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 compat_etree_fromstring
   7 from ..networking import Request
   8 from ..networking.exceptions import network_exceptions
   9 from ..utils import (
  10     ExtractorError,
  11     clean_html,
  12     determine_ext,
  13     float_or_none,
  14     format_field,
  15     get_element_by_id,
  16     get_first,
  17     int_or_none,
  18     join_nonempty,
  19     js_to_json,
  20     merge_dicts,
  21     parse_count,
  22     parse_qs,
  23     qualities,
  24     str_or_none,
  25     traverse_obj,
  26     try_get,
  27     url_or_none,
  28     urlencode_postdata,
  29     urljoin,
  30     variadic,
  31 )
  32
  33
  34 class FacebookIE(InfoExtractor):
  35     _VALID_URL = r'''(?x)
  36                 (?:
  37                     https?://
  38                         (?:[\w-]+\.)?(?:facebook\.com|facebookwkhpilnemxj7asaniu7vnjjbiltxjqhye3mhbshg7kx5tfyd\.onion)/
  39                         (?:[^#]*?\#!/)?
  40                         (?:
  41                             (?:
  42                                 permalink\.php|
  43                                 video/video\.php|
  44                                 photo\.php|
  45                                 video\.php|
  46                                 video/embed|
  47                                 story\.php|
  48                                 watch(?:/live)?/?
  49                             )\?(?:.*?)(?:v|video_id|story_fbid)=|
  50                             [^/]+/videos/(?:[^/]+/)?|
  51                             [^/]+/posts/|
  52                             events/(?:[^/]+/)?|
  53                             groups/[^/]+/(?:permalink|posts)/(?:[\da-f]+/)?|
  54                             watchparty/
  55                         )|
  56                     facebook:
  57                 )
  58                 (?P<id>pfbid[A-Za-z0-9]+|\d+)
  59                 '''
  60     _EMBED_REGEX = [
  61         r'<iframe[^>]+?src=(["\'])(?P<url>https?://www\.facebook\.com/(?:video/embed|plugins/video\.php).+?)\1',
  62         # Facebook API embed https://developers.facebook.com/docs/plugins/embedded-video-player
  63         r'''(?x)<div[^>]+
  64                 class=(?P<q1>[\'"])[^\'"]*\bfb-(?:video|post)\b[^\'"]*(?P=q1)[^>]+
  65                 data-href=(?P<q2>[\'"])(?P<url>(?:https?:)?//(?:www\.)?facebook.com/.+?)(?P=q2)''',
  66     ]
  67     _LOGIN_URL = 'https://www.facebook.com/login.php?next=http%3A%2F%2Ffacebook.com%2Fhome.php&login_attempt=1'
  68     _CHECKPOINT_URL = 'https://www.facebook.com/checkpoint/?next=http%3A%2F%2Ffacebook.com%2Fhome.php&_fb_noscript=1'
  69     _NETRC_MACHINE = 'facebook'
  70     IE_NAME = 'facebook'
  71
  72     _VIDEO_PAGE_TEMPLATE = 'https://www.facebook.com/video/video.php?v=%s'
  73     _VIDEO_PAGE_TAHOE_TEMPLATE = 'https://www.facebook.com/video/tahoe/async/%s/?chain=true&isvideo=true&payloadtype=primary'
  74
  75     _TESTS = [{
  76         'url': 'https://www.facebook.com/radiokicksfm/videos/3676516585958356/',
  77         'info_dict': {
  78             'id': '3676516585958356',
  79             'ext': 'mp4',
  80             'title': 'dr Adam Przygoda',
  81             'description': 'md5:34675bda53336b1d16400265c2bb9b3b',
  82             'uploader': 'RADIO KICKS FM',
  83             'upload_date': '20230818',
  84             'timestamp': 1692346159,
  85             'thumbnail': r're:^https?://.*',
  86             'uploader_id': '100063551323670',
  87             'duration': 3133.583,
  88             'view_count': int,
  89             'concurrent_view_count': 0,
  90         },
  91     }, {
  92         'url': 'https://www.facebook.com/video.php?v=637842556329505&fref=nf',
  93         'md5': '6a40d33c0eccbb1af76cf0485a052659',
  94         'info_dict': {
  95             'id': '637842556329505',
  96             'ext': 'mp4',
  97             'title': 're:Did you know Kei Nishikori is the first Asian man to ever reach a Grand Slam',
  98             'uploader': 'Tennis on Facebook',
  99             'upload_date': '20140908',
 100             'timestamp': 1410199200,
 101         },
 102         'skip': 'Requires logging in',
 103     }, {
 104         # data.video
 105         'url': 'https://www.facebook.com/video.php?v=274175099429670',
 106         'info_dict': {
 107             'id': '274175099429670',
 108             'ext': 'mp4',
 109             'title': 'Asif',
 110             'description': '',
 111             'uploader': 'Asif Nawab Butt',
 112             'upload_date': '20140506',
 113             'timestamp': 1399398998,
 114             'thumbnail': r're:^https?://.*',
 115             'uploader_id': 'pfbid05AzrFTXgY37tqwaSgbFTTEpCLBjjEJHkigogwGiRPtKEpAsJYJpzE94H1RxYXWEtl',
 116             'duration': 131.03,
 117             'concurrent_view_count': int,
 118             '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': '1659aa21fb3dd1585874f668e81a72c8',
 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': 44.181,
 185         },
 186     }, {
 187         # FIXME: unable to extract uploader, no formats found
 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: Cannot parse data error
 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': 'pfbid0FkkycT95ySNNyfCw4Cho6u5G7WbbZEcxT496Hq8rtx1K3LcTCATpR3wnyYhmyGC5l',
 247             'thumbnail': r're:^https?://.*',
 248             'duration': 148.224,
 249         },
 250     }, {
 251         # data.node.comet_sections.content.story.attachments[].styles.attachment.media
 252         'url': 'https://www.facebook.com/attn/posts/pfbid0j1Czf2gGDVqeQ8KiMLFm3pWN8GxsQmeRrVhimWDzMuKQoR8r4b1knNsejELmUgyhl',
