yt_dlp/extractor/facebook.py

   1 # coding: utf-8
   2 from __future__ import unicode_literals
   3
   4 import json
   5 import re
   6
   7 from .common import InfoExtractor
   8 from ..compat import (
   9     compat_etree_fromstring,
  10     compat_str,
  11     compat_urllib_parse_unquote,
  12     compat_urllib_parse_unquote_plus,
  13 )
  14 from ..utils import (
  15     clean_html,
  16     error_to_compat_str,
  17     ExtractorError,
  18     float_or_none,
  19     get_element_by_id,
  20     int_or_none,
  21     js_to_json,
  22     limit_length,
  23     network_exceptions,
  24     parse_count,
  25     qualities,
  26     sanitized_Request,
  27     try_get,
  28     urlencode_postdata,
  29     urljoin,
  30 )
  31
  32
  33 class FacebookIE(InfoExtractor):
  34     _VALID_URL = r'''(?x)
  35                 (?:
  36                     https?://
  37                         (?:[\w-]+\.)?(?:facebook\.com|facebookcorewwwi\.onion)/
  38                         (?:[^#]*?\#!/)?
  39                         (?:
  40                             (?:
  41                                 video/video\.php|
  42                                 photo\.php|
  43                                 video\.php|
  44                                 video/embed|
  45                                 story\.php|
  46                                 watch(?:/live)?/?
  47                             )\?(?:.*?)(?:v|video_id|story_fbid)=|
  48                             [^/]+/videos/(?:[^/]+/)?|
  49                             [^/]+/posts/|
  50                             groups/[^/]+/permalink/|
  51                             watchparty/
  52                         )|
  53                     facebook:
  54                 )
  55                 (?P<id>[0-9]+)
  56                 '''
  57     _LOGIN_URL = 'https://www.facebook.com/login.php?next=http%3A%2F%2Ffacebook.com%2Fhome.php&login_attempt=1'
  58     _CHECKPOINT_URL = 'https://www.facebook.com/checkpoint/?next=http%3A%2F%2Ffacebook.com%2Fhome.php&_fb_noscript=1'
  59     _NETRC_MACHINE = 'facebook'
  60     IE_NAME = 'facebook'
  61
  62     _VIDEO_PAGE_TEMPLATE = 'https://www.facebook.com/video/video.php?v=%s'
  63     _VIDEO_PAGE_TAHOE_TEMPLATE = 'https://www.facebook.com/video/tahoe/async/%s/?chain=true&isvideo=true&payloadtype=primary'
  64
  65     _TESTS = [{
  66         'url': 'https://www.facebook.com/video.php?v=637842556329505&fref=nf',
  67         'md5': '6a40d33c0eccbb1af76cf0485a052659',
  68         'info_dict': {
  69             'id': '637842556329505',
  70             'ext': 'mp4',
  71             'title': 're:Did you know Kei Nishikori is the first Asian man to ever reach a Grand Slam',
  72             'uploader': 'Tennis on Facebook',
  73             'upload_date': '20140908',
  74             'timestamp': 1410199200,
  75         },
  76         'skip': 'Requires logging in',
  77     }, {
  78         # data.video
  79         'url': 'https://www.facebook.com/video.php?v=274175099429670',
  80         'info_dict': {
  81             'id': '274175099429670',
  82             'ext': 'mp4',
  83             'title': 're:^Asif Nawab Butt posted a video',
  84             'uploader': 'Asif Nawab Butt',
  85             'upload_date': '20140506',
  86             'timestamp': 1399398998,
  87             'thumbnail': r're:^https?://.*',
  88         },
  89         'expected_warnings': [
  90             'title'
  91         ]
  92     }, {
  93         'note': 'Video with DASH manifest',
  94         'url': 'https://www.facebook.com/video.php?v=957955867617029',
  95         'md5': 'b2c28d528273b323abe5c6ab59f0f030',
  96         'info_dict': {
  97             'id': '957955867617029',
  98             'ext': 'mp4',
  99             'title': 'When you post epic content on instagram.com/433 8 million followers, this is ...',
 100             'uploader': 'Demy de Zeeuw',
