yt_dlp/extractor/brightcove.py

   1 import base64
   2 import re
   3 import struct
   4 import xml.etree.ElementTree
   5
   6 from .adobepass import AdobePassIE
   7 from .common import InfoExtractor
   8 from ..compat import (
   9     compat_etree_fromstring,
  10     compat_HTTPError,
  11     compat_parse_qs,
  12     compat_urlparse,
  13 )
  14 from ..utils import (
  15     clean_html,
  16     dict_get,
  17     extract_attributes,
  18     ExtractorError,
  19     find_xpath_attr,
  20     fix_xml_ampersands,
  21     float_or_none,
  22     int_or_none,
  23     js_to_json,
  24     mimetype2ext,
  25     parse_iso8601,
  26     parse_qs,
  27     smuggle_url,
  28     str_or_none,
  29     try_get,
  30     unescapeHTML,
  31     unsmuggle_url,
  32     UnsupportedError,
  33     update_url_query,
  34     url_or_none,
  35 )
  36
  37
  38 class BrightcoveLegacyIE(InfoExtractor):
  39     IE_NAME = 'brightcove:legacy'
  40     _VALID_URL = r'(?:https?://.*brightcove\.com/(services|viewer).*?\?|brightcove:)(?P<query>.*)'
  41
  42     _TESTS = [
  43         {
  44             # From http://www.8tv.cat/8aldia/videos/xavier-sala-i-martin-aquesta-tarda-a-8-al-dia/
  45             'url': 'http://c.brightcove.com/services/viewer/htmlFederated?playerID=1654948606001&flashID=myExperience&%40videoPlayer=2371591881001',
  46             'md5': '5423e113865d26e40624dce2e4b45d95',
  47             'note': 'Test Brightcove downloads and detection in GenericIE',
  48             'info_dict': {
  49                 'id': '2371591881001',
  50                 'ext': 'mp4',
  51                 'title': 'Xavier Sala i Martín: “Un banc que no presta és un banc zombi que no serveix per a res”',
  52                 'uploader': '8TV',
  53                 'description': 'md5:a950cc4285c43e44d763d036710cd9cd',
  54                 'timestamp': 1368213670,
  55                 'upload_date': '20130510',
  56                 'uploader_id': '1589608506001',
  57             },
  58             'skip': 'The player has been deactivated by the content owner',
  59         },
  60         {
  61             # From http://medianetwork.oracle.com/video/player/1785452137001
  62             'url': 'http://c.brightcove.com/services/viewer/htmlFederated?playerID=1217746023001&flashID=myPlayer&%40videoPlayer=1785452137001',
  63             'info_dict': {
  64                 'id': '1785452137001',
  65                 'ext': 'flv',
  66                 'title': 'JVMLS 2012: Arrays 2.0 - Opportunities and Challenges',
  67                 'description': 'John Rose speaks at the JVM Language Summit, August 1, 2012.',
  68                 'uploader': 'Oracle',
  69                 'timestamp': 1344975024,
  70                 'upload_date': '20120814',
  71                 'uploader_id': '1460825906',
  72             },
  73             'skip': 'video not playable',
  74         },
  75         {
  76             # From http://mashable.com/2013/10/26/thermoelectric-bracelet-lets-you-control-your-body-temperature/
  77             'url': 'http://c.brightcove.com/services/viewer/federated_f9?&playerID=1265504713001&publisherID=AQ%7E%7E%2CAAABBzUwv1E%7E%2CxP-xFHVUstiMFlNYfvF4G9yFnNaqCw_9&videoID=2750934548001',
  78             'info_dict': {
  79                 'id': '2750934548001',
  80                 'ext': 'mp4',
  81                 'title': 'This Bracelet Acts as a Personal Thermostat',
  82                 'description': 'md5:547b78c64f4112766ccf4e151c20b6a0',
  83                 # 'uploader': 'Mashable',
  84                 'timestamp': 1382041798,
  85                 'upload_date': '20131017',
  86                 'uploader_id': '1130468786001',
  87             },
  88         },
  89         {
  90             # test that the default referer works
  91             # from http://national.ballet.ca/interact/video/Lost_in_Motion_II/
  92             'url': 'http://link.brightcove.com/services/player/bcpid756015033001?bckey=AQ~~,AAAApYJi_Ck~,GxhXCegT1Dp39ilhXuxMJxasUhVNZiil&bctid=2878862109001',
  93             'info_dict': {
  94                 'id': '2878862109001',
  95                 'ext': 'mp4',
  96                 'title': 'Lost in Motion II',
  97                 'description': 'md5:363109c02998fee92ec02211bd8000df',
  98                 'uploader': 'National Ballet of Canada',
  99             },
 100             'skip': 'Video gone',
 101         },
 102         {
 103             # test flv videos served by akamaihd.net
 104             # From http://www.redbull.com/en/bike/stories/1331655643987/replay-uci-dh-world-cup-2014-from-fort-william
 105             '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',
 106             # The md5 checksum changes on each download
