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     merge_dicts,
  24     network_exceptions,
  25     parse_count,
  26     qualities,
  27     sanitized_Request,
  28     try_get,
  29     urlencode_postdata,
  30     urljoin,
  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                                 video/video\.php|
  43                                 photo\.php|
  44                                 video\.php|
  45                                 video/embed|
  46                                 story\.php|
  47                                 watch(?:/live)?/?
  48                             )\?(?:.*?)(?:v|video_id|story_fbid)=|
  49                             [^/]+/videos/(?:[^/]+/)?|
  50                             [^/]+/posts/|
  51                             groups/[^/]+/permalink/|
  52                             watchparty/
  53                         )|
  54                     facebook:
  55                 )
  56                 (?P<id>[0-9]+)
  57                 '''
  58     _LOGIN_URL = 'https://www.facebook.com/login.php?next=http%3A%2F%2Ffacebook.com%2Fhome.php&login_attempt=1'
  59     _CHECKPOINT_URL = 'https://www.facebook.com/checkpoint/?next=http%3A%2F%2Ffacebook.com%2Fhome.php&_fb_noscript=1'
  60     _NETRC_MACHINE = 'facebook'
  61     IE_NAME = 'facebook'
  62
  63     _VIDEO_PAGE_TEMPLATE = 'https://www.facebook.com/video/video.php?v=%s'
  64     _VIDEO_PAGE_TAHOE_TEMPLATE = 'https://www.facebook.com/video/tahoe/async/%s/?chain=true&isvideo=true&payloadtype=primary'
  65
  66     _TESTS = [{
  67         'url': 'https://www.facebook.com/video.php?v=637842556329505&fref=nf',
  68         'md5': '6a40d33c0eccbb1af76cf0485a052659',
  69         'info_dict': {
  70             'id': '637842556329505',
  71             'ext': 'mp4',
  72             'title': 're:Did you know Kei Nishikori is the first Asian man to ever reach a Grand Slam',
  73             'uploader': 'Tennis on Facebook',
  74             'upload_date': '20140908',
  75             'timestamp': 1410199200,
  76         },
  77         'skip': 'Requires logging in',
  78     }, {
  79         # data.video
  80         'url': 'https://www.facebook.com/video.php?v=274175099429670',
  81         'info_dict': {
  82             'id': '274175099429670',
  83             'ext': 'mp4',
  84             'title': 'Asif Nawab Butt',
  85             'description': 'Asif Nawab Butt',
  86             'uploader': 'Asif Nawab Butt',
  87             'upload_date': '20140506',
  88             'timestamp': 1399398998,
  89             'thumbnail': r're:^https?://.*',
  90         },
  91         'expected_warnings': [
  92             'title'
  93         ]
  94     }, {
  95         'note': 'Video with DASH manifest',
  96         'url': 'https://www.facebook.com/video.php?v=957955867617029',
  97         'md5': 'b2c28d528273b323abe5c6ab59f0f030',
  98         'info_dict': {
  99             'id': '957955867617029',
 100             'ext': 'mp4',
 101             'title': 'When you post epic content on instagram.com/433 8 million followers, this is ...',
 102             'uploader': 'Demy de Zeeuw',
 103             'upload_date': '20160110',
 104             'timestamp': 1452431627,
 105         },
 106         'skip': 'Requires logging in',
 107     }, {
 108         'url': 'https://www.facebook.com/maxlayn/posts/10153807558977570',
 109         'md5': '037b1fa7f3c2d02b7a0d7bc16031ecc6',
 110         'info_dict': {
 111             'id': '544765982287235',
 112             'ext': 'mp4',
 113             'title': '"What are you doing running in the snow?"',
 114             'uploader': 'FailArmy',
 115         },
 116         'skip': 'Video gone',
 117     }, {
 118         'url': 'https://m.facebook.com/story.php?story_fbid=1035862816472149&id=116132035111903',
 119         'md5': '1deb90b6ac27f7efcf6d747c8a27f5e3',
 120         'info_dict': {
 121             'id': '1035862816472149',
 122             'ext': 'mp4',
 123             'title': 'What the Flock Is Going On In New Zealand  Credit: ViralHog',
 124             'uploader': 'S. Saint',
 125         },
 126         'skip': 'Video gone',
 127     }, {
 128         'note': 'swf params escaped',