 101             'upload_date': '20160110',
 102             'timestamp': 1452431627,
 103         },
 104         'skip': 'Requires logging in',
 105     }, {
 106         'url': 'https://www.facebook.com/maxlayn/posts/10153807558977570',
 107         'md5': '037b1fa7f3c2d02b7a0d7bc16031ecc6',
 108         'info_dict': {
 109             'id': '544765982287235',
 110             'ext': 'mp4',
 111             'title': '"What are you doing running in the snow?"',
 112             'uploader': 'FailArmy',
 113         },
 114         'skip': 'Video gone',
 115     }, {
 116         'url': 'https://m.facebook.com/story.php?story_fbid=1035862816472149&id=116132035111903',
 117         'md5': '1deb90b6ac27f7efcf6d747c8a27f5e3',
 118         'info_dict': {
 119             'id': '1035862816472149',
 120             'ext': 'mp4',
 121             'title': 'What the Flock Is Going On In New Zealand  Credit: ViralHog',
 122             'uploader': 'S. Saint',
 123         },
 124         'skip': 'Video gone',
 125     }, {
 126         'note': 'swf params escaped',
 127         'url': 'https://www.facebook.com/barackobama/posts/10153664894881749',
 128         'md5': '97ba073838964d12c70566e0085c2b91',
 129         'info_dict': {
 130             'id': '10153664894881749',
 131             'ext': 'mp4',
 132             'title': 'Average time to confirm recent Supreme Court nominees: 67 days Longest it\'s t...',
 133             'thumbnail': r're:^https?://.*',
 134             'timestamp': 1456259628,
 135             'upload_date': '20160223',
 136             'uploader': 'Barack Obama',
 137         },
 138     }, {
 139         # have 1080P, but only up to 720p in swf params
 140         # data.video.story.attachments[].media
 141         'url': 'https://www.facebook.com/cnn/videos/10155529876156509/',
 142         'md5': '9571fae53d4165bbbadb17a94651dcdc',
 143         'info_dict': {
 144             'id': '10155529876156509',
 145             'ext': 'mp4',
 146             'title': 'She survived the holocaust — and years later, she’s getting her citizenship s...',
 147             'timestamp': 1477818095,
 148             'upload_date': '20161030',
 149             'uploader': 'CNN',
 150             'thumbnail': r're:^https?://.*',
 151             'view_count': int,
 152         },
 153     }, {
 154         # bigPipe.onPageletArrive ... onPageletArrive pagelet_group_mall
 155         # data.node.comet_sections.content.story.attachments[].style_type_renderer.attachment.media
 156         'url': 'https://www.facebook.com/yaroslav.korpan/videos/1417995061575415/',
 157         'info_dict': {
 158             'id': '1417995061575415',
 159             'ext': 'mp4',
 160             'title': 'md5:1db063d6a8c13faa8da727817339c857',
 161             'timestamp': 1486648217,
 162             'upload_date': '20170209',
 163             'uploader': 'Yaroslav Korpan',
 164         },
 165         'params': {
 166             'skip_download': True,
 167         },
 168     }, {
 169         'url': 'https://www.facebook.com/LaGuiaDelVaron/posts/1072691702860471',
 170         'info_dict': {
 171             'id': '1072691702860471',
 172             'ext': 'mp4',
 173             'title': 'md5:ae2d22a93fbb12dad20dc393a869739d',
 174             'timestamp': 1477305000,
 175             'upload_date': '20161024',
 176             'uploader': 'La Guía Del Varón',
 177             'thumbnail': r're:^https?://.*',
 178         },
 179         'params': {
 180             'skip_download': True,
 181         },
 182     }, {
 183         # data.node.comet_sections.content.story.attachments[].style_type_renderer.attachment.media
 184         'url': 'https://www.facebook.com/groups/1024490957622648/permalink/1396382447100162/',
 185         'info_dict': {
 186             'id': '1396382447100162',
 187             'ext': 'mp4',