 107             'info_dict': {
 108                 'id': '3750436379001',
 109                 'ext': 'flv',
 110                 'title': 'UCI MTB World Cup 2014: Fort William, UK - Downhill Finals',
 111                 'uploader': 'RBTV Old (do not use)',
 112                 'description': 'UCI MTB World Cup 2014: Fort William, UK - Downhill Finals',
 113                 'timestamp': 1409122195,
 114                 'upload_date': '20140827',
 115                 'uploader_id': '710858724001',
 116             },
 117             'skip': 'Video gone',
 118         },
 119         {
 120             # playlist with 'videoList'
 121             # from http://support.brightcove.com/en/video-cloud/docs/playlist-support-single-video-players
 122             'url': 'http://c.brightcove.com/services/viewer/htmlFederated?playerID=3550052898001&playerKey=AQ%7E%7E%2CAAABmA9XpXk%7E%2C-Kp7jNgisre1fG5OdqpAFUTcs0lP_ZoL',
 123             'info_dict': {
 124                 'title': 'Sealife',
 125                 'id': '3550319591001',
 126             },
 127             'playlist_mincount': 7,
 128             'skip': 'Unsupported URL',
 129         },
 130         {
 131             # playlist with 'playlistTab' (https://github.com/ytdl-org/youtube-dl/issues/9965)
 132             'url': 'http://c.brightcove.com/services/json/experience/runtime/?command=get_programming_for_experience&playerKey=AQ%7E%7E,AAABXlLMdok%7E,NJ4EoMlZ4rZdx9eU1rkMVd8EaYPBBUlg',
 133             'info_dict': {
 134                 'id': '1522758701001',
 135                 'title': 'Lesson 08',
 136             },
 137             'playlist_mincount': 10,
 138             'skip': 'Unsupported URL',
 139         },
 140         {
 141             # playerID inferred from bcpid
 142             # from http://www.un.org/chinese/News/story.asp?NewsID=27724
 143             'url': 'https://link.brightcove.com/services/player/bcpid1722935254001/?bctid=5360463607001&autoStart=false&secureConnections=true&width=650&height=350',
 144             'only_matching': True,  # Tested in GenericIE
 145         }
 146     ]
 147
 148     @classmethod
 149     def _build_brightcove_url(cls, object_str):
 150         """
 151         Build a Brightcove url from a xml string containing
 152         <object class="BrightcoveExperience">{params}</object>
 153         """
 154
 155         # Fix up some stupid HTML, see https://github.com/ytdl-org/youtube-dl/issues/1553
 156         object_str = re.sub(r'(<param(?:\s+[a-zA-Z0-9_]+="[^"]*")*)>',
 157                             lambda m: m.group(1) + '/>', object_str)
 158         # Fix up some stupid XML, see https://github.com/ytdl-org/youtube-dl/issues/1608
 159         object_str = object_str.replace('<--', '<!--')
 160         # remove namespace to simplify extraction
 161         object_str = re.sub(r'(<object[^>]*)(xmlns=".*?")', r'\1', object_str)
 162         object_str = fix_xml_ampersands(object_str)
 163
 164         try:
 165             object_doc = compat_etree_fromstring(object_str.encode('utf-8'))
 166         except xml.etree.ElementTree.ParseError:
 167             return
 168
 169         fv_el = find_xpath_attr(object_doc, './param', 'name', 'flashVars')
 170         if fv_el is not None:
 171             flashvars = dict(
 172                 (k, v[0])
 173                 for k, v in compat_parse_qs(fv_el.attrib['value']).items())
 174         else:
 175             flashvars = {}
 176
 177         data_url = object_doc.attrib.get('data', '')
 178         data_url_params = parse_qs(data_url)
 179
 180         def find_param(name):
 181             if name in flashvars:
 182                 return flashvars[name]
 183             node = find_xpath_attr(object_doc, './param', 'name', name)
 184             if node is not None:
 185                 return node.attrib['value']
 186             return data_url_params.get(name)
 187
 188         params = {}
 189
 190         playerID = find_param('playerID') or find_param('playerId')
 191         if playerID is None:
 192             raise ExtractorError('Cannot find player ID')
 193         params['playerID'] = playerID
 194
 195         playerKey = find_param('playerKey')
 196         # Not all pages define this value
 197         if playerKey is not None:
 198             params['playerKey'] = playerKey
 199         # These fields hold the id of the video
 200         videoPlayer = find_param('@videoPlayer') or find_param('videoId') or find_param('videoID') or find_param('@videoList')
 201         if videoPlayer is not None:
 202             if isinstance(videoPlayer, list):
 203                 videoPlayer = videoPlayer[0]
 204             videoPlayer = videoPlayer.strip()
 205             # UUID is also possible for videoPlayer (e.g.