 129         'url': 'https://www.facebook.com/barackobama/posts/10153664894881749',
 130         'md5': '97ba073838964d12c70566e0085c2b91',
 131         'info_dict': {
 132             'id': '10153664894881749',
 133             'ext': 'mp4',
 134             'title': 'Average time to confirm recent Supreme Court nominees: 67 days Longest it\'s t...',
 135             'thumbnail': r're:^https?://.*',
 136             'timestamp': 1456259628,
 137             'upload_date': '20160223',
 138             'uploader': 'Barack Obama',
 139         },
 140         'skip': 'Gif on giphy.com gone',
 141     }, {
 142         # have 1080P, but only up to 720p in swf params
 143         # data.video.story.attachments[].media
 144         'url': 'https://www.facebook.com/cnn/videos/10155529876156509/',
 145         'md5': '3f3798adb2b73423263e59376f1f5eb7',
 146         'info_dict': {
 147             'id': '10155529876156509',
 148             'ext': 'mp4',
 149             'title': 'Holocaust survivor becomes US citizen',
 150             'description': 'She survived the holocaust — and years later, she’s getting her citizenship so she can vote for Hillary Clinton http://cnn.it/2eERh5f',
 151             'timestamp': 1477818095,
 152             'upload_date': '20161030',
 153             'uploader': 'CNN',
 154             'thumbnail': r're:^https?://.*',
 155             'view_count': int,
 156         },
 157     }, {
 158         # bigPipe.onPageletArrive ... onPageletArrive pagelet_group_mall
 159         # data.node.comet_sections.content.story.attachments[].style_type_renderer.attachment.media
 160         'url': 'https://www.facebook.com/yaroslav.korpan/videos/1417995061575415/',
 161         'info_dict': {
 162             'id': '1417995061575415',
 163             'ext': 'mp4',
 164             'title': 'Yaroslav Korpan - Довгоочікуване відео',
 165             'description': 'Довгоочікуване відео',
 166             'timestamp': 1486648771,
 167             'upload_date': '20170209',
 168             'uploader': 'Yaroslav Korpan',
 169             'uploader_id': '100000948048708',
 170         },
 171         'params': {
 172             'skip_download': True,
 173         },
 174     }, {
 175         # FIXME
 176         'url': 'https://www.facebook.com/LaGuiaDelVaron/posts/1072691702860471',
 177         'info_dict': {
 178             'id': '1072691702860471',
 179             'ext': 'mp4',
 180             'title': 'md5:ae2d22a93fbb12dad20dc393a869739d',
 181             'timestamp': 1477305000,
 182             'upload_date': '20161024',
 183             'uploader': 'La Guía Del Varón',
 184             'thumbnail': r're:^https?://.*',
 185         },
 186         'params': {
 187             'skip_download': True,
 188         },
 189     }, {
 190         # data.node.comet_sections.content.story.attachments[].style_type_renderer.attachment.media
 191         'url': 'https://www.facebook.com/groups/1024490957622648/permalink/1396382447100162/',
 192         'info_dict': {
 193             'id': '202882990186699',
 194             'ext': 'mp4',
 195             'title': 'Elisabeth Ahtn - Hello? Yes your uber ride is here\n* Jukin...',
 196             'description': 'Hello? Yes your uber ride is here\n* Jukin Media Verified *\nFind this video and others like it by visiting...',
 197             'timestamp': 1486035513,
 198             'upload_date': '20170202',
 199             'uploader': 'Elisabeth Ahtn',
 200             'uploader_id': '100013949973717',
 201         },
 202         'params': {
 203             'skip_download': True,
 204         },
 205     }, {
 206         'url': 'https://www.facebook.com/video.php?v=10204634152394104',
 207         'only_matching': True,
 208     }, {
 209         'url': 'https://www.facebook.com/amogood/videos/1618742068337349/?fref=nf',
 210         'only_matching': True,
 211     }, {
 212         # data.mediaset.currMedia.edges
 213         'url': 'https://www.facebook.com/ChristyClarkForBC/videos/vb.22819070941/10153870694020942/?type=2&theater',
 214         'only_matching': True,
 215     }, {
 216         # data.video.story.attachments[].media
 217         'url': 'facebook:544765982287235',