 253         'info_dict': {
 254             'id': '6968553779868435',
 255             'ext': 'mp4',
 256             'description': 'md5:2f2fcf93e97ac00244fe64521bbdb0cb',
 257             'uploader': 'ATTN:',
 258             'upload_date': '20231207',
 259             'title': 'ATTN:',
 260             'duration': 132.675,
 261             'uploader_id': '100064451419378',
 262             'view_count': int,
 263             'thumbnail': r're:^https?://.*',
 264             'timestamp': 1701975646,
 265         },
 266     }, {
 267         # data.node.comet_sections.content.story.attachments[].styles.attachment.media
 268         'url': 'https://www.facebook.com/permalink.php?story_fbid=pfbid0fqQuVEQyXRa9Dp4RcaTR14KHU3uULHV1EK7eckNXSH63JMuoALsAvVCJ97zAGitil&id=100068861234290',
 269         'info_dict': {
 270             'id': '270103405756416',
 271             'ext': 'mp4',
 272             'title': 'Lela Evans',
 273             'description': 'Today Makkovik\'s own Pilot Mandy Smith made her inaugural landing on the airstrip in her hometown. What a proud moment as we all cheered and...',
 274             'thumbnail': r're:^https?://.*',
 275             'uploader': 'Lela Evans',
 276             'uploader_id': 'pfbid0swT2y7t6TAsZVBvcyeYPdhTMefGaS26mzUwML3vd1ma6ndGZKxsyS4Ssu3jitZLXl',
 277             'upload_date': '20231228',
 278             'timestamp': 1703804085,
 279             'duration': 394.347,
 280             'view_count': int,
 281         },
 282     }, {
 283         'url': 'https://www.facebook.com/story.php?story_fbid=pfbid0Fnzhm8UuzjBYpPMNFzaSpFE9UmLdU4fJN8qTANi1Dmtj5q7DNrL5NERXfsAzDEV7l&id=100073071055552',
 284         'only_matching': True,
 285     }, {
 286         'url': 'https://www.facebook.com/video.php?v=10204634152394104',
 287         'only_matching': True,
 288     }, {
 289         'url': 'https://www.facebook.com/amogood/videos/1618742068337349/?fref=nf',
 290         'only_matching': True,
 291     }, {
 292         # data.mediaset.currMedia.edges
 293         'url': 'https://www.facebook.com/ChristyClarkForBC/videos/vb.22819070941/10153870694020942/?type=2&theater',
 294         'only_matching': True,
 295     }, {
 296         # data.video.story.attachments[].media
 297         'url': 'facebook:544765982287235',
 298         'only_matching': True,
 299     }, {
 300         # data.node.comet_sections.content.story.attachments[].style_type_renderer.attachment.media
 301         'url': 'https://www.facebook.com/groups/164828000315060/permalink/764967300301124/',
 302         'only_matching': True,
 303     }, {
 304         # data.video.creation_story.attachments[].media
 305         'url': 'https://zh-hk.facebook.com/peoplespower/videos/1135894589806027/',
 306         'only_matching': True,
 307     }, {
 308         # data.video
 309         'url': 'https://www.facebookwkhpilnemxj7asaniu7vnjjbiltxjqhye3mhbshg7kx5tfyd.onion/video.php?v=274175099429670',
 310         'only_matching': True,
 311     }, {
 312         # no title
 313         'url': 'https://www.facebook.com/onlycleverentertainment/videos/1947995502095005/',
 314         'only_matching': True,
 315     }, {
 316         # data.video
 317         'url': 'https://www.facebook.com/WatchESLOne/videos/359649331226507/',
 318         'info_dict': {
 319             'id': '359649331226507',
 320             'ext': 'mp4',
 321             'title': 'Fnatic vs. EG - Group A - Opening Match - ESL One Birmingham Day 1',
 322             'description': '#ESLOne VoD - Birmingham Finals Day#1 Fnatic vs. @Evil Geniuses',
 323             'timestamp': 1527084179,
 324             'upload_date': '20180523',
 325             'uploader': 'ESL One Dota 2',
 326             'uploader_id': '100066514874195',
 327             'duration': 4524.001,
 328             'view_count': int,
 329             'thumbnail': r're:^https?://.*',
 330             'concurrent_view_count': int,
 331         },
 332         'params': {
 333             'skip_download': True,
 334         },
 335     }, {
 336         # data.node.comet_sections.content.story.attachments[].style_type_renderer.attachment.all_subattachments.nodes[].media
 337         'url': 'https://www.facebook.com/100033620354545/videos/106560053808006/',
 338         'info_dict': {
 339             'id': '106560053808006',
 340             'ext': 'mp4',
 341             'title': 'Josef',
 342             'thumbnail': r're:^https?://.*',
 343             'concurrent_view_count': int,
 344             'uploader_id': 'pfbid02gpfwRM2XvdEJfsERupwQiNmBiDArc38RMRYZnap372q6Vs7MtFTVy72mmFWpJBTKl',
 345             'timestamp': 1549275572,
 346             'duration': 3.283,
 347             'uploader': 'Josef Novak',
 348             'description': '',
 349             'upload_date': '20190204',
 350         },
 351     }, {
 352         # data.video.story.attachments[].media
 353         'url': 'https://www.facebook.com/watch/?v=647537299265662',
 354         'only_matching': True,
 355     }, {
 356         # FIXME: https://github.com/yt-dlp/yt-dlp/issues/542
 357         # data.node.comet_sections.content.story.attachments[].style_type_renderer.attachment.all_subattachments.nodes[].media
 358         'url': 'https://www.facebook.com/PankajShahLondon/posts/10157667649866271',
 359         'info_dict': {
 360             'id': '10157667649866271',
 361         },
 362         'playlist_count': 3,
 363         'skip': 'Requires logging in',
 364     }, {
 365         # data.nodes[].comet_sections.content.story.attachments[].style_type_renderer.attachment.media
 366         'url': 'https://m.facebook.com/Alliance.Police.Department/posts/4048563708499330',
