yt_dlp/extractor/brightcove.py

   1 import base64
   2 import re
   3 import struct
   4 import urllib.parse
   5 import xml.etree.ElementTree
   6
   7 from .adobepass import AdobePassIE
   8 from .common import InfoExtractor
   9 from ..compat import compat_etree_fromstring
  10 from ..networking.exceptions import HTTPError
  11 from ..utils import (
  12     ExtractorError,
  13     UnsupportedError,
  14     clean_html,
  15     dict_get,
  16     extract_attributes,
  17     find_xpath_attr,
  18     fix_xml_ampersands,
  19     float_or_none,
  20     int_or_none,
  21     join_nonempty,
  22     js_to_json,
  23     mimetype2ext,
  24     parse_iso8601,
  25     parse_qs,
  26     smuggle_url,
  27     str_or_none,
  28     try_get,
  29     unescapeHTML,
  30     unsmuggle_url,
  31     update_url_query,
  32     url_or_none,
  33 )
  34 from ..utils.traversal import traverse_obj
  35
  36
  37 class BrightcoveLegacyIE(InfoExtractor):
  38     IE_NAME = 'brightcove:legacy'
  39     _VALID_URL = r'(?:https?://.*brightcove\.com/(services|viewer).*?\?|brightcove:)(?P<query>.*)'
  40
  41     _TESTS = [
  42         {
  43             # From http://www.8tv.cat/8aldia/videos/xavier-sala-i-martin-aquesta-tarda-a-8-al-dia/
  44             'url': 'http://c.brightcove.com/services/viewer/htmlFederated?playerID=1654948606001&flashID=myExperience&%40videoPlayer=2371591881001',
  45             'md5': '5423e113865d26e40624dce2e4b45d95',
  46             'note': 'Test Brightcove downloads and detection in GenericIE',
  47             'info_dict': {
  48                 'id': '2371591881001',
  49                 'ext': 'mp4',
  50                 'title': 'Xavier Sala i Martín: “Un banc que no presta és un banc zombi que no serveix per a res”',
  51                 'uploader': '8TV',
  52                 'description': 'md5:a950cc4285c43e44d763d036710cd9cd',
  53                 'timestamp': 1368213670,
  54                 'upload_date': '20130510',
  55                 'uploader_id': '1589608506001',
  56             },
  57             'skip': 'The player has been deactivated by the content owner',
  58         },
  59         {
  60             # From http://medianetwork.oracle.com/video/player/1785452137001
  61             'url': 'http://c.brightcove.com/services/viewer/htmlFederated?playerID=1217746023001&flashID=myPlayer&%40videoPlayer=1785452137001',
  62             'info_dict': {
  63                 'id': '1785452137001',
  64                 'ext': 'flv',
  65                 'title': 'JVMLS 2012: Arrays 2.0 - Opportunities and Challenges',
  66                 'description': 'John Rose speaks at the JVM Language Summit, August 1, 2012.',
  67                 'uploader': 'Oracle',
  68                 'timestamp': 1344975024,
  69                 'upload_date': '20120814',
  70                 'uploader_id': '1460825906',
  71             },
  72             'skip': 'video not playable',
  73         },
  74         {
  75             # From http://mashable.com/2013/10/26/thermoelectric-bracelet-lets-you-control-your-body-temperature/
  76             'url': 'http://c.brightcove.com/services/viewer/federated_f9?&playerID=1265504713001&publisherID=AQ%7E%7E%2CAAABBzUwv1E%7E%2CxP-xFHVUstiMFlNYfvF4G9yFnNaqCw_9&videoID=2750934548001',
  77             'info_dict': {
  78                 'id': '2750934548001',
  79                 'ext': 'mp4',
  80                 'title': 'This Bracelet Acts as a Personal Thermostat',
  81                 'description': 'md5:547b78c64f4112766ccf4e151c20b6a0',
  82                 # 'uploader': 'Mashable',
  83                 'timestamp': 1382041798,
  84                 'upload_date': '20131017',
  85                 'uploader_id': '1130468786001',
  86             },
  87         },
  88         {
  89             # test that the default referer works
  90             # from http://national.ballet.ca/interact/video/Lost_in_Motion_II/
  91             'url': 'http://link.brightcove.com/services/player/bcpid756015033001?bckey=AQ~~,AAAApYJi_Ck~,GxhXCegT1Dp39ilhXuxMJxasUhVNZiil&bctid=2878862109001',
  92             'info_dict': {
  93                 'id': '2878862109001',
  94                 'ext': 'mp4',
  95                 'title': 'Lost in Motion II',
  96                 'description': 'md5:363109c02998fee92ec02211bd8000df',
  97                 'uploader': 'National Ballet of Canada',
  98             },
  99             'skip': 'Video gone',
 100         },
 101         {
 102             # test flv videos served by akamaihd.net
 103             # From http://www.redbull.com/en/bike/stories/1331655643987/replay-uci-dh-world-cup-2014-from-fort-william
 104             'url': 'http://c.brightcove.com/services/viewer/htmlFederated?%40videoPlayer=ref%3Aevent-stream-356&linkBaseURL=http%3A%2F%2Fwww.redbull.com%2Fen%2Fbike%2Fvideos%2F1331655630249%2Freplay-uci-fort-william-2014-dh&playerKey=AQ%7E%7E%2CAAAApYJ7UqE%7E%2Cxqr_zXk0I-zzNndy8NlHogrCb5QdyZRf&playerID=1398061561001#__youtubedl_smuggle=%7B%22Referer%22%3A+%22http%3A%2F%2Fwww.redbull.com%2Fen%2Fbike%2Fstories%2F1331655643987%2Freplay-uci-dh-world-cup-2014-from-fort-william%22%7D',
 105             # The md5 checksum changes on each download
 106             'info_dict': {
 107                 'id': '3750436379001',
 108                 'ext': 'flv',
 109                 'title': 'UCI MTB World Cup 2014: Fort William, UK - Downhill Finals',
 110                 'uploader': 'RBTV Old (do not use)',
 111                 'description': 'UCI MTB World Cup 2014: Fort William, UK - Downhill Finals',
 112                 'timestamp': 1409122195,
 113                 'upload_date': '20140827',