 218         'only_matching': True,
 219     }, {
 220         # data.node.comet_sections.content.story.attachments[].style_type_renderer.attachment.media
 221         'url': 'https://www.facebook.com/groups/164828000315060/permalink/764967300301124/',
 222         'only_matching': True,
 223     }, {
 224         # data.video.creation_story.attachments[].media
 225         'url': 'https://zh-hk.facebook.com/peoplespower/videos/1135894589806027/',
 226         'only_matching': True,
 227     }, {
 228         # data.video
 229         'url': 'https://www.facebookwkhpilnemxj7asaniu7vnjjbiltxjqhye3mhbshg7kx5tfyd.onion/video.php?v=274175099429670',
 230         'only_matching': True,
 231     }, {
 232         # no title
 233         'url': 'https://www.facebook.com/onlycleverentertainment/videos/1947995502095005/',
 234         'only_matching': True,
 235     }, {
 236         # data.video
 237         'url': 'https://www.facebook.com/WatchESLOne/videos/359649331226507/',
 238         'info_dict': {
 239             'id': '359649331226507',
 240             'ext': 'mp4',
 241             'title': 'Fnatic vs. EG - Group A - Opening Match - ESL One Birmingham Day 1',
 242             'description': '#ESLOne VoD - Birmingham Finals Day#1 Fnatic vs. @Evil Geniuses',
 243             'timestamp': 1527084179,
 244             'upload_date': '20180523',
 245             'uploader': 'ESL One Dota 2',
 246             'uploader_id': '234218833769558',
 247         },
 248         'params': {
 249             'skip_download': True,
 250         },
 251     }, {
 252         # data.node.comet_sections.content.story.attachments[].style_type_renderer.attachment.all_subattachments.nodes[].media
 253         'url': 'https://www.facebook.com/100033620354545/videos/106560053808006/',
 254         'info_dict': {
 255             'id': '106560053808006',
 256         },
 257         'playlist_count': 2,
 258     }, {
 259         # data.video.story.attachments[].media
 260         'url': 'https://www.facebook.com/watch/?v=647537299265662',
 261         'only_matching': True,
 262     }, {
 263         # FIXME: https://github.com/yt-dlp/yt-dlp/issues/542
 264         # data.node.comet_sections.content.story.attachments[].style_type_renderer.attachment.all_subattachments.nodes[].media
 265         'url': 'https://www.facebook.com/PankajShahLondon/posts/10157667649866271',
 266         'info_dict': {
 267             'id': '10157667649866271',
 268         },
 269         'playlist_count': 3,
 270     }, {
 271         # data.nodes[].comet_sections.content.story.attachments[].style_type_renderer.attachment.media
 272         'url': 'https://m.facebook.com/Alliance.Police.Department/posts/4048563708499330',
 273         'info_dict': {
 274             'id': '117576630041613',
 275             'ext': 'mp4',
 276             # TODO: title can be extracted from video page
 277             'title': 'Facebook video #117576630041613',
 278             'uploader_id': '189393014416438',
 279             'upload_date': '20201123',
 280             'timestamp': 1606162592,
 281         },
 282         'skip': 'Requires logging in',
 283     }, {
 284         # node.comet_sections.content.story.attached_story.attachments.style_type_renderer.attachment.media
 285         'url': 'https://www.facebook.com/groups/ateistiskselskab/permalink/10154930137678856/',
 286         'info_dict': {
 287             'id': '211567722618337',
 288             'ext': 'mp4',
 289             'title': 'Facebook video #211567722618337',
 290             'uploader_id': '127875227654254',
 291             'upload_date': '20161122',
 292             'timestamp': 1479793574,
 293         },
 294         'skip': 'No video',
 295     }, {
 296         # data.video.creation_story.attachments[].media
 297         'url': 'https://www.facebook.com/watch/live/?v=1823658634322275',
 298         'only_matching': True,
 299     }, {
 300         'url': 'https://www.facebook.com/watchparty/211641140192478',
 301         'info_dict': {
 302             'id': '211641140192478',
 303         },
 304         'playlist_count': 1,
 305         'skip': 'Requires logging in',
 306     }]
 307     _SUPPORTED_PAGLETS_REGEX = r'(?:pagelet_group_mall|permalink_video_pagelet|hyperfeed_story_id_[0-9a-f]+)'