 188             'title': 'md5:19a428bbde91364e3de815383b54a235',
 189             'timestamp': 1486035494,
 190             'upload_date': '20170202',
 191             'uploader': 'Elisabeth Ahtn',
 192         },
 193         'params': {
 194             'skip_download': True,
 195         },
 196     }, {
 197         'url': 'https://www.facebook.com/video.php?v=10204634152394104',
 198         'only_matching': True,
 199     }, {
 200         'url': 'https://www.facebook.com/amogood/videos/1618742068337349/?fref=nf',
 201         'only_matching': True,
 202     }, {
 203         # data.mediaset.currMedia.edges
 204         'url': 'https://www.facebook.com/ChristyClarkForBC/videos/vb.22819070941/10153870694020942/?type=2&theater',
 205         'only_matching': True,
 206     }, {
 207         # data.video.story.attachments[].media
 208         'url': 'facebook:544765982287235',
 209         'only_matching': True,
 210     }, {
 211         # data.node.comet_sections.content.story.attachments[].style_type_renderer.attachment.media
 212         'url': 'https://www.facebook.com/groups/164828000315060/permalink/764967300301124/',
 213         'only_matching': True,
 214     }, {
 215         # data.video.creation_story.attachments[].media
 216         'url': 'https://zh-hk.facebook.com/peoplespower/videos/1135894589806027/',
 217         'only_matching': True,
 218     }, {
 219         # data.video
 220         'url': 'https://www.facebookcorewwwi.onion/video.php?v=274175099429670',
 221         'only_matching': True,
 222     }, {
 223         # no title
 224         'url': 'https://www.facebook.com/onlycleverentertainment/videos/1947995502095005/',
 225         'only_matching': True,
 226     }, {
 227         # data.video
 228         'url': 'https://www.facebook.com/WatchESLOne/videos/359649331226507/',
 229         'info_dict': {
 230             'id': '359649331226507',
 231             'ext': 'mp4',
 232             'title': '#ESLOne VoD - Birmingham Finals Day#1 Fnatic vs. @Evil Geniuses',
 233             'uploader': 'ESL One Dota 2',
 234         },
 235         'params': {
 236             'skip_download': True,
 237         },
 238     }, {
 239         # data.node.comet_sections.content.story.attachments[].style_type_renderer.attachment.all_subattachments.nodes[].media
 240         'url': 'https://www.facebook.com/100033620354545/videos/106560053808006/',
 241         'info_dict': {
 242             'id': '106560053808006',
 243         },
 244         'playlist_count': 2,
 245     }, {
 246         # data.video.story.attachments[].media
 247         'url': 'https://www.facebook.com/watch/?v=647537299265662',
 248         'only_matching': True,
 249     }, {
 250         # data.node.comet_sections.content.story.attachments[].style_type_renderer.attachment.all_subattachments.nodes[].media
 251         'url': 'https://www.facebook.com/PankajShahLondon/posts/10157667649866271',
 252         'info_dict': {
 253             'id': '10157667649866271',
 254         },
 255         'playlist_count': 3,
 256     }, {
 257         # data.nodes[].comet_sections.content.story.attachments[].style_type_renderer.attachment.media
 258         'url': 'https://m.facebook.com/Alliance.Police.Department/posts/4048563708499330',
 259         'info_dict': {
 260             'id': '117576630041613',
 261             'ext': 'mp4',
 262             # TODO: title can be extracted from video page
 263             'title': 'Facebook video #117576630041613',
 264             'uploader_id': '189393014416438',
 265             'upload_date': '20201123',
 266             'timestamp': 1606162592,
 267         },
 268         'skip': 'Requires logging in',
 269     }, {
 270         # node.comet_sections.content.story.attached_story.attachments.style_type_renderer.attachment.media
 271         'url': 'https://www.facebook.com/groups/ateistiskselskab/permalink/10154930137678856/',