 114                 'uploader_id': '710858724001',
 115             },
 116             'skip': 'Video gone',
 117         },
 118         {
 119             # playlist with 'videoList'
 120             # from http://support.brightcove.com/en/video-cloud/docs/playlist-support-single-video-players
 121             'url': 'http://c.brightcove.com/services/viewer/htmlFederated?playerID=3550052898001&playerKey=AQ%7E%7E%2CAAABmA9XpXk%7E%2C-Kp7jNgisre1fG5OdqpAFUTcs0lP_ZoL',
 122             'info_dict': {
 123                 'title': 'Sealife',
 124                 'id': '3550319591001',
 125             },
 126             'playlist_mincount': 7,
 127             'skip': 'Unsupported URL',
 128         },
 129         {
 130             # playlist with 'playlistTab' (https://github.com/ytdl-org/youtube-dl/issues/9965)
 131             'url': 'http://c.brightcove.com/services/json/experience/runtime/?command=get_programming_for_experience&playerKey=AQ%7E%7E,AAABXlLMdok%7E,NJ4EoMlZ4rZdx9eU1rkMVd8EaYPBBUlg',
 132             'info_dict': {
 133                 'id': '1522758701001',
 134                 'title': 'Lesson 08',
 135             },
 136             'playlist_mincount': 10,
 137             'skip': 'Unsupported URL',
 138         },
 139         {
 140             # playerID inferred from bcpid
 141             # from http://www.un.org/chinese/News/story.asp?NewsID=27724
 142             'url': 'https://link.brightcove.com/services/player/bcpid1722935254001/?bctid=5360463607001&autoStart=false&secureConnections=true&width=650&height=350',
 143             'only_matching': True,  # Tested in GenericIE
 144         },
 145     ]
 146
 147     _WEBPAGE_TESTS = [{
 148         # embedded brightcove video
 149         # it also tests brightcove videos that need to set the 'Referer'
 150         # in the http requests
 151         'url': 'http://www.bfmtv.com/video/bfmbusiness/cours-bourse/cours-bourse-l-analyse-technique-154522/',
 152         'info_dict': {
 153             'id': '2765128793001',
 154             'ext': 'mp4',
 155             'title': 'Le cours de bourse : l’analyse technique',
 156             'description': 'md5:7e9ad046e968cb2d1114004aba466fd9',
 157             'uploader': 'BFM BUSINESS',
 158         },
 159         'params': {
 160             'skip_download': True,
 161         },
 162         'skip': '404 Not Found',
 163     }, {
 164         # embedded with itemprop embedURL and video id spelled as `idVideo`
 165         'url': 'http://bfmbusiness.bfmtv.com/mediaplayer/chroniques/olivier-delamarche/',
 166         'info_dict': {
 167             'id': '5255628253001',
 168             'ext': 'mp4',
 169             'title': 'md5:37c519b1128915607601e75a87995fc0',
 170             'description': 'md5:37f7f888b434bb8f8cc8dbd4f7a4cf26',
 171             'uploader': 'BFM BUSINESS',
 172             'uploader_id': '876450612001',
 173             'timestamp': 1482255315,
 174             'upload_date': '20161220',
 175         },
 176         'params': {
 177             'skip_download': True,
 178         },
 179         'skip': 'Redirects, page gone',
 180     }, {
 181         # https://github.com/ytdl-org/youtube-dl/issues/2253
 182         'url': 'http://bcove.me/i6nfkrc3',
 183         'md5': '0ba9446db037002366bab3b3eb30c88c',
 184         'info_dict': {
 185             'id': '3101154703001',
 186             'ext': 'mp4',
 187             'title': 'Still no power',
 188             'uploader': 'thestar.com',
 189             'description': 'Mississauga resident David Farmer is still out of power as a result of the ice storm a month ago. To keep the house warm, Farmer cuts wood from his property for a wood burning stove downstairs.',
 190         },
 191         'skip': 'video gone',
 192     }, {
 193         # https://github.com/ytdl-org/youtube-dl/issues/3541
 194         'url': 'http://www.kijk.nl/sbs6/leermijvrouwenkennen/videos/jqMiXKAYan2S/aflevering-1',
 195         'info_dict': {
 196             'id': '3866516442001',
 197             'ext': 'mp4',
 198             'title': 'Leer mij vrouwen kennen: Aflevering 1',
 199             'description': 'Leer mij vrouwen kennen: Aflevering 1',
 200             'uploader': 'SBS Broadcasting',
 201         },
 202         'skip': 'Restricted to Netherlands, 404 Not Found',
 203         'params': {
 204             'skip_download': True,  # m3u8 download
 205         },
 206     }, {
 207         # Brightcove video in <iframe>
 208         'url': 'http://www.un.org/chinese/News/story.asp?NewsID=27724',
 209         'md5': '36d74ef5e37c8b4a2ce92880d208b968',
 210         'info_dict': {
 211             'id': '5360463607001',
 212             'ext': 'mp4',
 213             'title': '叙利亚失明儿童在废墟上演唱《心跳》  呼吁获得正常童年生活',
 214             'description': '联合国儿童基金会中东和北非区域大使、作曲家扎德·迪拉尼（Zade Dirani）在3月15日叙利亚冲突爆发7周年纪念日之际发布了为叙利亚谱写的歌曲《心跳》（HEARTBEAT），为受到六年冲突影响的叙利亚儿童发出强烈呐喊，呼吁世界做出共同努力，使叙利亚儿童重新获得享有正常童年生活的权利。',
 215             'uploader': 'United Nations',
 216             'uploader_id': '1362235914001',
 217             'timestamp': 1489593889,
 218             'upload_date': '20170315',
 219         },
 220         'skip': '404 Not Found',
 221     }, {
 222         # Brightcove with UUID in videoPlayer
 223         'url': 'http://www8.hp.com/cn/zh/home.html',
 224         'info_dict': {
 225             'id': '5255815316001',
 226             'ext': 'mp4',
 227             'title': 'Sprocket Video - China',
 228             'description': 'Sprocket Video - China',
 229             'uploader': 'HP-Video Gallery',
 230             'timestamp': 1482263210,
 231             'upload_date': '20161220',
 232             'uploader_id': '1107601872001',