 308     _api_config = {
 309         'graphURI': '/api/graphql/'
 310     }
 311
 312     @staticmethod
 313     def _extract_urls(webpage):
 314         urls = []
 315         for mobj in re.finditer(
 316                 r'<iframe[^>]+?src=(["\'])(?P<url>https?://www\.facebook\.com/(?:video/embed|plugins/video\.php).+?)\1',
 317                 webpage):
 318             urls.append(mobj.group('url'))
 319         # Facebook API embed
 320         # see https://developers.facebook.com/docs/plugins/embedded-video-player
 321         for mobj in re.finditer(r'''(?x)<div[^>]+
 322                 class=(?P<q1>[\'"])[^\'"]*\bfb-(?:video|post)\b[^\'"]*(?P=q1)[^>]+
 323                 data-href=(?P<q2>[\'"])(?P<url>(?:https?:)?//(?:www\.)?facebook.com/.+?)(?P=q2)''', webpage):
 324             urls.append(mobj.group('url'))
 325         return urls
 326
 327     def _login(self):
 328         useremail, password = self._get_login_info()
 329         if useremail is None:
 330             return
 331
 332         login_page_req = sanitized_Request(self._LOGIN_URL)
 333         self._set_cookie('facebook.com', 'locale', 'en_US')
 334         login_page = self._download_webpage(login_page_req, None,
 335                                             note='Downloading login page',
 336                                             errnote='Unable to download login page')
 337         lsd = self._search_regex(
 338             r'<input type="hidden" name="lsd" value="([^"]*)"',
 339             login_page, 'lsd')
 340         lgnrnd = self._search_regex(r'name="lgnrnd" value="([^"]*?)"', login_page, 'lgnrnd')
 341
 342         login_form = {
 343             'email': useremail,
 344             'pass': password,
 345             'lsd': lsd,
 346             'lgnrnd': lgnrnd,
 347             'next': 'http://facebook.com/home.php',
 348             'default_persistent': '0',
 349             'legacy_return': '1',
 350             'timezone': '-60',
 351             'trynum': '1',
 352         }
 353         request = sanitized_Request(self._LOGIN_URL, urlencode_postdata(login_form))
 354         request.add_header('Content-Type', 'application/x-www-form-urlencoded')
 355         try:
 356             login_results = self._download_webpage(request, None,
 357                                                    note='Logging in', errnote='unable to fetch login page')
 358             if re.search(r'<form(.*)name="login"(.*)</form>', login_results) is not None:
 359                 error = self._html_search_regex(
 360                     r'(?s)<div[^>]+class=(["\']).*?login_error_box.*?\1[^>]*><div[^>]*>.*?</div><div[^>]*>(?P<error>.+?)</div>',
 361                     login_results, 'login error', default=None, group='error')
 362                 if error:
 363                     raise ExtractorError('Unable to login: %s' % error, expected=True)
 364                 self.report_warning('unable to log in: bad username/password, or exceeded login rate limit (~3/min). Check credentials or wait.')
 365                 return
 366
 367             fb_dtsg = self._search_regex(
 368                 r'name="fb_dtsg" value="(.+?)"', login_results, 'fb_dtsg', default=None)
 369             h = self._search_regex(
 370                 r'name="h"\s+(?:\w+="[^"]+"\s+)*?value="([^"]+)"', login_results, 'h', default=None)
 371
 372             if not fb_dtsg or not h:
 373                 return
 374
 375             check_form = {
 376                 'fb_dtsg': fb_dtsg,
 377                 'h': h,
 378                 'name_action_selected': 'dont_save',
 379             }
 380             check_req = sanitized_Request(self._CHECKPOINT_URL, urlencode_postdata(check_form))
 381             check_req.add_header('Content-Type', 'application/x-www-form-urlencoded')
 382             check_response = self._download_webpage(check_req, None,
 383                                                     note='Confirming login')
 384             if re.search(r'id="checkpointSubmitButton"', check_response) is not None:
 385                 self.report_warning('Unable to confirm login, you have to login in your browser and authorize the login.')