 367         'info_dict': {
 368             'id': '117576630041613',
 369             'ext': 'mp4',
 370             # TODO: title can be extracted from video page
 371             'title': 'Facebook video #117576630041613',
 372             'uploader_id': '189393014416438',
 373             'upload_date': '20201123',
 374             'timestamp': 1606162592,
 375         },
 376         'skip': 'Requires logging in',
 377     }, {
 378         # node.comet_sections.content.story.attached_story.attachments.style_type_renderer.attachment.media
 379         'url': 'https://www.facebook.com/groups/ateistiskselskab/permalink/10154930137678856/',
 380         'info_dict': {
 381             'id': '211567722618337',
 382             'ext': 'mp4',
 383             'title': 'Facebook video #211567722618337',
 384             'uploader_id': '127875227654254',
 385             'upload_date': '20161122',
 386             'timestamp': 1479793574,
 387         },
 388         'skip': 'No video',
 389     }, {
 390         # data.video.creation_story.attachments[].media
 391         'url': 'https://www.facebook.com/watch/live/?v=1823658634322275',
 392         'only_matching': True,
 393     }, {
 394         'url': 'https://www.facebook.com/watchparty/211641140192478',
 395         'info_dict': {
 396             'id': '211641140192478',
 397         },
 398         'playlist_count': 1,
 399         'skip': 'Requires logging in',
 400     }, {
 401         # FIXME: Cannot parse data error
 402         # data.event.cover_media_renderer.cover_video
 403         'url': 'https://m.facebook.com/events/1509582499515440',
 404         'info_dict': {
 405             'id': '637246984455045',
 406             'ext': 'mp4',
 407             'title': 'ANALISI IN CAMPO OSCURO " Coaguli nel sangue dei vaccinati"',
 408             'description': 'Other event by Comitato Liberi Pensatori on Tuesday, October 18 2022',
 409             'thumbnail': r're:^https?://.*',
 410             'uploader': 'Comitato Liberi Pensatori',
 411             'uploader_id': '100065709540881',
 412         },
 413     }, {
 414         'url': 'https://www.facebook.com/groups/1513990329015294/posts/d41d8cd9/2013209885760000/?app=fbl',
 415         'only_matching': True,
 416     }]
 417     _SUPPORTED_PAGLETS_REGEX = r'(?:pagelet_group_mall|permalink_video_pagelet|hyperfeed_story_id_[0-9a-f]+)'
 418     _api_config = {
 419         'graphURI': '/api/graphql/',
 420     }
 421
 422     def _perform_login(self, username, password):
 423         login_page_req = Request(self._LOGIN_URL)
 424         self._set_cookie('facebook.com', 'locale', 'en_US')
 425         login_page = self._download_webpage(login_page_req, None,
 426                                             note='Downloading login page',
 427                                             errnote='Unable to download login page')
 428         lsd = self._search_regex(
 429             r'<input type="hidden" name="lsd" value="([^"]*)"',
 430             login_page, 'lsd')
 431         lgnrnd = self._search_regex(r'name="lgnrnd" value="([^"]*?)"', login_page, 'lgnrnd')
 432
 433         login_form = {
 434             'email': username,
 435             'pass': password,
 436             'lsd': lsd,
 437             'lgnrnd': lgnrnd,
 438             'next': 'http://facebook.com/home.php',
 439             'default_persistent': '0',
 440             'legacy_return': '1',
 441             'timezone': '-60',
 442             'trynum': '1',
 443         }
 444         request = Request(self._LOGIN_URL, urlencode_postdata(login_form))
 445         request.headers['Content-Type'] = 'application/x-www-form-urlencoded'
 446         try:
 447             login_results = self._download_webpage(request, None,
 448                                                    note='Logging in', errnote='unable to fetch login page')
 449             if re.search(r'<form(.*)name="login"(.*)</form>', login_results) is not None:
 450                 error = self._html_search_regex(
 451                     r'(?s)<div[^>]+class=(["\']).*?login_error_box.*?\1[^>]*><div[^>]*>.*?</div><div[^>]*>(?P<error>.+?)</div>',
 452                     login_results, 'login error', default=None, group='error')
 453                 if error:
 454                     raise ExtractorError(f'Unable to login: {error}', expected=True)
 455                 self.report_warning('unable to log in: bad username/password, or exceeded login rate limit (~3/min). Check credentials or wait.')
 456                 return
 457
 458             fb_dtsg = self._search_regex(
 459                 r'name="fb_dtsg" value="(.+?)"', login_results, 'fb_dtsg', default=None)
 460             h = self._search_regex(
 461                 r'name="h"\s+(?:\w+="[^"]+"\s+)*?value="([^"]+)"', login_results, 'h', default=None)
 462
 463             if not fb_dtsg or not h:
 464                 return
 465
 466             check_form = {
 467                 'fb_dtsg': fb_dtsg,
 468                 'h': h,
 469                 'name_action_selected': 'dont_save',
 470             }
 471             check_req = Request(self._CHECKPOINT_URL, urlencode_postdata(check_form))
 472             check_req.headers['Content-Type'] = 'application/x-www-form-urlencoded'
 473             check_response = self._download_webpage(check_req, None,
 474                                                     note='Confirming login')
 475             if re.search(r'id="checkpointSubmitButton"', check_response) is not None:
 476                 self.report_warning('Unable to confirm login, you have to login in your browser and authorize the login.')