 233         },
 234         'params': {
 235             'skip_download': True,  # m3u8 download
 236         },
 237         'skip': 'video rotates...weekly?',
 238     }, {
 239         # Multiple brightcove videos
 240         # https://github.com/ytdl-org/youtube-dl/issues/2283
 241         'url': 'http://www.newyorker.com/online/blogs/newsdesk/2014/01/always-never-nuclear-command-and-control.html',
 242         'info_dict': {
 243             'id': 'always-never',
 244             'title': 'Always / Never - The New Yorker',
 245         },
 246         'playlist_count': 3,
 247         'params': {
 248             'extract_flat': False,
 249             'skip_download': True,
 250         },
 251         'skip': 'Redirects, page gone',
 252     }, {
 253         # BrightcoveInPageEmbed embed
 254         'url': 'http://www.geekandsundry.com/tabletop-bonus-wils-final-thoughts-on-dread/',
 255         'info_dict': {
 256             'id': '4238694884001',
 257             'ext': 'flv',
 258             'title': 'Tabletop: Dread, Last Thoughts',
 259             'description': 'Tabletop: Dread, Last Thoughts',
 260             'duration': 51690,
 261         },
 262         'skip': 'Redirects, page gone',
 263     }, {
 264         # Brightcove embed, with no valid 'renditions' but valid 'IOSRenditions'
 265         # This video can't be played in browsers if Flash disabled and UA set to iPhone, which is actually a false alarm
 266         'url': 'https://dl.dropboxusercontent.com/u/29092637/interview.html',
 267         'info_dict': {
 268             'id': '4785848093001',
 269             'ext': 'mp4',
 270             'title': 'The Cardinal Pell Interview',
 271             'description': 'Sky News Contributor Andrew Bolt interviews George Pell in Rome, following the Cardinal\'s evidence before the Royal Commission into Child Abuse. ',
 272             'uploader': 'GlobeCast Australia - GlobeStream',
 273             'uploader_id': '2733773828001',
 274             'upload_date': '20160304',
 275             'timestamp': 1457083087,
 276         },
 277         'params': {
 278             # m3u8 downloads
 279             'skip_download': True,
 280         },
 281         'skip': '404 Not Found',
 282     }, {
 283         # Brightcove embed with whitespace around attribute names
 284         'url': 'http://www.stack.com/video/3167554373001/learn-to-hit-open-three-pointers-with-damian-lillard-s-baseline-drift-drill',
 285         'info_dict': {
 286             'id': '3167554373001',
 287             'ext': 'mp4',
 288             'title': "Learn to Hit Open Three-Pointers With Damian Lillard's Baseline Drift Drill",
 289             'description': 'md5:57bacb0e0f29349de4972bfda3191713',
 290             'uploader_id': '1079349493',
 291             'upload_date': '20140207',
 292             'timestamp': 1391810548,
 293         },
 294         'params': {
 295             'skip_download': True,
 296         },
 297         'skip': '410 Gone',
 298     }]
 299
 300     @classmethod
 301     def _build_brightcove_url(cls, object_str):
 302         """
 303         Build a Brightcove url from a xml string containing
 304         <object class="BrightcoveExperience">{params}</object>
 305         """
 306
 307         # Fix up some stupid HTML, see https://github.com/ytdl-org/youtube-dl/issues/1553
 308         object_str = re.sub(r'(<param(?:\s+[a-zA-Z0-9_]+="[^"]*")*)>',
 309                             lambda m: m.group(1) + '/>', object_str)
 310         # Fix up some stupid XML, see https://github.com/ytdl-org/youtube-dl/issues/1608
 311         object_str = object_str.replace('<--', '<!--')
 312         # remove namespace to simplify extraction
 313         object_str = re.sub(r'(<object[^>]*)(xmlns=".*?")', r'\1', object_str)
 314         object_str = fix_xml_ampersands(object_str)
 315
 316         try:
 317             object_doc = compat_etree_fromstring(object_str.encode())
 318         except xml.etree.ElementTree.ParseError:
 319             return
 320
 321         fv_el = find_xpath_attr(object_doc, './param', 'name', 'flashVars')
 322         if fv_el is not None:
 323             flashvars = dict(
 324                 (k, v[0])
 325                 for k, v in urllib.parse.parse_qs(fv_el.attrib['value']).items())
 326         else:
 327             flashvars = {}
 328
 329         data_url = object_doc.attrib.get('data', '')
 330         data_url_params = parse_qs(data_url)
 331
 332         def find_param(name):
 333             if name in flashvars:
 334                 return flashvars[name]
 335             node = find_xpath_attr(object_doc, './param', 'name', name)
 336             if node is not None:
 337                 return node.attrib['value']
 338             return data_url_params.get(name)
 339
 340         params = {}
 341
 342         player_id = find_param('playerID') or find_param('playerId')
 343         if player_id is None:
 344             raise ExtractorError('Cannot find player ID')
 345         params['playerID'] = player_id
 346
 347         player_key = find_param('playerKey')
 348         # Not all pages define this value
 349         if player_key is not None:
 350             params['playerKey'] = player_key
 351         # These fields hold the id of the video
 352         video_player = find_param('@videoPlayer') or find_param('videoId') or find_param('videoID') or find_param('@videoList')
 353         if video_player is not None:
 354             if isinstance(video_player, list):
 355                 video_player = video_player[0]
 356             video_player = video_player.strip()
 357             # UUID is also possible for videoPlayer (e.g.