 386         except network_exceptions as err:
 387             self.report_warning('unable to log in: %s' % error_to_compat_str(err))
 388             return
 389
 390     def _real_initialize(self):
 391         self._login()
 392
 393     def _extract_from_url(self, url, video_id):
 394         webpage = self._download_webpage(
 395             url.replace('://m.facebook.com/', '://www.facebook.com/'), video_id)
 396
 397         def extract_metadata(webpage):
 398             video_title = self._html_search_regex(
 399                 r'<h2\s+[^>]*class="uiHeaderTitle"[^>]*>([^<]*)</h2>', webpage,
 400                 'title', default=None)
 401             if not video_title:
 402                 video_title = self._html_search_regex(
 403                     r'(?s)<span class="fbPhotosPhotoCaption".*?id="fbPhotoPageCaption"><span class="hasCaption">(.*?)</span>',
 404                     webpage, 'alternative title', default=None)
 405             if not video_title:
 406                 video_title = self._html_search_meta(
 407                     ['og:title', 'twitter:title', 'description'],
 408                     webpage, 'title', default=None)
 409             if video_title:
 410                 video_title = limit_length(video_title, 80)
 411             else:
 412                 video_title = 'Facebook video #%s' % video_id
 413             description = self._html_search_meta(
 414                 ['description', 'og:description', 'twitter:description'],
 415                 webpage, 'description', default=None)
 416             uploader = clean_html(get_element_by_id(
 417                 'fbPhotoPageAuthorName', webpage)) or self._search_regex(
 418                 r'ownerName\s*:\s*"([^"]+)"', webpage, 'uploader',
 419                 default=None) or self._og_search_title(webpage, fatal=False)
 420             timestamp = int_or_none(self._search_regex(
 421                 r'<abbr[^>]+data-utime=["\'](\d+)', webpage,
 422                 'timestamp', default=None))
 423             thumbnail = self._html_search_meta(
 424                 ['og:image', 'twitter:image'], webpage, 'thumbnail', default=None)
 425             # some webpages contain unretrievable thumbnail urls
 426             # like https://lookaside.fbsbx.com/lookaside/crawler/media/?media_id=10155168902769113&get_thumbnail=1
 427             # in https://www.facebook.com/yaroslav.korpan/videos/1417995061575415/
 428             if thumbnail and not re.search(r'\.(?:jpg|png)', thumbnail):
 429                 thumbnail = None
 430             view_count = parse_count(self._search_regex(
 431                 r'\bviewCount\s*:\s*["\']([\d,.]+)', webpage, 'view count',
 432                 default=None))
 433             info_dict = {
 434                 'title': video_title,
 435                 'description': description,
 436                 'uploader': uploader,
 437                 'timestamp': timestamp,
 438                 'thumbnail': thumbnail,
 439                 'view_count': view_count,
 440             }
 441             info_json_ld = self._search_json_ld(webpage, video_id, default={})
 442             if info_json_ld.get('title'):
 443                 info_json_ld['title'] = limit_length(
 444                     re.sub(r'\s*\|\s*Facebook$', '', info_json_ld['title']), 80)
 445             return merge_dicts(info_json_ld, info_dict)
 446
 447         video_data = None
 448
 449         def extract_video_data(instances):
 450             video_data = []
 451             for item in instances:
 452                 if try_get(item, lambda x: x[1][0]) == 'VideoConfig':
 453                     video_item = item[2][0]
 454                     if video_item.get('video_id'):
 455                         video_data.append(video_item['videoData'])
 456             return video_data
 457
 458         server_js_data = self._parse_json(self._search_regex(
 459             [r'handleServerJS\(({.+})(?:\);|,")', r'\bs\.handle\(({.+?})\);'],
 460             webpage, 'server js data', default='{}'), video_id, fatal=False)
 461
 462         if server_js_data:
 463             video_data = extract_video_data(server_js_data.get('instances', []))
 464
 465         def extract_from_jsmods_instances(js_data):
 466             if js_data:
 467                 return extract_video_data(try_get(
 468                     js_data, lambda x: x['jsmods']['instances'], list) or [])
 469
 470         def extract_dash_manifest(video, formats):
 471             dash_manifest = video.get('dash_manifest')
 472             if dash_manifest:
 473                 formats.extend(self._parse_mpd_formats(
 474                     compat_etree_fromstring(compat_urllib_parse_unquote_plus(dash_manifest))))
 475
 476         def process_formats(formats):
 477             # Downloads with browser's User-Agent are rate limited. Working around
 478             # with non-browser User-Agent.