 477         except network_exceptions as err:
 478             self.report_warning(f'unable to log in: {err}')
 479             return
 480
 481     def _extract_from_url(self, url, video_id):
 482         webpage = self._download_webpage(
 483             url.replace('://m.facebook.com/', '://www.facebook.com/'), video_id)
 484
 485         def extract_metadata(webpage):
 486             post_data = [self._parse_json(j, video_id, fatal=False) for j in re.findall(
 487                 r'data-sjs>({.*?ScheduledServerJS.*?})</script>', webpage)]
 488             post = traverse_obj(post_data, (
 489                 ..., 'require', ..., ..., ..., '__bbox', 'require', ..., ..., ..., '__bbox', 'result', 'data'), expected_type=dict) or []
 490             media = traverse_obj(post, (..., 'attachments', ..., lambda k, v: (
 491                 k == 'media' and str(v['id']) == video_id and v['__typename'] == 'Video')), expected_type=dict)
 492             title = get_first(media, ('title', 'text'))
 493             description = get_first(media, ('creation_story', 'comet_sections', 'message', 'story', 'message', 'text'))
 494             page_title = title or self._html_search_regex((
 495                 r'<h2\s+[^>]*class="uiHeaderTitle"[^>]*>(?P<content>[^<]*)</h2>',
 496                 r'(?s)<span class="fbPhotosPhotoCaption".*?id="fbPhotoPageCaption"><span class="hasCaption">(?P<content>.*?)</span>',
 497                 self._meta_regex('og:title'), self._meta_regex('twitter:title'), r'<title>(?P<content>.+?)</title>',
 498             ), webpage, 'title', default=None, group='content')
 499             description = description or self._html_search_meta(
 500                 ['description', 'og:description', 'twitter:description'],
 501                 webpage, 'description', default=None)
 502             uploader_data = (
 503                 get_first(media, ('owner', {dict}))
 504                 or get_first(post, ('video', 'creation_story', 'attachments', ..., 'media', lambda k, v: k == 'owner' and v['name']))
 505                 or get_first(post, (..., 'video', lambda k, v: k == 'owner' and v['name']))
 506                 or get_first(post, ('node', 'actors', ..., {dict}))
 507                 or get_first(post, ('event', 'event_creator', {dict}))
 508                 or get_first(post, ('video', 'creation_story', 'short_form_video_context', 'video_owner', {dict})) or {})
 509             uploader = uploader_data.get('name') or (
 510                 clean_html(get_element_by_id('fbPhotoPageAuthorName', webpage))
 511                 or self._search_regex(
 512                     (r'ownerName\s*:\s*"([^"]+)"', *self._og_regexes('title')), webpage, 'uploader', fatal=False))
 513             timestamp = int_or_none(self._search_regex(
 514                 r'<abbr[^>]+data-utime=["\'](\d+)', webpage,
 515                 'timestamp', default=None))
 516             thumbnail = self._html_search_meta(
 517                 ['og:image', 'twitter:image'], webpage, 'thumbnail', default=None)
 518             # some webpages contain unretrievable thumbnail urls
 519             # like https://lookaside.fbsbx.com/lookaside/crawler/media/?media_id=10155168902769113&get_thumbnail=1
 520             # in https://www.facebook.com/yaroslav.korpan/videos/1417995061575415/
 521             if thumbnail and not re.search(r'\.(?:jpg|png)', thumbnail):
 522                 thumbnail = None
 523             info_dict = {
 524                 'description': description,
 525                 'uploader': uploader,
 526                 'uploader_id': uploader_data.get('id'),
 527                 'timestamp': timestamp,
 528                 'thumbnail': thumbnail,
 529                 'view_count': parse_count(self._search_regex(
 530                     (r'\bviewCount\s*:\s*["\']([\d,.]+)', r'video_view_count["\']\s*:\s*(\d+)'),
 531                     webpage, 'view count', default=None)),
 532                 'concurrent_view_count': get_first(post, (
 533                     ('video', (..., ..., 'attachments', ..., 'media')), 'liveViewerCount', {int_or_none})),
 534                 **traverse_obj(post, (lambda _, v: video_id in v['url'], 'feedback', {
 535                     'like_count': ('likers', 'count', {int}),
 536                     'comment_count': ('total_comment_count', {int}),
 537                     'repost_count': ('share_count_reduced', {parse_count}),
 538                 }), get_all=False),
 539             }
 540
 541             info_json_ld = self._search_json_ld(webpage, video_id, default={})
 542             info_json_ld['title'] = (re.sub(r'\s*\|\s*Facebook$', '', title or info_json_ld.get('title') or page_title or '')
 543                                      or (description or '').replace('\n', ' ') or f'Facebook video #{video_id}')
 544             return merge_dicts(info_json_ld, info_dict)
 545
 546         video_data = None
 547
 548         def extract_video_data(instances):
 549             video_data = []
 550             for item in instances:
 551                 if try_get(item, lambda x: x[1][0]) == 'VideoConfig':
 552                     video_item = item[2][0]
 553                     if video_item.get('video_id'):
 554                         video_data.append(video_item['videoData'])
 555             return video_data
 556
 557         server_js_data = self._parse_json(self._search_regex(
 558             [r'handleServerJS\(({.+})(?:\);|,")', r'\bs\.handle\(({.+?})\);'],
 559             webpage, 'server js data', default='{}'), video_id, fatal=False)
 560
 561         if server_js_data:
 562             video_data = extract_video_data(server_js_data.get('instances', []))
 563
 564         def extract_from_jsmods_instances(js_data):
 565             if js_data:
 566                 return extract_video_data(try_get(
 567                     js_data, lambda x: x['jsmods']['instances'], list) or [])
 568
 569         def extract_dash_manifest(vid_data, formats, mpd_url=None):
 570             dash_manifest = traverse_obj(
 571                 vid_data, 'dash_manifest', 'playlist', 'dash_manifest_xml_string', 'manifest_xml', expected_type=str)
 572             if dash_manifest:
 573                 formats.extend(self._parse_mpd_formats(
 574                     compat_etree_fromstring(urllib.parse.unquote_plus(dash_manifest)),
 575                     mpd_url=url_or_none(vid_data.get('dash_manifest_url')) or mpd_url))
 576
 577         def process_formats(info):
 578             # Downloads with browser's User-Agent are rate limited. Working around
 579             # with non-browser User-Agent.
 580             for f in info['formats']:
 581                 # Downloads with browser's User-Agent are rate limited. Working around
 582                 # with non-browser User-Agent.