 358             # http://www.popcornflix.com/hoodies-vs-hooligans/7f2d2b87-bbf2-4623-acfb-ea942b4f01dd
 359             # or http://www8.hp.com/cn/zh/home.html)
 360             if not (re.match(
 361                     r'^(?:\d+|[\da-fA-F]{8}-?[\da-fA-F]{4}-?[\da-fA-F]{4}-?[\da-fA-F]{4}-?[\da-fA-F]{12})$',
 362                     video_player) or video_player.startswith('ref:')):
 363                 return None
 364             params['@videoPlayer'] = video_player
 365         link_base = find_param('linkBaseURL')
 366         if link_base is not None:
 367             params['linkBaseURL'] = link_base
 368         return cls._make_brightcove_url(params)
 369
 370     @classmethod
 371     def _build_brightcove_url_from_js(cls, object_js):
 372         # The layout of JS is as follows:
 373         # customBC.createVideo = function (width, height, playerID, playerKey, videoPlayer, VideoRandomID) {
 374         #   // build Brightcove <object /> XML
 375         # }
 376         m = re.search(
 377             r'''(?x)customBC\.createVideo\(
 378                 .*?                                                  # skipping width and height
 379                 ["\'](?P<playerID>\d+)["\']\s*,\s*                   # playerID
 380                 ["\'](?P<playerKey>AQ[^"\']{48})[^"\']*["\']\s*,\s*  # playerKey begins with AQ and is 50 characters
 381                                                                      # in length, however it's appended to itself
 382                                                                      # in places, so truncate
 383                 ["\'](?P<videoID>\d+)["\']                           # @videoPlayer
 384             ''', object_js)
 385         if m:
 386             return cls._make_brightcove_url(m.groupdict())
 387
 388     @classmethod
 389     def _make_brightcove_url(cls, params):
 390         return update_url_query(
 391             'https://c.brightcove.com/services/viewer/htmlFederated', params)
 392
 393     @classmethod
 394     def _extract_brightcove_url(cls, webpage):
 395         """Try to extract the brightcove url from the webpage, returns None
 396         if it can't be found
 397         """
 398         urls = cls._extract_brightcove_urls(webpage)
 399         return urls[0] if urls else None
 400
 401     @classmethod
 402     def _extract_brightcove_urls(cls, webpage):
 403         """Return a list of all Brightcove URLs from the webpage """
 404
 405         url_m = re.search(
 406             r'''(?x)
 407                 <meta\s+
 408                     (?:property|itemprop)=([\'"])(?:og:video|embedURL)\1[^>]+
 409                     content=([\'"])(?P<url>https?://(?:secure|c)\.brightcove.com/(?:(?!\2).)+)\2
 410             ''', webpage)
 411         if url_m:
 412             url = unescapeHTML(url_m.group('url'))
 413             # Some sites don't add it, we can't download with this url, for example:
 414             # http://www.ktvu.com/videos/news/raw-video-caltrain-releases-video-of-man-almost/vCTZdY/
 415             if 'playerKey' in url or 'videoId' in url or 'idVideo' in url:
 416                 return [url]
 417
 418         matches = re.findall(
 419             r'''(?sx)<object
 420             (?:
 421                 [^>]+?class=[\'"][^>]*?BrightcoveExperience.*?[\'"] |
 422                 [^>]*?>\s*<param\s+name="movie"\s+value="https?://[^/]*brightcove\.com/
 423             ).+?>\s*</object>''',
 424             webpage)
 425         if matches:
 426             return list(filter(None, [cls._build_brightcove_url(m) for m in matches]))
 427
 428         matches = re.findall(r'(customBC\.createVideo\(.+?\);)', webpage)
 429         if matches:
 430             return list(filter(None, [
 431                 cls._build_brightcove_url_from_js(custom_bc)
 432                 for custom_bc in matches]))
 433         return [src for _, src in re.findall(
 434             r'<iframe[^>]+src=([\'"])((?:https?:)?//link\.brightcove\.com/services/player/(?!\1).+)\1', webpage)]
 435
 436     def _extract_from_webpage(self, url, webpage):
 437         bc_urls = self._extract_brightcove_urls(webpage)
 438         for bc_url in bc_urls:
 439             yield self.url_result(smuggle_url(bc_url, {'Referer': url}), BrightcoveLegacyIE)
 440
 441     def _real_extract(self, url):
 442         url, smuggled_data = unsmuggle_url(url, {})
 443
 444         # Change the 'videoId' and others field to '@videoPlayer'
 445         url = re.sub(r'(?<=[?&])(videoI(d|D)|idVideo|bctid)', '%40videoPlayer', url)
 446         # Change bckey (used by bcove.me urls) to playerKey
 447         url = re.sub(r'(?<=[?&])bckey', 'playerKey', url)
 448         mobj = self._match_valid_url(url)
 449         query_str = mobj.group('query')
 450         query = urllib.parse.parse_qs(query_str)
 451
 452         video_player = query.get('@videoPlayer')
 453         if video_player:
 454             # We set the original url as the default 'Referer' header
 455             referer = query.get('linkBaseURL', [None])[0] or smuggled_data.get('Referer', url)
 456             video_id = video_player[0]
 457             if 'playerID' not in query:
 458                 mobj = re.search(r'/bcpid(\d+)', url)
 459                 if mobj is not None:
 460                     query['playerID'] = [mobj.group(1)]
 461             publisher_id = query.get('publisherId')
 462             if publisher_id and publisher_id[0].isdigit():
 463                 publisher_id = publisher_id[0]
 464             if not publisher_id:
 465                 player_key = query.get('playerKey')
 466                 if player_key and ',' in player_key[0]:
 467                     player_key = player_key[0]
 468                 else:
 469                     player_id = query.get('playerID')
 470                     if player_id and player_id[0].isdigit():
 471                         headers = {}
 472                         if referer:
 473                             headers['Referer'] = referer
 474                         player_page = self._download_webpage(
 475                             'https://link.brightcove.com/services/player/bcpid' + player_id[0],