 206             # http://www.popcornflix.com/hoodies-vs-hooligans/7f2d2b87-bbf2-4623-acfb-ea942b4f01dd
 207             # or http://www8.hp.com/cn/zh/home.html)
 208             if not (re.match(
 209                     r'^(?:\d+|[\da-fA-F]{8}-?[\da-fA-F]{4}-?[\da-fA-F]{4}-?[\da-fA-F]{4}-?[\da-fA-F]{12})$',
 210                     videoPlayer) or videoPlayer.startswith('ref:')):
 211                 return None
 212             params['@videoPlayer'] = videoPlayer
 213         linkBase = find_param('linkBaseURL')
 214         if linkBase is not None:
 215             params['linkBaseURL'] = linkBase
 216         return cls._make_brightcove_url(params)
 217
 218     @classmethod
 219     def _build_brightcove_url_from_js(cls, object_js):
 220         # The layout of JS is as follows:
 221         # customBC.createVideo = function (width, height, playerID, playerKey, videoPlayer, VideoRandomID) {
 222         #   // build Brightcove <object /> XML
 223         # }
 224         m = re.search(
 225             r'''(?x)customBC\.createVideo\(
 226                 .*?                                                  # skipping width and height
 227                 ["\'](?P<playerID>\d+)["\']\s*,\s*                   # playerID
 228                 ["\'](?P<playerKey>AQ[^"\']{48})[^"\']*["\']\s*,\s*  # playerKey begins with AQ and is 50 characters
 229                                                                      # in length, however it's appended to itself
 230                                                                      # in places, so truncate
 231                 ["\'](?P<videoID>\d+)["\']                           # @videoPlayer
 232             ''', object_js)
 233         if m:
 234             return cls._make_brightcove_url(m.groupdict())
 235
 236     @classmethod
 237     def _make_brightcove_url(cls, params):
 238         return update_url_query(
 239             'http://c.brightcove.com/services/viewer/htmlFederated', params)
 240
 241     @classmethod
 242     def _extract_brightcove_url(cls, webpage):
 243         """Try to extract the brightcove url from the webpage, returns None
 244         if it can't be found
 245         """
 246         urls = cls._extract_brightcove_urls(webpage)
 247         return urls[0] if urls else None
 248
 249     @classmethod
 250     def _extract_brightcove_urls(cls, webpage):
 251         """Return a list of all Brightcove URLs from the webpage """
 252
 253         url_m = re.search(
 254             r'''(?x)
 255                 <meta\s+
 256                     (?:property|itemprop)=([\'"])(?:og:video|embedURL)\1[^>]+
 257                     content=([\'"])(?P<url>https?://(?:secure|c)\.brightcove.com/(?:(?!\2).)+)\2
 258             ''', webpage)
 259         if url_m:
 260             url = unescapeHTML(url_m.group('url'))
 261             # Some sites don't add it, we can't download with this url, for example:
 262             # http://www.ktvu.com/videos/news/raw-video-caltrain-releases-video-of-man-almost/vCTZdY/
 263             if 'playerKey' in url or 'videoId' in url or 'idVideo' in url:
 264                 return [url]
 265
 266         matches = re.findall(
 267             r'''(?sx)<object
 268             (?:
 269                 [^>]+?class=[\'"][^>]*?BrightcoveExperience.*?[\'"] |
 270                 [^>]*?>\s*<param\s+name="movie"\s+value="https?://[^/]*brightcove\.com/
 271             ).+?>\s*</object>''',
 272             webpage)
 273         if matches:
 274             return list(filter(None, [cls._build_brightcove_url(m) for m in matches]))
 275
 276         matches = re.findall(r'(customBC\.createVideo\(.+?\);)', webpage)
 277         if matches:
 278             return list(filter(None, [
 279                 cls._build_brightcove_url_from_js(custom_bc)
 280                 for custom_bc in matches]))
 281         return [src for _, src in re.findall(
 282             r'<iframe[^>]+src=([\'"])((?:https?:)?//link\.brightcove\.com/services/player/(?!\1).+)\1', webpage)]
