Set',
 855             'description': 'md5:2d713ccbb45b686a1888397b2c77ca6b',
 856             'channel_id': 'UCGBpxWJr9FNOcFYA5GkKrMg',
 857             'playable_in_embed': True,
 858             'categories': ['Music'],
 859             'channel': 'Boiler Room',
 860             'uploader_id': 'brtvofficial',
 861             'uploader': 'Boiler Room',
 862             'tags': 'count:11',
 863             'duration': 3332,
 864             'live_status': 'not_live',
 865             'thumbnail': 'https://i.ytimg.com/vi/pO8h3EaFRdo/maxresdefault.jpg',
 866             'channel_url': 'https://www.youtube.com/channel/UCGBpxWJr9FNOcFYA5GkKrMg',
 867             'availability': 'public',
 868             'uploader_url': 'http://www.youtube.com/user/brtvofficial',
 869             'upload_date': '20150917',
 870             'age_limit': 0,
 871             'view_count': int,
 872             'like_count': int,
 873         },
 874         'add_ie': ['Youtube'],
 875         'params': {'skip_download': 'Youtube'},
 876     }]
 877
 878     def _real_extract(self, url):
 879         redirect_url = url_or_none(parse_qs(url).get('u', [None])[-1])
 880         if not redirect_url:
 881             raise ExtractorError('Invalid facebook redirect URL', expected=True)
 882         return self.url_result(redirect_url)
 883
 884
 885 class FacebookReelIE(InfoExtractor):
 886     _VALID_URL = r'https?://(?:[\w-]+\.)?facebook\.com/reel/(?P<id>\d+)'
 887     IE_NAME = 'facebook:reel'
 888
 889     _TESTS = [{
 890         'url': 'https://www.facebook.com/reel/1195289147628387',
 891         'md5': 'f13dd37f2633595982db5ed8765474d3',
 892         'info_dict': {
 893             'id': '1195289147628387',
 894             'ext': 'mp4',
 895             'title': 'md5:b05800b5b1ad56c0ca78bd3807b6a61e',
 896             'description': 'md5:22f03309b216ac84720183961441d8db',
 897             'uploader': 'md5:723e6cb3091241160f20b3c5dc282af1',
 898             'uploader_id': '100040874179269',
 899             'duration': 9.579,
 900             'timestamp': 1637502609,
 901             'upload_date': '20211121',
 902             'thumbnail': r're:^https?://.*',
 903         }
 904     }]
 905
 906     def _real_extract(self, url):
 907         video_id = self._match_id(url)
 908         return self.url_result(
 909             f'https://m.facebook.com/watch/?v={video_id}&_rdr', FacebookIE, video_id)