 476                             video_id, headers=headers, fatal=False)
 477                         if player_page:
 478                             player_key = self._search_regex(
 479                                 r'<param\s+name="playerKey"\s+value="([\w~,-]+)"',
 480                                 player_page, 'player key', fatal=False)
 481                 if player_key:
 482                     enc_pub_id = player_key.split(',')[1].replace('~', '=')
 483                     publisher_id = struct.unpack('>Q', base64.urlsafe_b64decode(enc_pub_id))[0]
 484             if publisher_id:
 485                 brightcove_new_url = f'https://players.brightcove.net/{publisher_id}/default_default/index.html?videoId={video_id}'
 486                 if referer:
 487                     brightcove_new_url = smuggle_url(brightcove_new_url, {'referrer': referer})
 488                 return self.url_result(brightcove_new_url, BrightcoveNewIE.ie_key(), video_id)
 489         # TODO: figure out if it's possible to extract playlistId from playerKey
 490         # elif 'playerKey' in query:
 491         #     player_key = query['playerKey']
 492         #     return self._get_playlist_info(player_key[0])
 493         raise UnsupportedError(url)
 494
 495
 496 class BrightcoveNewBaseIE(AdobePassIE):
 497     def _parse_brightcove_metadata(self, json_data, video_id, headers={}):
 498         title = json_data['name'].strip()
 499
 500         formats, subtitles = [], {}
 501         sources = json_data.get('sources') or []
 502         for source in sources:
 503             container = source.get('container')
 504             ext = mimetype2ext(source.get('type'))
 505             src = source.get('src')
 506             if ext == 'm3u8' or container == 'M2TS':
 507                 if not src:
 508                     continue
 509                 fmts, subs = self._extract_m3u8_formats_and_subtitles(
 510                     src, video_id, 'mp4', 'm3u8_native', m3u8_id='hls', fatal=False)
 511                 subtitles = self._merge_subtitles(subtitles, subs)
 512             elif ext == 'mpd':
 513                 if not src:
 514                     continue
 515                 fmts, subs = self._extract_mpd_formats_and_subtitles(src, video_id, 'dash', fatal=False)
 516                 subtitles = self._merge_subtitles(subtitles, subs)
 517             else:
 518                 streaming_src = source.get('streaming_src')
 519                 stream_name, app_name = source.get('stream_name'), source.get('app_name')
 520                 if not src and not streaming_src and (not stream_name or not app_name):
 521                     continue
 522                 tbr = float_or_none(source.get('avg_bitrate'), 1000)
 523                 height = int_or_none(source.get('height'))
 524                 width = int_or_none(source.get('width'))
 525                 f = {
 526                     'tbr': tbr,
 527                     'filesize': int_or_none(source.get('size')),
 528                     'container': container,
 529                     'ext': ext or container.lower(),
 530                 }
 531                 if width == 0 and height == 0:
 532                     f.update({
 533                         'vcodec': 'none',
 534                     })
 535                 else:
 536                     f.update({
 537                         'width': width,
 538                         'height': height,
 539                         'vcodec': source.get('codec'),
 540                     })
 541
 542                 def build_format_id(kind):
 543                     return join_nonempty(kind, tbr and f'{int(tbr)}k', height and f'{height}p')
 544
 545                 if src or streaming_src:
 546                     f.update({
 547                         'url': src or streaming_src,
 548                         'format_id': build_format_id('http' if src else 'http-streaming'),
 549                         'source_preference': 0 if src else -1,
 550                     })
 551                 else:
 552                     f.update({
 553                         'url': app_name,
 554                         'play_path': stream_name,
 555                         'format_id': build_format_id('rtmp'),
 556                     })
 557                 fmts = [f]
 558
 559             # https://support.brightcove.com/playback-api-video-fields-reference#key_systems_object
 560             if container == 'WVM' or source.get('key_systems') or ext == 'ism':
 561                 for f in fmts:
 562                     f['has_drm'] = True
 563             formats.extend(fmts)
 564
 565         if not formats:
 566             errors = json_data.get('errors')
 567             if errors:
 568                 error = errors[0]
 569                 self.raise_no_formats(
 570                     error.get('message') or error.get('error_subcode') or error['error_code'], expected=True)
 571
 572         headers.pop('Authorization', None)  # or else http formats will give error 400
 573         for f in formats:
 574             f.setdefault('http_headers', {}).update(headers)
 575
 576         for text_track in json_data.get('text_tracks', []):
 577             if text_track.get('kind') != 'captions':
 578                 continue
 579             text_track_url = url_or_none(text_track.get('src'))
 580             if not text_track_url:
 581                 continue
 582             lang = (str_or_none(text_track.get('srclang'))
 583                     or str_or_none(text_track.get('label')) or 'en').lower()
 584             subtitles.setdefault(lang, []).append({
 585                 'url': text_track_url,
 586             })
 587
 588         is_live = False
 589         duration = float_or_none(json_data.get('duration'), 1000)
 590         if duration is not None and duration <= 0:
 591             is_live = True
 592
 593         common_res = [(160, 90), (320, 180), (480, 720), (640, 360), (768, 432), (1024, 576), (1280, 720), (1366, 768), (1920, 1080)]
 594         thumb_base_url = dict_get(json_data, ('poster', 'thumbnail'))