 272         'info_dict': {
 273             'id': '211567722618337',
 274             'ext': 'mp4',
 275             'title': 'Facebook video #211567722618337',
 276             'uploader_id': '127875227654254',
 277             'upload_date': '20161122',
 278             'timestamp': 1479793574,
 279         },
 280     }, {
 281         # data.video.creation_story.attachments[].media
 282         'url': 'https://www.facebook.com/watch/live/?v=1823658634322275',
 283         'only_matching': True,
 284     }, {
 285         'url': 'https://www.facebook.com/watchparty/211641140192478',
 286         'info_dict': {
 287             'id': '211641140192478',
 288         },
 289         'playlist_count': 1,
 290         'skip': 'Requires logging in',
 291     }]
 292     _SUPPORTED_PAGLETS_REGEX = r'(?:pagelet_group_mall|permalink_video_pagelet|hyperfeed_story_id_[0-9a-f]+)'
 293     _api_config = {
 294         'graphURI': '/api/graphql/'
 295     }
 296
 297     @staticmethod
 298     def _extract_urls(webpage):
 299         urls = []
 300         for mobj in re.finditer(
 301                 r'<iframe[^>]+?src=(["\'])(?P<url>https?://www\.facebook\.com/(?:video/embed|plugins/video\.php).+?)\1',
 302                 webpage):
 303             urls.append(mobj.group('url'))
 304         # Facebook API embed
 305         # see https://developers.facebook.com/docs/plugins/embedded-video-player
 306         for mobj in re.finditer(r'''(?x)<div[^>]+
 307                 class=(?P<q1>[\'"])[^\'"]*\bfb-(?:video|post)\b[^\'"]*(?P=q1)[^>]+
 308                 data-href=(?P<q2>[\'"])(?P<url>(?:https?:)?//(?:www\.)?facebook.com/.+?)(?P=q2)''', webpage):
 309             urls.append(mobj.group('url'))
 310         return urls
 311
 312     def _login(self):
 313         useremail, password = self._get_login_info()
 314         if useremail is None:
 315             return
 316
 317         login_page_req = sanitized_Request(self._LOGIN_URL)
 318         self._set_cookie('facebook.com', 'locale', 'en_US')
 319         login_page = self._download_webpage(login_page_req, None,
 320                                             note='Downloading login page',
 321                                             errnote='Unable to download login page')
 322         lsd = self._search_regex(
 323             r'<input type="hidden" name="lsd" value="([^"]*)"',
 324             login_page, 'lsd')
 325         lgnrnd = self._search_regex(r'name="lgnrnd" value="([^"]*?)"', login_page, 'lgnrnd')
 326
 327         login_form = {
 328             'email': useremail,
 329             'pass': password,
 330             'lsd': lsd,
 331             'lgnrnd': lgnrnd,
 332             'next': 'http://facebook.com/home.php',
 333             'default_persistent': '0',
 334             'legacy_return': '1',
 335             'timezone': '-60',
 336             'trynum': '1',
 337         }
 338         request = sanitized_Request(self._LOGIN_URL, urlencode_postdata(login_form))
 339         request.add_header('Content-Type', 'application/x-www-form-urlencoded')
 340         try:
 341             login_results = self._download_webpage(request, None,
 342                                                    note='Logging in', errnote='unable to fetch login page')
 343             if re.search(r'<form(.*)name="login"(.*)</form>', login_results) is not None:
 344                 error = self._html_search_regex(
 345                     r'(?s)<div[^>]+class=(["\']).*?login_error_box.*?\1[^>]*><div[^>]*>.*?</div><div[^>]*>(?P<error>.+?)</div>',
 346                     login_results, 'login error', default=None, group='error')
 347                 if error:
 348                     raise ExtractorError('Unable to login: %s' % error, expected=True)
 349                 self.report_warning('unable to log in: bad username/password, or exceeded login rate limit (~3/min). Check credentials or wait.')