 583                 f.setdefault('http_headers', {})['User-Agent'] = 'facebookexternalhit/1.1'
 584                 # Formats larger than ~500MB will return error 403 unless chunk size is regulated
 585                 f.setdefault('downloader_options', {})['http_chunk_size'] = 250 << 20
 586
 587         def yield_all_relay_data(_filter):
 588             for relay_data in re.findall(rf'data-sjs>({{.*?{_filter}.*?}})</script>', webpage):
 589                 yield self._parse_json(relay_data, video_id, fatal=False) or {}
 590
 591         def extract_relay_data(_filter):
 592             return next(filter(None, yield_all_relay_data(_filter)), {})
 593
 594         def extract_relay_prefetched_data(_filter, target_keys=None):
 595             path = 'data'
 596             if target_keys is not None:
 597                 path = lambda k, v: k == 'data' and any(target in v for target in variadic(target_keys))
 598             return traverse_obj(yield_all_relay_data(_filter), (
 599                 ..., 'require', (None, (..., ..., ..., '__bbox', 'require')),
 600                 lambda _, v: any(key.startswith('RelayPrefetchedStreamCache') for key in v),
 601                 ..., ..., '__bbox', 'result', path, {dict}), get_all=False) or {}
 602
 603         if not video_data:
 604             server_js_data = self._parse_json(self._search_regex([
 605                 r'bigPipe\.onPageletArrive\(({.+?})\)\s*;\s*}\s*\)\s*,\s*["\']onPageletArrive\s+' + self._SUPPORTED_PAGLETS_REGEX,
 606                 rf'bigPipe\.onPageletArrive\(({{.*?id\s*:\s*"{self._SUPPORTED_PAGLETS_REGEX}".*?}})\);',
 607             ], webpage, 'js data', default='{}'), video_id, js_to_json, False)
 608             video_data = extract_from_jsmods_instances(server_js_data)
 609
 610         if not video_data:
 611             data = extract_relay_prefetched_data(
 612                 r'"(?:dash_manifest|playable_url(?:_quality_hd)?)',
 613                 target_keys=('video', 'event', 'nodes', 'node', 'mediaset'))
 614             if data:
 615                 entries = []
 616
 617                 def parse_graphql_video(video):
 618                     v_id = video.get('videoId') or video.get('id') or video_id
 619                     reel_info = traverse_obj(
 620                         video, ('creation_story', 'short_form_video_context', 'playback_video', {dict}))
 621                     if reel_info:
 622                         video = video['creation_story']
 623                         video['owner'] = traverse_obj(video, ('short_form_video_context', 'video_owner'))
 624                         video.update(reel_info)
 625
 626                     formats = []
 627                     q = qualities(['sd', 'hd'])
 628
 629                     # Legacy formats extraction
 630                     fmt_data = traverse_obj(video, ('videoDeliveryLegacyFields', {dict})) or video
 631                     for key, format_id in (('playable_url', 'sd'), ('playable_url_quality_hd', 'hd'),
 632                                            ('playable_url_dash', ''), ('browser_native_hd_url', 'hd'),
 633                                            ('browser_native_sd_url', 'sd')):
 634                         playable_url = fmt_data.get(key)
 635                         if not playable_url:
 636                             continue
 637                         if determine_ext(playable_url) == 'mpd':
 638                             formats.extend(self._extract_mpd_formats(playable_url, video_id, fatal=False))
 639                         else:
 640                             formats.append({
 641                                 'format_id': format_id,
 642                                 # sd, hd formats w/o resolution info should be deprioritized below DASH
 643                                 'quality': q(format_id) - 3,
 644                                 'url': playable_url,
 645                             })
 646                     extract_dash_manifest(fmt_data, formats)
 647
 648                     # New videoDeliveryResponse formats extraction
 649                     fmt_data = traverse_obj(video, ('videoDeliveryResponseFragment', 'videoDeliveryResponseResult'))
 650                     mpd_urls = traverse_obj(fmt_data, ('dash_manifest_urls', ..., 'manifest_url', {url_or_none}))
 651                     dash_manifests = traverse_obj(fmt_data, ('dash_manifests', lambda _, v: v['manifest_xml']))
 652                     for idx, dash_manifest in enumerate(dash_manifests):
 653                         extract_dash_manifest(dash_manifest, formats, mpd_url=traverse_obj(mpd_urls, idx))
 654                     if not dash_manifests:
 655                         # Only extract from MPD URLs if the manifests are not already provided
 656                         for mpd_url in mpd_urls:
 657                             formats.extend(self._extract_mpd_formats(mpd_url, video_id, fatal=False))
 658                     for prog_fmt in traverse_obj(fmt_data, ('progressive_urls', lambda _, v: v['progressive_url'])):
 659                         format_id = traverse_obj(prog_fmt, ('metadata', 'quality', {str.lower}))
 660                         formats.append({
 661                             'format_id': format_id,
 662                             # sd, hd formats w/o resolution info should be deprioritized below DASH
 663                             'quality': q(format_id) - 3,
 664                             'url': prog_fmt['progressive_url'],
 665                         })
 666                     for m3u8_url in traverse_obj(fmt_data, ('hls_playlist_urls', ..., 'hls_playlist_url', {url_or_none})):
 667                         formats.extend(self._extract_m3u8_formats(m3u8_url, video_id, 'mp4', fatal=False, m3u8_id='hls'))
 668
 669                     if not formats:
 670                         # Do not append false positive entry w/o any formats
 671                         return
 672
 673                     automatic_captions, subtitles = {}, {}
 674                     is_broadcast = traverse_obj(video, ('is_video_broadcast', {bool}))
 675                     for caption in traverse_obj(video, (
 676                         'video_available_captions_locales',
 677                         {lambda x: sorted(x, key=lambda c: c['locale'])},
 678                         lambda _, v: url_or_none(v['captions_url']),
 679                     )):
 680                         lang = caption.get('localized_language') or 'und'
 681                         subs = {
 682                             'url': caption['captions_url'],
 683                             'name': format_field(caption, 'localized_country', f'{lang} (%s)', default=lang),
 684                         }
 685                         if caption.get('localized_creation_method') or is_broadcast:
 686                             automatic_captions.setdefault(caption['locale'], []).append(subs)
 687                         else:
 688                             subtitles.setdefault(caption['locale'], []).append(subs)
 689                     captions_url = traverse_obj(video, ('captions_url', {url_or_none}))