 595         thumbnails = [{
 596             'url': re.sub(r'\d+x\d+', f'{w}x{h}', thumb_base_url),
 597             'width': w,
 598             'height': h,
 599         } for w, h in common_res] if thumb_base_url else None
 600
 601         return {
 602             'id': video_id,
 603             'title': title,
 604             'description': clean_html(json_data.get('description')),
 605             'thumbnails': thumbnails,
 606             'duration': duration,
 607             'timestamp': parse_iso8601(json_data.get('published_at')),
 608             'uploader_id': json_data.get('account_id'),
 609             'formats': formats,
 610             'subtitles': subtitles,
 611             'tags': json_data.get('tags', []),
 612             'is_live': is_live,
 613         }
 614
 615
 616 class BrightcoveNewIE(BrightcoveNewBaseIE):
 617     IE_NAME = 'brightcove:new'
 618     _VALID_URL = r'https?://players\.brightcove\.net/(?P<account_id>\d+)/(?P<player_id>[^/]+)_(?P<embed>[^/]+)/index\.html\?.*(?P<content_type>video|playlist)Id=(?P<video_id>\d+|ref:[^&]+)'
 619     _TESTS = [{
 620         'url': 'http://players.brightcove.net/929656772001/e41d32dc-ec74-459e-a845-6c69f7b724ea_default/index.html?videoId=4463358922001',
 621         'md5': 'c8100925723840d4b0d243f7025703be',
 622         'info_dict': {
 623             'id': '4463358922001',
 624             'ext': 'mp4',
 625             'title': 'Meet the man behind Popcorn Time',
 626             'description': 'md5:eac376a4fe366edc70279bfb681aea16',
 627             'duration': 165.768,
 628             'timestamp': 1441391203,
 629             'upload_date': '20150904',
 630             'uploader_id': '929656772001',
 631             'formats': 'mincount:20',
 632         },
 633         'skip': '404 Not Found',
 634     }, {
 635         # with rtmp streams
 636         'url': 'http://players.brightcove.net/4036320279001/5d112ed9-283f-485f-a7f9-33f42e8bc042_default/index.html?videoId=4279049078001',
 637         'info_dict': {
 638             'id': '4279049078001',
 639             'ext': 'mp4',
 640             'title': 'Titansgrave: Chapter 0',
 641             'description': 'Titansgrave: Chapter 0',
 642             'duration': 1242.058,
 643             'timestamp': 1433556729,
 644             'upload_date': '20150606',
 645             'uploader_id': '4036320279001',
 646             'formats': 'mincount:39',
 647         },
 648         'params': {
 649             # m3u8 download
 650             'skip_download': True,
 651         },
 652     }, {
 653         # playlist stream
 654         'url': 'https://players.brightcove.net/1752604059001/S13cJdUBz_default/index.html?playlistId=5718313430001',
 655         'info_dict': {
 656             'id': '5718313430001',
 657             'title': 'No Audio Playlist',
 658         },
 659         'playlist_count': 7,
 660         'params': {
 661             # m3u8 download
 662             'skip_download': True,
 663         },
 664     }, {
 665         'url': 'http://players.brightcove.net/5690807595001/HyZNerRl7_default/index.html?playlistId=5743160747001',
 666         'only_matching': True,
 667     }, {
 668         # ref: prefixed video id
 669         'url': 'http://players.brightcove.net/3910869709001/21519b5c-4b3b-4363-accb-bdc8f358f823_default/index.html?videoId=ref:7069442',
 670         'only_matching': True,
 671     }, {
 672         # non numeric ref: prefixed video id
 673         'url': 'http://players.brightcove.net/710858724001/default_default/index.html?videoId=ref:event-stream-356',
 674         'only_matching': True,
 675     }, {
 676         # unavailable video without message but with error_code
 677         'url': 'http://players.brightcove.net/1305187701/c832abfb-641b-44eb-9da0-2fe76786505f_default/index.html?videoId=4377407326001',
 678         'only_matching': True,
 679     }]
 680
 681     _WEBPAGE_TESTS = [{
 682         # brightcove player url embed
 683         'url': 'https://nbc-2.com/weather/forecast/2022/11/16/forecast-warmest-day-of-the-week/',
 684         'md5': '2934d5372b354d27083ccf8575dbfee2',
 685         'info_dict': {
 686             'id': '6315650313112',
 687             'title': 'First Alert Forecast: November 15, 2022',
 688             'ext': 'mp4',
 689             'tags': ['nbc2', 'forecast'],
 690             'uploader_id': '6146886170001',
 691             'thumbnail': r're:^https?://.*\.jpg$',
 692             'timestamp': 1668574571,
 693             'duration': 233.375,
 694             'upload_date': '20221116',
 695         },
 696     }, {
 697         # embedded with video tag only
 698         'url': 'https://www.gooddishtv.com/tiktok-rapping-chef-mr-pyrex',
 699         'info_dict': {
 700             'id': 'tiktok-rapping-chef-mr-pyrex',
 701             'title': 'TikTok\'s Rapping Chef Makes Jambalaya for the Hosts',
 702             'thumbnail': r're:^https?://.*\.jpg$',
 703             'age_limit': 0,
 704             'description': 'Just in time for Mardi Gras',
 705         },
 706         'playlist': [{
 707             'info_dict': {
 708                 'id': '6299189544001',
 709                 'ext': 'mp4',
 710                 'title': 'TGD_01-032_5',
 711                 'thumbnail': r're:^https?://.*\.jpg$',
 712                 'tags': [],
 713                 'timestamp': 1646078943,
 714                 'uploader_id': '1569565978001',
 715                 'upload_date': '20220228',
 716                 'duration': 217.195,
 717             },
 718         }, {
 719             'info_dict': {
 720                 'id': '6305565995112',
 721                 'ext': 'mp4',
 722                 'title': 'TGD 01-087 (Airs 05.25.22)_Segment 5',
 723                 'thumbnail': r're:^https?://.*\.jpg$',
 724                 'tags': [],
 725                 'timestamp': 1651604591,
 726                 'uploader_id': '1569565978001',
 727                 'upload_date': '20220503',
 728                 'duration': 310.421,
 729             },
 730         }],
 731     }, {
 732         # Brightcove:new type [2].