 479             for f in formats:
 480                 f.setdefault('http_headers', {})['User-Agent'] = 'facebookexternalhit/1.1'
 481
 482             self._sort_formats(formats)
 483
 484         def extract_relay_data(_filter):
 485             return self._parse_json(self._search_regex(
 486                 r'handleWithCustomApplyEach\([^,]+,\s*({.*?%s.*?})\);' % _filter,
 487                 webpage, 'replay data', default='{}'), video_id, fatal=False) or {}
 488
 489         def extract_relay_prefetched_data(_filter):
 490             replay_data = extract_relay_data(_filter)
 491             for require in (replay_data.get('require') or []):
 492                 if require[0] == 'RelayPrefetchedStreamCache':
 493                     return try_get(require, lambda x: x[3][1]['__bbox']['result']['data'], dict) or {}
 494
 495         if not video_data:
 496             server_js_data = self._parse_json(self._search_regex([
 497                 r'bigPipe\.onPageletArrive\(({.+?})\)\s*;\s*}\s*\)\s*,\s*["\']onPageletArrive\s+' + self._SUPPORTED_PAGLETS_REGEX,
 498                 r'bigPipe\.onPageletArrive\(({.*?id\s*:\s*"%s".*?})\);' % self._SUPPORTED_PAGLETS_REGEX
 499             ], webpage, 'js data', default='{}'), video_id, js_to_json, False)
 500             video_data = extract_from_jsmods_instances(server_js_data)
 501
 502         if not video_data:
 503             data = extract_relay_prefetched_data(
 504                 r'"(?:dash_manifest|playable_url(?:_quality_hd)?)"\s*:\s*"[^"]+"')
 505             if data:
 506                 entries = []
 507
 508                 def parse_graphql_video(video):
 509                     formats = []
 510                     q = qualities(['sd', 'hd'])
 511                     for (suffix, format_id) in [('', 'sd'), ('_quality_hd', 'hd')]:
 512                         playable_url = video.get('playable_url' + suffix)
 513                         if not playable_url:
 514                             continue
 515                         formats.append({
 516                             'format_id': format_id,
 517                             'quality': q(format_id),
 518                             'url': playable_url,
 519                         })
 520                     extract_dash_manifest(video, formats)
 521                     process_formats(formats)
 522                     v_id = video.get('videoId') or video.get('id') or video_id
 523                     info = {
 524                         'id': v_id,
 525                         'formats': formats,
 526                         'thumbnail': try_get(video, lambda x: x['thumbnailImage']['uri']),
 527                         'uploader_id': try_get(video, lambda x: x['owner']['id']),
 528                         'timestamp': int_or_none(video.get('publish_time')),
 529                         'duration': float_or_none(video.get('playable_duration_in_ms'), 1000),
 530                     }
 531                     description = try_get(video, lambda x: x['savable_description']['text'])
 532                     title = video.get('name')
 533                     if title:
 534                         info.update({
 535                             'title': title,
 536                             'description': description,
 537                         })
 538                     else:
 539                         info['title'] = description or 'Facebook video #%s' % v_id
 540                     entries.append(info)
 541
 542                 def parse_attachment(attachment, key='media'):
 543                     media = attachment.get(key) or {}
 544                     if media.get('__typename') == 'Video':
 545                         return parse_graphql_video(media)
 546
 547                 nodes = data.get('nodes') or []
 548                 node = data.get('node') or {}
 549                 if not nodes and node:
 550                     nodes.append(node)
 551                 for node in nodes:
 552                     story = try_get(node, lambda x: x['comet_sections']['content']['story'], dict) or {}
 553                     attachments = try_get(story, [
 554                         lambda x: x['attached_story']['attachments'],
 555                         lambda x: x['attachments']
 556                     ], list) or []
 557                     for attachment in attachments:
 558                         attachment = try_get(attachment, lambda x: x['style_type_renderer']['attachment'], dict)