 690                     if captions_url and not automatic_captions and not subtitles:
 691                         locale = self._html_search_meta(
 692                             ['og:locale', 'twitter:locale'], webpage, 'locale', default='en_US')
 693                         (automatic_captions if is_broadcast else subtitles)[locale] = [{'url': captions_url}]
 694
 695                     info = {
 696                         'id': v_id,
 697                         'formats': formats,
 698                         'thumbnail': traverse_obj(
 699                             video, ('thumbnailImage', 'uri'), ('preferred_thumbnail', 'image', 'uri')),
 700                         'uploader_id': traverse_obj(video, ('owner', 'id', {str_or_none})),
 701                         'timestamp': traverse_obj(video, 'publish_time', 'creation_time', expected_type=int_or_none),
 702                         'duration': (float_or_none(video.get('playable_duration_in_ms'), 1000)
 703                                      or float_or_none(video.get('length_in_second'))),
 704                         'automatic_captions': automatic_captions,
 705                         'subtitles': subtitles,
 706                     }
 707                     process_formats(info)
 708                     description = try_get(video, lambda x: x['savable_description']['text'])
 709                     title = video.get('name')
 710                     if title:
 711                         info.update({
 712                             'title': title,
 713                             'description': description,
 714                         })
 715                     else:
 716                         info['title'] = description or f'Facebook video #{v_id}'
 717                     entries.append(info)
 718
 719                 def parse_attachment(attachment, key='media'):
 720                     media = attachment.get(key) or {}
 721                     if media.get('__typename') == 'Video':
 722                         return parse_graphql_video(media)
 723
 724                 nodes = variadic(traverse_obj(data, 'nodes', 'node') or [])
 725                 attachments = traverse_obj(nodes, (
 726                     ..., 'comet_sections', 'content', 'story', (None, 'attached_story'), 'attachments',
 727                     ..., ('styles', 'style_type_renderer', ('throwbackStyles', 'attachment_target_renderer')),
 728                     'attachment', {dict}))
 729                 for attachment in attachments:
 730                     ns = traverse_obj(attachment, ('all_subattachments', 'nodes', ..., {dict}),
 731                                       ('target', 'attachments', ..., 'styles', 'attachment', {dict}))
 732                     for n in ns:
 733                         parse_attachment(n)
 734                     parse_attachment(attachment)
 735
 736                 edges = try_get(data, lambda x: x['mediaset']['currMedia']['edges'], list) or []
 737                 for edge in edges:
 738                     parse_attachment(edge, key='node')
 739
 740                 video = traverse_obj(data, (
 741                     'event', 'cover_media_renderer', 'cover_video'), 'video', expected_type=dict) or {}
 742                 if video:
 743                     attachments = try_get(video, [
 744                         lambda x: x['story']['attachments'],
 745                         lambda x: x['creation_story']['attachments'],
 746                     ], list) or []
 747                     for attachment in attachments:
 748                         parse_attachment(attachment)
 749                     if not entries:
 750                         parse_graphql_video(video)
 751
 752                 if len(entries) > 1:
 753                     return self.playlist_result(entries, video_id)
 754
 755                 video_info = entries[0] if entries else {'id': video_id}
 756                 webpage_info = extract_metadata(webpage)
 757                 # honor precise duration in video info
 758                 if video_info.get('duration'):
 759                     webpage_info['duration'] = video_info['duration']
 760                 # preserve preferred_thumbnail in video info
 761                 if video_info.get('thumbnail'):
 762                     webpage_info['thumbnail'] = video_info['thumbnail']
 763                 return merge_dicts(webpage_info, video_info)
 764
 765         if not video_data:
 766             m_msg = re.search(r'class="[^"]*uiInterstitialContent[^"]*"><div>(.*?)</div>', webpage)
 767             if m_msg is not None:
 768                 raise ExtractorError(
 769                     f'The video is not available, Facebook said: "{m_msg.group(1)}"',
 770                     expected=True)
 771             elif any(p in webpage for p in (
 772                     '>You must log in to continue',
 773                     'id="login_form"',
 774                     'id="loginbutton"')):
 775                 self.raise_login_required()
 776
 777         if not video_data and '/watchparty/' in url:
 778             post_data = {
 779                 'doc_id': 3731964053542869,
 780                 'variables': json.dumps({
 781                     'livingRoomID': video_id,
 782                 }),
 783             }
 784
 785             prefetched_data = extract_relay_prefetched_data(r'"login_data"\s*:\s*{')
 786             if prefetched_data:
 787                 lsd = try_get(prefetched_data, lambda x: x['login_data']['lsd'], dict)
 788                 if lsd:
 789                     post_data[lsd['name']] = lsd['value']
 790
 791             relay_data = extract_relay_data(r'\[\s*"RelayAPIConfigDefaults"\s*,')
 792             for define in (relay_data.get('define') or []):
 793                 if define[0] == 'RelayAPIConfigDefaults':
 794                     self._api_config = define[2]
 795
 796             living_room = self._download_json(
 797                 urljoin(url, self._api_config['graphURI']), video_id,
 798                 data=urlencode_postdata(post_data))['data']['living_room']
 799
 800             entries = []
 801             for edge in (try_get(living_room, lambda x: x['recap']['watched_content']['edges']) or []):
 802                 video = try_get(edge, lambda x: x['node']['video']) or {}
 803                 v_id = video.get('id')
 804                 if not v_id:
 805                     continue
 806                 v_id = str(v_id)
 807                 entries.append(self.url_result(
 808                     self._VIDEO_PAGE_TEMPLATE % v_id,
 809                     self.ie_key(), v_id, video.get('name')))
 810
 811             return self.playlist_result(entries, video_id)
 812
 813         if not video_data:
 814             # Video info not in first request, do a secondary request using
 815             # tahoe player specific URL
 816             tahoe_data = self._download_webpage(
 817                 self._VIDEO_PAGE_TAHOE_TEMPLATE % video_id, video_id,
 818                 data=urlencode_postdata({