 350                 return
 351
 352             fb_dtsg = self._search_regex(
 353                 r'name="fb_dtsg" value="(.+?)"', login_results, 'fb_dtsg', default=None)
 354             h = self._search_regex(
 355                 r'name="h"\s+(?:\w+="[^"]+"\s+)*?value="([^"]+)"', login_results, 'h', default=None)
 356
 357             if not fb_dtsg or not h:
 358                 return
 359
 360             check_form = {
 361                 'fb_dtsg': fb_dtsg,
 362                 'h': h,
 363                 'name_action_selected': 'dont_save',
 364             }
 365             check_req = sanitized_Request(self._CHECKPOINT_URL, urlencode_postdata(check_form))
 366             check_req.add_header('Content-Type', 'application/x-www-form-urlencoded')
 367             check_response = self._download_webpage(check_req, None,
 368                                                     note='Confirming login')
 369             if re.search(r'id="checkpointSubmitButton"', check_response) is not None:
 370                 self.report_warning('Unable to confirm login, you have to login in your browser and authorize the login.')
 371         except network_exceptions as err:
 372             self.report_warning('unable to log in: %s' % error_to_compat_str(err))
 373             return
 374
 375     def _real_initialize(self):
 376         self._login()
 377
 378     def _extract_from_url(self, url, video_id):
 379         webpage = self._download_webpage(
 380             url.replace('://m.facebook.com/', '://www.facebook.com/'), video_id)
 381
 382         video_data = None
 383
 384         def extract_video_data(instances):
 385             video_data = []
 386             for item in instances:
 387                 if try_get(item, lambda x: x[1][0]) == 'VideoConfig':
 388                     video_item = item[2][0]
 389                     if video_item.get('video_id'):
 390                         video_data.append(video_item['videoData'])
 391             return video_data
 392
 393         server_js_data = self._parse_json(self._search_regex(
 394             [r'handleServerJS\(({.+})(?:\);|,")', r'\bs\.handle\(({.+?})\);'],
 395             webpage, 'server js data', default='{}'), video_id, fatal=False)
 396
 397         if server_js_data:
 398             video_data = extract_video_data(server_js_data.get('instances', []))
 399
 400         def extract_from_jsmods_instances(js_data):
 401             if js_data:
 402                 return extract_video_data(try_get(
 403                     js_data, lambda x: x['jsmods']['instances'], list) or [])
 404
 405         def extract_dash_manifest(video, formats):
 406             dash_manifest = video.get('dash_manifest')
 407             if dash_manifest:
 408                 formats.extend(self._parse_mpd_formats(
 409                     compat_etree_fromstring(compat_urllib_parse_unquote_plus(dash_manifest))))
 410
 411         def process_formats(formats):
 412             # Downloads with browser's User-Agent are rate limited. Working around
 413             # with non-browser User-Agent.
 414             for f in formats:
 415                 f.setdefault('http_headers', {})['User-Agent'] = 'facebookexternalhit/1.1'
 416
 417             self._sort_formats(formats)
 418
 419         def extract_relay_data(_filter):
 420             return self._parse_json(self._search_regex(
 421                 r'handleWithCustomApplyEach\([^,]+,\s*({.*?%s.*?})\);' % _filter,
 422                 webpage, 'replay data', default='{}'), video_id, fatal=False) or {}
 423
 424         def extract_relay_prefetched_data(_filter):
 425             replay_data = extract_relay_data(_filter)
 426             for require in (replay_data.get('require') or []):
 427                 if require[0] == 'RelayPrefetchedStreamCache':
 428                     return try_get(require, lambda x: x[3][1]['__bbox']['result']['data'], dict) or {}
 429
 430         if not video_data:
 431             server_js_data = self._parse_json(self._search_regex([
 432                 r'bigPipe\.onPageletArrive\(({.+?})\)\s*;\s*}\s*\)\s*,\s*["\']onPageletArrive\s+' + self._SUPPORTED_PAGLETS_REGEX,
 433                 r'bigPipe\.onPageletArrive\(({.*?id\s*:\s*"%s".*?})\);' % self._SUPPORTED_PAGLETS_REGEX
 434             ], webpage, 'js data', default='{}'), video_id, js_to_json, False)