 559                         ns = try_get(attachment, lambda x: x['all_subattachments']['nodes'], list) or []
 560                         for n in ns:
 561                             parse_attachment(n)
 562                         parse_attachment(attachment)
 563
 564                 edges = try_get(data, lambda x: x['mediaset']['currMedia']['edges'], list) or []
 565                 for edge in edges:
 566                     parse_attachment(edge, key='node')
 567
 568                 video = data.get('video') or {}
 569                 if video:
 570                     attachments = try_get(video, [
 571                         lambda x: x['story']['attachments'],
 572                         lambda x: x['creation_story']['attachments']
 573                     ], list) or []
 574                     for attachment in attachments:
 575                         parse_attachment(attachment)
 576                     if not entries:
 577                         parse_graphql_video(video)
 578
 579                 if len(entries) > 1:
 580                     return self.playlist_result(entries, video_id)
 581
 582                 video_info = entries[0]
 583                 webpage_info = extract_metadata(webpage)
 584                 # honor precise duration in video info
 585                 if video_info.get('duration'):
 586                     webpage_info['duration'] = video_info['duration']
 587                 return merge_dicts(webpage_info, video_info)
 588
 589         if not video_data:
 590             m_msg = re.search(r'class="[^"]*uiInterstitialContent[^"]*"><div>(.*?)</div>', webpage)
 591             if m_msg is not None:
 592                 raise ExtractorError(
 593                     'The video is not available, Facebook said: "%s"' % m_msg.group(1),
 594                     expected=True)
 595             elif any(p in webpage for p in (
 596                     '>You must log in to continue',
 597                     'id="login_form"',
 598                     'id="loginbutton"')):
 599                 self.raise_login_required()
 600
 601         if not video_data and '/watchparty/' in url:
 602             post_data = {
 603                 'doc_id': 3731964053542869,
 604                 'variables': json.dumps({
 605                     'livingRoomID': video_id,
 606                 }),
 607             }
 608
 609             prefetched_data = extract_relay_prefetched_data(r'"login_data"\s*:\s*{')
 610             if prefetched_data:
 611                 lsd = try_get(prefetched_data, lambda x: x['login_data']['lsd'], dict)
 612                 if lsd:
 613                     post_data[lsd['name']] = lsd['value']
 614
 615             relay_data = extract_relay_data(r'\[\s*"RelayAPIConfigDefaults"\s*,')
 616             for define in (relay_data.get('define') or []):
 617                 if define[0] == 'RelayAPIConfigDefaults':
 618                     self._api_config = define[2]
 619
 620             living_room = self._download_json(
 621                 urljoin(url, self._api_config['graphURI']), video_id,
 622                 data=urlencode_postdata(post_data))['data']['living_room']
 623
 624             entries = []
 625             for edge in (try_get(living_room, lambda x: x['recap']['watched_content']['edges']) or []):
 626                 video = try_get(edge, lambda x: x['node']['video']) or {}
 627                 v_id = video.get('id')
 628                 if not v_id:
 629                     continue
 630                 v_id = compat_str(v_id)
 631                 entries.append(self.url_result(
 632                     self._VIDEO_PAGE_TEMPLATE % v_id,
 633                     self.ie_key(), v_id, video.get('name')))
 634
 635             return self.playlist_result(entries, video_id)
 636
 637         if not video_data:
 638             # Video info not in first request, do a secondary request using
 639             # tahoe player specific URL
 640             tahoe_data = self._download_webpage(
 641                 self._VIDEO_PAGE_TAHOE_TEMPLATE % video_id, video_id,
 642                 data=urlencode_postdata({
 643                     '__a': 1,
 644                     '__pc': self._search_regex(
 645                         r'pkg_cohort["\']\s*:\s*["\'](.+?)["\']', webpage,
 646                         'pkg cohort', default='PHASED:DEFAULT'),
 647                     '__rev': self._search_regex(