 819                     '__a': 1,
 820                     '__pc': self._search_regex(
 821                         r'pkg_cohort["\']\s*:\s*["\'](.+?)["\']', webpage,
 822                         'pkg cohort', default='PHASED:DEFAULT'),
 823                     '__rev': self._search_regex(
 824                         r'client_revision["\']\s*:\s*(\d+),', webpage,
 825                         'client revision', default='3944515'),
 826                     'fb_dtsg': self._search_regex(
 827                         r'"DTSGInitialData"\s*,\s*\[\]\s*,\s*{\s*"token"\s*:\s*"([^"]+)"',
 828                         webpage, 'dtsg token', default=''),
 829                 }),
 830                 headers={
 831                     'Content-Type': 'application/x-www-form-urlencoded',
 832                 })
 833             tahoe_js_data = self._parse_json(
 834                 self._search_regex(
 835                     r'for\s+\(\s*;\s*;\s*\)\s*;(.+)', tahoe_data,
 836                     'tahoe js data', default='{}'),
 837                 video_id, fatal=False)
 838             video_data = extract_from_jsmods_instances(tahoe_js_data)
 839
 840         if not video_data:
 841             raise ExtractorError('Cannot parse data')
 842
 843         if len(video_data) > 1:
 844             entries = []
 845             for v in video_data:
 846                 video_url = v[0].get('video_url')
 847                 if not video_url:
 848                     continue
 849                 entries.append(self.url_result(urljoin(
 850                     url, video_url), self.ie_key(), v[0].get('video_id')))
 851             return self.playlist_result(entries, video_id)
 852         video_data = video_data[0]
 853
 854         formats = []
 855         subtitles = {}
 856         for f in video_data:
 857             format_id = f['stream_type']
 858             if f and isinstance(f, dict):
 859                 f = [f]
 860             if not f or not isinstance(f, list):
 861                 continue
 862             for quality in ('sd', 'hd'):
 863                 for src_type in ('src', 'src_no_ratelimit'):
 864                     src = f[0].get(f'{quality}_{src_type}')
 865                     if src:
 866                         # sd, hd formats w/o resolution info should be deprioritized below DASH
 867                         # TODO: investigate if progressive or src formats still exist
 868                         preference = -10 if format_id == 'progressive' else -3
 869                         if quality == 'hd':
 870                             preference += 1
 871                         formats.append({
 872                             'format_id': f'{format_id}_{quality}_{src_type}',
 873                             'url': src,
 874                             'quality': preference,
 875                             'height': 720 if quality == 'hd' else None,
 876                         })
 877             extract_dash_manifest(f[0], formats)
 878             subtitles_src = f[0].get('subtitles_src')
 879             if subtitles_src:
 880                 subtitles.setdefault('en', []).append({'url': subtitles_src})
 881
 882         info_dict = {
 883             'id': video_id,
 884             'formats': formats,
 885             'subtitles': subtitles,
 886         }
 887         process_formats(info_dict)
 888         info_dict.update(extract_metadata(webpage))
 889
 890         return info_dict
 891
 892     def _real_extract(self, url):
 893         video_id = self._match_id(url)
 894
 895         real_url = self._VIDEO_PAGE_TEMPLATE % video_id if url.startswith('facebook:') else url
 896         return self._extract_from_url(real_url, video_id)
 897
 898
 899 class FacebookPluginsVideoIE(InfoExtractor):
 900     _VALID_URL = r'https?://(?:[\w-]+\.)?facebook\.com/plugins/video\.php\?.*?\bhref=(?P<id>https.+)'
 901
 902     _TESTS = [{
 903         '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',
 904         'md5': '5954e92cdfe51fe5782ae9bda7058a07',
 905         'info_dict': {
 906             'id': '10154383743583686',
 907             'ext': 'mp4',
 908             # TODO: Fix title, uploader
 909             'title': 'What to do during the haze?',
 910             'uploader': 'Gov.sg',
 911             'upload_date': '20160826',
 912             'timestamp': 1472184808,
 913         },
 914         'add_ie': [FacebookIE.ie_key()],
 915     }, {
 916         'url': 'https://www.facebook.com/plugins/video.php?href=https%3A%2F%2Fwww.facebook.com%2Fvideo.php%3Fv%3D10204634152394104',
 917         'only_matching': True,
 918     }, {
 919         'url': 'https://www.facebook.com/plugins/video.php?href=https://www.facebook.com/gov.sg/videos/10154383743583686/&show_text=0&width=560',
 920         'only_matching': True,
 921     }]
 922
 923     def _real_extract(self, url):
 924         return self.url_result(
 925             urllib.parse.unquote(self._match_id(url)),
 926             FacebookIE.ie_key())
 927
 928
 929 class FacebookRedirectURLIE(InfoExtractor):
 930     IE_DESC = False  # Do not list
 931     _VALID_URL = r'https?://(?:[\w-]+\.)?facebook\.com/flx/warn[/?]'
 932     _TESTS = [{
 933         'url': 'https://www.facebook.com/flx/warn/?h=TAQHsoToz&u=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DpO8h3EaFRdo&s=1',
 934         'info_dict': {
 935             'id': 'pO8h3EaFRdo',
 936             'ext': 'mp4',
 937             'title': 'Tripeo Boiler Room x Dekmantel Festival DJ Set',
 938             'description': 'md5:2d713ccbb45b686a1888397b2c77ca6b',
 939             'channel_id': 'UCGBpxWJr9FNOcFYA5GkKrMg',
 940             'playable_in_embed': True,
 941             'categories': ['Music'],
 942             'channel': 'Boiler Room',
 943             'uploader_id': 'brtvofficial',
 944             'uploader': 'Boiler Room',
 945             'tags': 'count:11',
 946             'duration': 3332,
 947             'live_status': 'not_live',
 948             'thumbnail': 'https://i.ytimg.com/vi/pO8h3EaFRdo/maxresdefault.jpg',
 949             'channel_url': 'https://www.youtube.com/channel/UCGBpxWJr9FNOcFYA5GkKrMg',
 950             'availability': 'public',
 951             'uploader_url': 'http://www.youtube.com/user/brtvofficial',
 952             'upload_date': '20150917',
 953             'age_limit': 0,
 954             'view_count': int,
 955             'like_count': int,
 956         },
 957         'add_ie': ['Youtube'],
 958         'params': {'skip_download': 'Youtube'},
 959     }]
 960
 961     def _real_extract(self, url):
 962         redirect_url = url_or_none(parse_qs(url).get('u', [None])[-1])
 963         if not redirect_url:
 964             raise ExtractorError('Invalid facebook redirect URL', expected=True)
 965         return self.url_result(redirect_url)
 966
 967
 968 class FacebookReelIE(InfoExtractor):
 969     _VALID_URL = r'https?://(?:[\w-]+\.)?facebook\.com/reel/(?P<id>\d+)'
 970     IE_NAME = 'facebook:reel'
 971
 972     _TESTS = [{
 973         'url': 'https://www.facebook.com/reel/1195289147628387',