 435             video_data = extract_from_jsmods_instances(server_js_data)
 436
 437         if not video_data:
 438             data = extract_relay_prefetched_data(
 439                 r'"(?:dash_manifest|playable_url(?:_quality_hd)?)"\s*:\s*"[^"]+"')
 440             if data:
 441                 entries = []
 442
 443                 def parse_graphql_video(video):
 444                     formats = []
 445                     q = qualities(['sd', 'hd'])
 446                     for (suffix, format_id) in [('', 'sd'), ('_quality_hd', 'hd')]:
 447                         playable_url = video.get('playable_url' + suffix)
 448                         if not playable_url:
 449                             continue
 450                         formats.append({
 451                             'format_id': format_id,
 452                             'quality': q(format_id),
 453                             'url': playable_url,
 454                         })
 455                     extract_dash_manifest(video, formats)
 456                     process_formats(formats)
 457                     v_id = video.get('videoId') or video.get('id') or video_id
 458                     info = {
 459                         'id': v_id,
 460                         'formats': formats,
 461                         'thumbnail': try_get(video, lambda x: x['thumbnailImage']['uri']),
 462                         'uploader_id': try_get(video, lambda x: x['owner']['id']),
 463                         'timestamp': int_or_none(video.get('publish_time')),
 464                         'duration': float_or_none(video.get('playable_duration_in_ms'), 1000),
 465                     }
 466                     description = try_get(video, lambda x: x['savable_description']['text'])
 467                     title = video.get('name')
 468                     if title:
 469                         info.update({
 470                             'title': title,
 471                             'description': description,
 472                         })
 473                     else:
 474                         info['title'] = description or 'Facebook video #%s' % v_id
 475                     entries.append(info)
 476
 477                 def parse_attachment(attachment, key='media'):
 478                     media = attachment.get(key) or {}
 479                     if media.get('__typename') == 'Video':
 480                         return parse_graphql_video(media)
 481
 482                 nodes = data.get('nodes') or []
 483                 node = data.get('node') or {}
 484                 if not nodes and node:
 485                     nodes.append(node)
 486                 for node in nodes:
 487                     story = try_get(node, lambda x: x['comet_sections']['content']['story'], dict) or {}
 488                     attachments = try_get(story, [
 489                         lambda x: x['attached_story']['attachments'],
 490                         lambda x: x['attachments']
 491                     ], list) or []
 492                     for attachment in attachments:
 493                         attachment = try_get(attachment, lambda x: x['style_type_renderer']['attachment'], dict)
 494                         ns = try_get(attachment, lambda x: x['all_subattachments']['nodes'], list) or []
 495                         for n in ns:
 496                             parse_attachment(n)
 497                         parse_attachment(attachment)
 498
 499                 edges = try_get(data, lambda x: x['mediaset']['currMedia']['edges'], list) or []
 500                 for edge in edges:
 501                     parse_attachment(edge, key='node')
 502
 503                 video = data.get('video') or {}
 504                 if video:
 505                     attachments = try_get(video, [
 506                         lambda x: x['story']['attachments'],
 507                         lambda x: x['creation_story']['attachments']
 508                     ], list) or []
 509                     for attachment in attachments:
 510                         parse_attachment(attachment)
 511                     if not entries:
 512                         parse_graphql_video(video)
 513
 514                 return self.playlist_result(entries, video_id)
 515
 516         if not video_data:
 517             m_msg = re.search(r'class="[^"]*uiInterstitialContent[^"]*"><div>(.*?)</div>', webpage)