 648                         r'client_revision["\']\s*:\s*(\d+),', webpage,
 649                         'client revision', default='3944515'),
 650                     'fb_dtsg': self._search_regex(
 651                         r'"DTSGInitialData"\s*,\s*\[\]\s*,\s*{\s*"token"\s*:\s*"([^"]+)"',
 652                         webpage, 'dtsg token', default=''),
 653                 }),
 654                 headers={
 655                     'Content-Type': 'application/x-www-form-urlencoded',
 656                 })
 657             tahoe_js_data = self._parse_json(
 658                 self._search_regex(
 659                     r'for\s+\(\s*;\s*;\s*\)\s*;(.+)', tahoe_data,
 660                     'tahoe js data', default='{}'),
 661                 video_id, fatal=False)
 662             video_data = extract_from_jsmods_instances(tahoe_js_data)
 663
 664         if not video_data:
 665             raise ExtractorError('Cannot parse data')
 666
 667         if len(video_data) > 1:
 668             entries = []
 669             for v in video_data:
 670                 video_url = v[0].get('video_url')
 671                 if not video_url:
 672                     continue
 673                 entries.append(self.url_result(urljoin(
 674                     url, video_url), self.ie_key(), v[0].get('video_id')))
 675             return self.playlist_result(entries, video_id)
 676         video_data = video_data[0]
 677
 678         formats = []
 679         subtitles = {}
 680         for f in video_data:
 681             format_id = f['stream_type']
 682             if f and isinstance(f, dict):
 683                 f = [f]
 684             if not f or not isinstance(f, list):
 685                 continue
 686             for quality in ('sd', 'hd'):
 687                 for src_type in ('src', 'src_no_ratelimit'):
 688                     src = f[0].get('%s_%s' % (quality, src_type))
 689                     if src:
 690                         preference = -10 if format_id == 'progressive' else 0
 691                         if quality == 'hd':
 692                             preference += 5
 693                         formats.append({
 694                             'format_id': '%s_%s_%s' % (format_id, quality, src_type),
 695                             'url': src,
 696                             'quality': preference,
 697                         })
 698             extract_dash_manifest(f[0], formats)
 699             subtitles_src = f[0].get('subtitles_src')
 700             if subtitles_src:
 701                 subtitles.setdefault('en', []).append({'url': subtitles_src})
 702
 703         process_formats(formats)
 704
 705         info_dict = {
 706             'id': video_id,
 707             'formats': formats,
 708             'subtitles': subtitles,
 709         }
 710         info_dict.update(extract_metadata(webpage))
 711
 712         return info_dict
 713
 714     def _real_extract(self, url):
 715         video_id = self._match_id(url)
 716
 717         real_url = self._VIDEO_PAGE_TEMPLATE % video_id if url.startswith('facebook:') else url
 718         return self._extract_from_url(real_url, video_id)
 719
 720
 721 class FacebookPluginsVideoIE(InfoExtractor):
 722     _VALID_URL = r'https?://(?:[\w-]+\.)?facebook\.com/plugins/video\.php\?.*?\bhref=(?P<id>https.+)'
 723
 724     _TESTS = [{
 725         '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',
 726         'md5': '5954e92cdfe51fe5782ae9bda7058a07',
 727         'info_dict': {
 728             'id': '10154383743583686',
 729             'ext': 'mp4',
 730             'title': 'What to do during the haze?',
 731             'uploader': 'Gov.sg',
 732             'upload_date': '20160826',
 733             'timestamp': 1472184808,
 734         },
 735         'add_ie': [FacebookIE.ie_key()],
 736     }, {
 737         'url': 'https://www.facebook.com/plugins/video.php?href=https%3A%2F%2Fwww.facebook.com%2Fvideo.php%3Fv%3D10204634152394104',
 738         'only_matching': True,
 739     }, {
 740         'url': 'https://www.facebook.com/plugins/video.php?href=https://www.facebook.com/gov.sg/videos/10154383743583686/&show_text=0&width=560',
 741         'only_matching': True,
 742     }]
 743
 744     def _real_extract(self, url):
 745         return self.url_result(
 746             compat_urllib_parse_unquote(self._match_id(url)),
 747             FacebookIE.ie_key())