 283
 284     def _real_extract(self, url):
 285         url, smuggled_data = unsmuggle_url(url, {})
 286
 287         # Change the 'videoId' and others field to '@videoPlayer'
 288         url = re.sub(r'(?<=[?&])(videoI(d|D)|idVideo|bctid)', '%40videoPlayer', url)
 289         # Change bckey (used by bcove.me urls) to playerKey
 290         url = re.sub(r'(?<=[?&])bckey', 'playerKey', url)
 291         mobj = self._match_valid_url(url)
 292         query_str = mobj.group('query')
 293         query = compat_urlparse.parse_qs(query_str)
 294
 295         videoPlayer = query.get('@videoPlayer')
 296         if videoPlayer:
 297             # We set the original url as the default 'Referer' header
 298             referer = query.get('linkBaseURL', [None])[0] or smuggled_data.get('Referer', url)
 299             video_id = videoPlayer[0]
 300             if 'playerID' not in query:
 301                 mobj = re.search(r'/bcpid(\d+)', url)
 302                 if mobj is not None:
 303                     query['playerID'] = [mobj.group(1)]
 304             publisher_id = query.get('publisherId')
 305             if publisher_id and publisher_id[0].isdigit():
 306                 publisher_id = publisher_id[0]
 307             if not publisher_id:
 308                 player_key = query.get('playerKey')
 309                 if player_key and ',' in player_key[0]:
 310                     player_key = player_key[0]
 311                 else:
 312                     player_id = query.get('playerID')
 313                     if player_id and player_id[0].isdigit():
 314                         headers = {}
 315                         if referer:
 316                             headers['Referer'] = referer
 317                         player_page = self._download_webpage(
 318                             'http://link.brightcove.com/services/player/bcpid' + player_id[0],
 319                             video_id, headers=headers, fatal=False)
 320                         if player_page:
 321                             player_key = self._search_regex(
 322                                 r'<param\s+name="playerKey"\s+value="([\w~,-]+)"',
 323                                 player_page, 'player key', fatal=False)
 324                 if player_key:
 325                     enc_pub_id = player_key.split(',')[1].replace('~', '=')
 326                     publisher_id = struct.unpack('>Q', base64.urlsafe_b64decode(enc_pub_id))[0]
 327             if publisher_id:
 328                 brightcove_new_url = 'http://players.brightcove.net/%s/default_default/index.html?videoId=%s' % (publisher_id, video_id)
 329                 if referer:
 330                     brightcove_new_url = smuggle_url(brightcove_new_url, {'referrer': referer})
 331                 return self.url_result(brightcove_new_url, BrightcoveNewIE.ie_key(), video_id)
 332         # TODO: figure out if it's possible to extract playlistId from playerKey
 333         # elif 'playerKey' in query:
 334         #     player_key = query['playerKey']
 335         #     return self._get_playlist_info(player_key[0])
 336         raise UnsupportedError(url)
 337
 338
 339 class BrightcoveNewIE(AdobePassIE):
 340     IE_NAME = 'brightcove:new'
 341     _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:[^&]+)'
 342     _TESTS = [{
 343         'url': 'http://players.brightcove.net/929656772001/e41d32dc-ec74-459e-a845-6c69f7b724ea_default/index.html?videoId=4463358922001',
 344         'md5': 'c8100925723840d4b0d243f7025703be',
 345         'info_dict': {
 346             'id': '4463358922001',
 347             'ext': 'mp4',
 348             'title': 'Meet the man behind Popcorn Time',
 349             'description': 'md5:eac376a4fe366edc70279bfb681aea16',
 350             'duration': 165.768,
 351             'timestamp': 1441391203,
 352             'upload_date': '20150904',
 353             'uploader_id': '929656772001',
 354             'formats': 'mincount:20',