 518             if m_msg is not None:
 519                 raise ExtractorError(
 520                     'The video is not available, Facebook said: "%s"' % m_msg.group(1),
 521                     expected=True)
 522             elif '>You must log in to continue' in webpage:
 523                 self.raise_login_required()
 524
 525         if not video_data and '/watchparty/' in url:
 526             post_data = {
 527                 'doc_id': 3731964053542869,
 528                 'variables': json.dumps({
 529                     'livingRoomID': video_id,
 530                 }),
 531             }
 532
 533             prefetched_data = extract_relay_prefetched_data(r'"login_data"\s*:\s*{')
 534             if prefetched_data:
 535                 lsd = try_get(prefetched_data, lambda x: x['login_data']['lsd'], dict)
 536                 if lsd:
 537                     post_data[lsd['name']] = lsd['value']
 538
 539             relay_data = extract_relay_data(r'\[\s*"RelayAPIConfigDefaults"\s*,')
 540             for define in (relay_data.get('define') or []):
 541                 if define[0] == 'RelayAPIConfigDefaults':
 542                     self._api_config = define[2]
 543
 544             living_room = self._download_json(
 545                 urljoin(url, self._api_config['graphURI']), video_id,
 546                 data=urlencode_postdata(post_data))['data']['living_room']
 547
 548             entries = []
 549             for edge in (try_get(living_room, lambda x: x['recap']['watched_content']['edges']) or []):
 550                 video = try_get(edge, lambda x: x['node']['video']) or {}
 551                 v_id = video.get('id')
 552                 if not v_id:
 553                     continue
 554                 v_id = compat_str(v_id)
 555                 entries.append(self.url_result(
 556                     self._VIDEO_PAGE_TEMPLATE % v_id,
 557                     self.ie_key(), v_id, video.get('name')))
 558
 559             return self.playlist_result(entries, video_id)
 560
 561         if not video_data:
 562             # Video info not in first request, do a secondary request using
 563             # tahoe player specific URL
 564             tahoe_data = self._download_webpage(
 565                 self._VIDEO_PAGE_TAHOE_TEMPLATE % video_id, video_id,
 566                 data=urlencode_postdata({
 567                     '__a': 1,
 568                     '__pc': self._search_regex(
 569                         r'pkg_cohort["\']\s*:\s*["\'](.+?)["\']', webpage,
 570                         'pkg cohort', default='PHASED:DEFAULT'),
 571                     '__rev': self._search_regex(
 572                         r'client_revision["\']\s*:\s*(\d+),', webpage,
 573                         'client revision', default='3944515'),
 574                     'fb_dtsg': self._search_regex(
 575                         r'"DTSGInitialData"\s*,\s*\[\]\s*,\s*{\s*"token"\s*:\s*"([^"]+)"',
 576                         webpage, 'dtsg token', default=''),
 577                 }),
 578                 headers={
 579                     'Content-Type': 'application/x-www-form-urlencoded',
 580                 })
 581             tahoe_js_data = self._parse_json(
 582                 self._search_regex(
 583                     r'for\s+\(\s*;\s*;\s*\)\s*;(.+)', tahoe_data,
 584                     'tahoe js data', default='{}'),
 585                 video_id, fatal=False)
 586             video_data = extract_from_jsmods_instances(tahoe_js_data)
 587
 588         if not video_data:
 589             raise ExtractorError('Cannot parse data')
 590
 591         if len(video_data) > 1:
 592             entries = []
 593             for v in video_data:
 594                 video_url = v[0].get('video_url')
 595                 if not video_url:
 596                     continue
 597                 entries.append(self.url_result(urljoin(
 598                     url, video_url), self.ie_key(), v[0].get('video_id')))
 599             return self.playlist_result(entries, video_id)
 600         video_data = video_data[0]
 601
 602         formats = []
 603         subtitles = {}
 604         for f in video_data:
 605             format_id = f['stream_type']
 606             if f and isinstance(f, dict):
 607                 f = [f]