 733         'url': 'http://www.delawaresportszone.com/video-st-thomas-more-earns-first-trip-to-basketball-semis',
 734         'md5': '2b35148fcf48da41c9fb4591650784f3',
 735         'info_dict': {
 736             'id': '5348741021001',
 737             'ext': 'mp4',
 738             'upload_date': '20170306',
 739             'uploader_id': '4191638492001',
 740             'timestamp': 1488769918,
 741             'title': 'VIDEO:  St. Thomas More earns first trip to basketball semis',
 742         },
 743         'skip': '404 Not Found',
 744     }, {
 745         # Alternative brightcove <video> attributes
 746         'url': 'http://www.programme-tv.net/videos/extraits/81095-guillaume-canet-evoque-les-rumeurs-d-infidelite-de-marion-cotillard-avec-brad-pitt-dans-vivement-dimanche/',
 747         'info_dict': {
 748             'id': '81095-guillaume-canet-evoque-les-rumeurs-d-infidelite-de-marion-cotillard-avec-brad-pitt-dans-vivement-dimanche',
 749             'title': "Guillaume Canet évoque les rumeurs d'infidélité de Marion Cotillard avec Brad Pitt dans Vivement Dimanche, Extraits : toutes les vidéos avec Télé-Loisirs",
 750         },
 751         'playlist': [{
 752             'md5': '732d22ba3d33f2f3fc253c39f8f36523',
 753             'info_dict': {
 754                 'id': '5311302538001',
 755                 'ext': 'mp4',
 756                 'title': "Guillaume Canet évoque les rumeurs d'infidélité de Marion Cotillard avec Brad Pitt dans Vivement Dimanche",
 757                 'description': "Guillaume Canet évoque les rumeurs d'infidélité de Marion Cotillard avec Brad Pitt dans Vivement Dimanche (France 2, 5 février 2017)",
 758                 'timestamp': 1486321708,
 759                 'upload_date': '20170205',
 760                 'uploader_id': '800000640001',
 761             },
 762             'only_matching': True,
 763         }],
 764         'skip': '404 Not Found',
 765     }, {
 766         # Brightcove URL in single quotes
 767         'url': 'http://www.sportsnet.ca/baseball/mlb/sn-presents-russell-martin-world-citizen/',
 768         'md5': '4ae374f1f8b91c889c4b9203c8c752af',
 769         'info_dict': {
 770             'id': '4255764656001',
 771             'ext': 'mp4',
 772             'title': 'SN Presents: Russell Martin, World Citizen',
 773             'description': 'To understand why he was the Toronto Blue Jays’ top off-season priority is to appreciate his background and upbringing in Montreal, where he first developed his baseball skills. Written and narrated by Stephen Brunt.',
 774             'uploader': 'Rogers Sportsnet',
 775             'uploader_id': '1704050871',
 776             'upload_date': '20150525',
 777             'timestamp': 1432570283,
 778         },
 779         'skip': 'Page no longer has URL, now has javascript',
 780     }]
 781
 782     @staticmethod
 783     def _extract_url(ie, webpage):
 784         urls = BrightcoveNewIE._extract_brightcove_urls(ie, webpage)
 785         return urls[0] if urls else None
 786
 787     @staticmethod
 788     def _extract_brightcove_urls(ie, webpage):
 789         # Reference:
 790         # 1. http://docs.brightcove.com/en/video-cloud/brightcove-player/guides/publish-video.html#setvideoiniframe
 791         # 2. http://docs.brightcove.com/en/video-cloud/brightcove-player/guides/publish-video.html#tag
 792         # 3. http://docs.brightcove.com/en/video-cloud/brightcove-player/guides/publish-video.html#setvideousingjavascript
 793         # 4. http://docs.brightcove.com/en/video-cloud/brightcove-player/guides/in-page-embed-player-implementation.html
 794         # 5. https://support.brightcove.com/en/video-cloud/docs/dynamically-assigning-videos-player
 795
 796         entries = []
 797
 798         # Look for iframe embeds [1]
 799         for _, url in re.findall(
 800                 r'<iframe[^>]+src=(["\'])((?:https?:)?//players\.brightcove\.net/\d+/[^/]+/index\.html.+?)\1', webpage):
 801             entries.append(url if url.startswith(('http:', 'https:')) else 'https:' + url)
 802
 803         # Look for <video> tags [2] and embed_in_page embeds [3]
 804         # [2] looks like:
 805         for video, script_tag, account_id, player_id, embed in re.findall(
 806                 r'''(?isx)
 807                     (<video(?:-js)?\s+[^>]*\bdata-video-id\s*=\s*['"]?[^>]+>)
 808                     (?:.*?
 809                         (<script[^>]+
 810                             src=["\'](?:https?:)?//players\.brightcove\.net/
 811                             (\d+)/([^/]+)_([^/]+)/index(?:\.min)?\.js
 812                         )
 813                     )?