 974         'md5': 'a53256d10fc2105441fe0c4212ed8cea',
 975         'info_dict': {
 976             'id': '1195289147628387',
 977             'ext': 'mp4',
 978             'title': r're:9\.6K views · 355 reactions .+ Let the “Slapathon” commence!! .+ LL COOL J · Mama Said Knock You Out$',
 979             'description': r're:When your trying to help your partner .+ LL COOL J · Mama Said Knock You Out$',
 980             'uploader': 'Beast Camp Training',
 981             'uploader_id': '100040874179269',
 982             'duration': 9.579,
 983             'timestamp': 1637502609,
 984             'upload_date': '20211121',
 985             'thumbnail': r're:^https?://.*',
 986             'like_count': int,
 987             'comment_count': int,
 988             'repost_count': int,
 989         },
 990     }]
 991
 992     def _real_extract(self, url):
 993         video_id = self._match_id(url)
 994         return self.url_result(
 995             f'https://m.facebook.com/watch/?v={video_id}&_rdr', FacebookIE, video_id)
 996
 997
 998 class FacebookAdsIE(InfoExtractor):
 999     _VALID_URL = r'https?://(?:[\w-]+\.)?facebook\.com/ads/library/?\?(?:[^#]+&)?id=(?P<id>\d+)'
1000     IE_NAME = 'facebook:ads'
1001
1002     _TESTS = [{
1003         'url': 'https://www.facebook.com/ads/library/?id=899206155126718',
1004         'info_dict': {
1005             'id': '899206155126718',
1006             'ext': 'mp4',
1007             'title': 'video by Kandao',
1008             'description': 'md5:0822724069e3aca97cbed5dabbab282e',
1009             'uploader': 'Kandao',
1010             'uploader_id': '774114102743284',
1011             'uploader_url': r're:^https?://.*',
1012             'timestamp': 1702548330,
1013             'thumbnail': r're:^https?://.*',
1014             'upload_date': '20231214',
1015             'like_count': int,
1016         },
1017     }, {
1018         # key 'watermarked_video_sd_url' missing
1019         'url': 'https://www.facebook.com/ads/library/?id=501152689226254',
1020         'info_dict': {
1021             'id': '501152689226254',
1022             'ext': 'mp4',
1023             'title': 'video by mat.nawrocki',
1024             'description': 'md5:02a446ace7ff8c3c37a2892922492490',
1025             'uploader': 'mat.nawrocki',
1026             'uploader_id': '148586968341456',
1027             'uploader_url': r're:^https?://.*',
1028             'timestamp': 1723452305,
1029             'thumbnail': r're:^https?://.*',
1030             'upload_date': '20240812',
1031             'like_count': int,
1032         },
1033     }, {
1034         'url': 'https://www.facebook.com/ads/library/?id=893637265423481',
1035         'info_dict': {
1036             'id': '893637265423481',
1037             'title': 'Jusqu\u2019\u00e0 -25% sur une s\u00e9lection de vins p\u00e9tillants italiens ',
1038             'uploader': 'Eataly Paris Marais',
1039             'uploader_id': '2086668958314152',
1040             'uploader_url': r're:^https?://.*',
1041             'timestamp': 1703571529,
1042             'upload_date': '20231226',
1043             'like_count': int,
1044         },
1045         'playlist_count': 3,
1046     }, {
1047         'url': 'https://es-la.facebook.com/ads/library/?id=901230958115569',
1048         'only_matching': True,
1049     }, {
1050         'url': 'https://m.facebook.com/ads/library/?id=901230958115569',
1051         'only_matching': True,
1052     }]
1053
1054     _FORMATS_MAP = {
1055         'watermarked_video_sd_url': ('sd-wmk', 'SD, watermarked'),
1056         'video_sd_url': ('sd', None),
1057         'watermarked_video_hd_url': ('hd-wmk', 'HD, watermarked'),
1058         'video_hd_url': ('hd', None),
1059     }
1060
1061     def _extract_formats(self, video_dict):
1062         formats = []
1063         for format_key, format_url in traverse_obj(video_dict, (
1064             {dict.items}, lambda _, v: v[0] in self._FORMATS_MAP and url_or_none(v[1]),
1065         )):
1066             formats.append({
1067                 'format_id': self._FORMATS_MAP[format_key][0],
1068                 'format_note': self._FORMATS_MAP[format_key][1],
1069                 'url': format_url,
1070                 'ext': 'mp4',
1071                 'quality': qualities(tuple(self._FORMATS_MAP))(format_key),
1072             })
1073         return formats
1074
1075     def _real_extract(self, url):
1076         video_id = self._match_id(url)
1077         webpage = self._download_webpage(url, video_id)
1078
1079         post_data = traverse_obj(
1080             re.findall(r'data-sjs>({.*?ScheduledServerJS.*?})</script>', webpage), (..., {json.loads}))
1081         data = get_first(post_data, (
1082             'require', ..., ..., ..., '__bbox', 'require', ..., ..., ...,
1083             'entryPointRoot', 'otherProps', 'deeplinkAdCard', 'snapshot', {dict}))
1084         if not data:
1085             raise ExtractorError('Unable to extract ad data')
1086
1087         title = data.get('title')
1088         if not title or title == '{{product.name}}':
1089             title = join_nonempty('display_format', 'page_name', delim=' by ', from_dict=data)
1090         markup_id = traverse_obj(data, ('body', '__m', {str}))
1091         markup = traverse_obj(post_data, (
1092             ..., 'require', ..., ..., ..., '__bbox', 'markup', lambda _, v: v[0].startswith(markup_id),
1093             ..., '__html', {clean_html}, {lambda x: not x.startswith('{{product.') and x}, any))
1094
1095         info_dict = merge_dicts({
1096             'title': title,
1097             'description': markup or None,
1098         }, traverse_obj(data, {
1099             'description': ('link_description', {lambda x: x if not x.startswith('{{product.') else None}),
1100             'uploader': ('page_name', {str}),
1101             'uploader_id': ('page_id', {str_or_none}),
1102             'uploader_url': ('page_profile_uri', {url_or_none}),
1103             'timestamp': ('creation_time', {int_or_none}),
1104             'like_count': ('page_like_count', {int_or_none}),
1105         }))
1106
1107         entries = []
1108         for idx, entry in enumerate(traverse_obj(
1109             data, (('videos', 'cards'), lambda _, v: any(url_or_none(v.get(f)) for f in self._FORMATS_MAP))), 1,
1110         ):
1111             entries.append({
1112                 'id': f'{video_id}_{idx}',
1113                 'title': entry.get('title') or title,
1114                 'description': traverse_obj(entry, 'body', 'link_description') or info_dict.get('description'),
1115                 'thumbnail': url_or_none(entry.get('video_preview_image_url')),
1116                 'formats': self._extract_formats(entry),
1117             })
1118
1119         if len(entries) == 1:
1120             info_dict.update(entries[0])
1121
1122         elif len(entries) > 1:
1123             info_dict.update({
1124                 'title': entries[0]['title'],
1125                 'entries': entries,
1126                 '_type': 'playlist',
1127             })
1128
1129         info_dict['id'] = video_id
1130
1131         return info_dict