 608             if not f or not isinstance(f, list):
 609                 continue
 610             for quality in ('sd', 'hd'):
 611                 for src_type in ('src', 'src_no_ratelimit'):
 612                     src = f[0].get('%s_%s' % (quality, src_type))
 613                     if src:
 614                         preference = -10 if format_id == 'progressive' else 0
 615                         if quality == 'hd':
 616                             preference += 5
 617                         formats.append({
 618                             'format_id': '%s_%s_%s' % (format_id, quality, src_type),
 619                             'url': src,
 620                             'quality': preference,
 621                         })
 622             extract_dash_manifest(f[0], formats)
 623             subtitles_src = f[0].get('subtitles_src')
 624             if subtitles_src:
 625                 subtitles.setdefault('en', []).append({'url': subtitles_src})
 626
 627         process_formats(formats)
 628
 629         video_title = self._html_search_regex(
 630             r'<h2\s+[^>]*class="uiHeaderTitle"[^>]*>([^<]*)</h2>', webpage,
 631             'title', default=None)
 632         if not video_title:
 633             video_title = self._html_search_regex(
 634                 r'(?s)<span class="fbPhotosPhotoCaption".*?id="fbPhotoPageCaption"><span class="hasCaption">(.*?)</span>',
 635                 webpage, 'alternative title', default=None)
 636         if not video_title:
 637             video_title = self._html_search_meta(
 638                 'description', webpage, 'title', default=None)
 639         if video_title:
 640             video_title = limit_length(video_title, 80)
 641         else:
 642             video_title = 'Facebook video #%s' % video_id
 643         uploader = clean_html(get_element_by_id(
 644             'fbPhotoPageAuthorName', webpage)) or self._search_regex(
 645             r'ownerName\s*:\s*"([^"]+)"', webpage, 'uploader',
 646             default=None) or self._og_search_title(webpage, fatal=False)
 647         timestamp = int_or_none(self._search_regex(
 648             r'<abbr[^>]+data-utime=["\'](\d+)', webpage,
 649             'timestamp', default=None))
 650         thumbnail = self._html_search_meta(['og:image', 'twitter:image'], webpage)
 651
 652         view_count = parse_count(self._search_regex(
 653             r'\bviewCount\s*:\s*["\']([\d,.]+)', webpage, 'view count',
 654             default=None))
 655
 656         info_dict = {
 657             'id': video_id,
 658             'title': video_title,
 659             'formats': formats,
 660             'uploader': uploader,
 661             'timestamp': timestamp,
 662             'thumbnail': thumbnail,
 663             'view_count': view_count,
 664             'subtitles': subtitles,
 665         }
 666
 667         return info_dict
 668
 669     def _real_extract(self, url):
 670         video_id = self._match_id(url)
 671
 672         real_url = self._VIDEO_PAGE_TEMPLATE % video_id if url.startswith('facebook:') else url
 673         return self._extract_from_url(real_url, video_id)
 674
 675
 676 class FacebookPluginsVideoIE(InfoExtractor):
 677     _VALID_URL = r'https?://(?:[\w-]+\.)?facebook\.com/plugins/video\.php\?.*?\bhref=(?P<id>https.+)'
 678
 679     _TESTS = [{
 680         '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',
 681         'md5': '5954e92cdfe51fe5782ae9bda7058a07',
 682         'info_dict': {
 683             'id': '10154383743583686',
 684             'ext': 'mp4',
 685             'title': 'What to do during the haze?',
 686             'uploader': 'Gov.sg',
 687             'upload_date': '20160826',
 688             'timestamp': 1472184808,
 689         },
 690         'add_ie': [FacebookIE.ie_key()],
 691     }, {
 692         'url': 'https://www.facebook.com/plugins/video.php?href=https%3A%2F%2Fwww.facebook.com%2Fvideo.php%3Fv%3D10204634152394104',
 693         'only_matching': True,
 694     }, {
 695         'url': 'https://www.facebook.com/plugins/video.php?href=https://www.facebook.com/gov.sg/videos/10154383743583686/&show_text=0&width=560',
 696         'only_matching': True,
 697     }]
 698
 699     def _real_extract(self, url):
 700         return self.url_result(
 701             compat_urllib_parse_unquote(self._match_id(url)),
 702             FacebookIE.ie_key())