 814                 ''', webpage):
 815             attrs = extract_attributes(video)
 816
 817             # According to examples from [4] it's unclear whether video id
 818             # may be optional and what to do when it is
 819             video_id = attrs.get('data-video-id')
 820             if not video_id:
 821                 continue
 822
 823             account_id = account_id or attrs.get('data-account')
 824             if not account_id:
 825                 continue
 826
 827             player_id = player_id or attrs.get('data-player') or 'default'
 828             embed = embed or attrs.get('data-embed') or 'default'
 829
 830             bc_url = f'https://players.brightcove.net/{account_id}/{player_id}_{embed}/index.html?videoId={video_id}'
 831
 832             # Some brightcove videos may be embedded with video tag only and
 833             # without script tag or any mentioning of brightcove at all. Such
 834             # embeds are considered ambiguous since they are matched based only
 835             # on data-video-id and data-account attributes and in the wild may
 836             # not be brightcove embeds at all. Let's check reconstructed
 837             # brightcove URLs in case of such embeds and only process valid
 838             # ones. By this we ensure there is indeed a brightcove embed.
 839             if not script_tag and not ie._is_valid_url(
 840                     bc_url, video_id, 'possible brightcove video'):
 841                 continue
 842
 843             entries.append(bc_url)
 844
 845         return entries
 846
 847     def _extract_from_webpage(self, url, webpage):
 848         bc_urls = self._extract_brightcove_urls(self, webpage)
 849         for bc_url in bc_urls:
 850             yield self.url_result(smuggle_url(bc_url, {'referrer': url}), BrightcoveNewIE)
 851
 852     def _real_extract(self, url):
 853         url, smuggled_data = unsmuggle_url(url, {})
 854         self._initialize_geo_bypass({
 855             'countries': smuggled_data.get('geo_countries'),
 856             'ip_blocks': smuggled_data.get('geo_ip_blocks'),
 857         })
 858
 859         account_id, player_id, embed, content_type, video_id = self._match_valid_url(url).groups()
 860
 861         policy_key_id = f'{account_id}_{player_id}'
 862         policy_key = self.cache.load('brightcove', policy_key_id)
 863         policy_key_extracted = False
 864         store_pk = lambda x: self.cache.store('brightcove', policy_key_id, x)
 865
 866         def extract_policy_key():
 867             base_url = f'https://players.brightcove.net/{account_id}/{player_id}_{embed}/'
 868             config = self._download_json(
 869                 base_url + 'config.json', video_id, fatal=False) or {}
 870             policy_key = try_get(
 871                 config, lambda x: x['video_cloud']['policy_key'])
 872             if not policy_key:
 873                 webpage = self._download_webpage(
 874                     base_url + 'index.min.js', video_id)
 875
 876                 catalog = self._search_regex(
 877                     r'catalog\(({.+?})\);', webpage, 'catalog', default=None)
 878                 if catalog:
 879                     catalog = self._parse_json(
 880                         js_to_json(catalog), video_id, fatal=False)
 881                     if catalog:
 882                         policy_key = catalog.get('policyKey')
 883
 884                 if not policy_key:
 885                     policy_key = self._search_regex(
 886                         r'policyKey\s*:\s*(["\'])(?P<pk>.+?)\1',
 887                         webpage, 'policy key', group='pk')
 888
 889             store_pk(policy_key)
 890             return policy_key
 891
 892         token = smuggled_data.get('token')
 893         api_url = f'https://{"edge-auth" if token else "edge"}.api.brightcove.com/playback/v1/accounts/{account_id}/{content_type}s/{video_id}'
 894         headers = {'Authorization': f'Bearer {token}'} if token else {}
 895         referrer = smuggled_data.get('referrer')  # XXX: notice the spelling/case of the key
 896         if referrer:
 897             headers.update({
 898                 'Referer': referrer,
 899                 'Origin': re.search(r'https?://[^/]+', referrer).group(0),
 900             })
 901
 902         for _ in range(2):
 903             if not policy_key:
 904                 policy_key = extract_policy_key()
 905                 policy_key_extracted = True
 906             headers['Accept'] = f'application/json;pk={policy_key}'
 907             try:
 908                 json_data = self._download_json(api_url, video_id, headers=headers)
 909                 break
 910             except ExtractorError as e:
 911                 if isinstance(e.cause, HTTPError) and e.cause.status in (401, 403):
 912                     json_data = self._parse_json(e.cause.response.read().decode(), video_id)[0]
 913                     message = json_data.get('message') or json_data['error_code']
 914                     if json_data.get('error_subcode') == 'CLIENT_GEO':
 915                         self.raise_geo_restricted(msg=message)
 916                     elif json_data.get('error_code') == 'INVALID_POLICY_KEY' and not policy_key_extracted:
 917                         policy_key = None
 918                         store_pk(None)
 919                         continue
 920                     raise ExtractorError(message, expected=True)
 921                 raise
 922
 923         errors = json_data.get('errors')
 924         if errors and errors[0].get('error_subcode') == 'TVE_AUTH':
 925             custom_fields = json_data['custom_fields']
 926             tve_token = self._extract_mvpd_auth(
 927                 smuggled_data['source_url'], video_id,
 928                 custom_fields['bcadobepassrequestorid'],
 929                 custom_fields['bcadobepassresourceid'])
 930             json_data = self._download_json(
 931                 api_url, video_id, headers={
 932                     'Accept': f'application/json;pk={policy_key}',
 933                 }, query={
 934                     'tveToken': tve_token,
 935                 })
 936
 937         if content_type == 'playlist':
 938             return self.playlist_result(
 939                 (self._parse_brightcove_metadata(vid, vid['id'], headers)
 940                  for vid in traverse_obj(json_data, ('videos', lambda _, v: v['id']))),
 941                 json_data.get('id'), json_data.get('name'),
 942                 json_data.get('description'))
 943
 944         return self._parse_brightcove_metadata(
 945             json_data, video_id, headers=headers)