 355         },
 356     }, {
 357         # with rtmp streams
 358         'url': 'http://players.brightcove.net/4036320279001/5d112ed9-283f-485f-a7f9-33f42e8bc042_default/index.html?videoId=4279049078001',
 359         'info_dict': {
 360             'id': '4279049078001',
 361             'ext': 'mp4',
 362             'title': 'Titansgrave: Chapter 0',
 363             'description': 'Titansgrave: Chapter 0',
 364             'duration': 1242.058,
 365             'timestamp': 1433556729,
 366             'upload_date': '20150606',
 367             'uploader_id': '4036320279001',
 368             'formats': 'mincount:39',
 369         },
 370         'params': {
 371             # m3u8 download
 372             'skip_download': True,
 373         }
 374     }, {
 375         # playlist stream
 376         'url': 'https://players.brightcove.net/1752604059001/S13cJdUBz_default/index.html?playlistId=5718313430001',
 377         'info_dict': {
 378             'id': '5718313430001',
 379             'title': 'No Audio Playlist',
 380         },
 381         'playlist_count': 7,
 382         'params': {
 383             # m3u8 download
 384             'skip_download': True,
 385         }
 386     }, {
 387         'url': 'http://players.brightcove.net/5690807595001/HyZNerRl7_default/index.html?playlistId=5743160747001',
 388         'only_matching': True,
 389     }, {
 390         # ref: prefixed video id
 391         'url': 'http://players.brightcove.net/3910869709001/21519b5c-4b3b-4363-accb-bdc8f358f823_default/index.html?videoId=ref:7069442',
 392         'only_matching': True,
 393     }, {
 394         # non numeric ref: prefixed video id
 395         'url': 'http://players.brightcove.net/710858724001/default_default/index.html?videoId=ref:event-stream-356',
 396         'only_matching': True,
 397     }, {
 398         # unavailable video without message but with error_code
 399         'url': 'http://players.brightcove.net/1305187701/c832abfb-641b-44eb-9da0-2fe76786505f_default/index.html?videoId=4377407326001',
 400         'only_matching': True,
 401     }]
 402
 403     @staticmethod
 404     def _extract_url(ie, webpage):
 405         urls = BrightcoveNewIE._extract_brightcove_urls(ie, webpage)
 406         return urls[0] if urls else None
 407
 408     @staticmethod
 409     def _extract_brightcove_urls(ie, webpage):
 410         # Reference:
 411         # 1. http://docs.brightcove.com/en/video-cloud/brightcove-player/guides/publish-video.html#setvideoiniframe
 412         # 2. http://docs.brightcove.com/en/video-cloud/brightcove-player/guides/publish-video.html#tag
 413         # 3. http://docs.brightcove.com/en/video-cloud/brightcove-player/guides/publish-video.html#setvideousingjavascript
 414         # 4. http://docs.brightcove.com/en/video-cloud/brightcove-player/guides/in-page-embed-player-implementation.html
 415         # 5. https://support.brightcove.com/en/video-cloud/docs/dynamically-assigning-videos-player
 416
 417         entries = []
 418
 419         # Look for iframe embeds [1]
 420         for _, url in re.findall(
 421                 r'<iframe[^>]+src=(["\'])((?:https?:)?//players\.brightcove\.net/\d+/[^/]+/index\.html.+?)\1', webpage):
 422             entries.append(url if url.startswith('http') else 'http:' + url)
 423
 424         # Look for <video> tags [2] and embed_in_page embeds [3]
 425         # [2] looks like:
 426         for video, script_tag, account_id, player_id, embed in re.findall(
 427                 r'''(?isx)
 428                     (<video(?:-js)?\s+[^>]*\bdata-video-id\s*=\s*['"]?[^>]+>)
 429                     (?:.*?
 430                         (<script[^>]+
 431                             src=["\'](?:https?:)?//players\.brightcove\.net/
 432                             (\d+)/([^/]+)_([^/]+)/index(?:\.min)?\.js
 433                         )
 434                     )?
 435                 ''', webpage):
 436             attrs = extract_attributes(video)
 437
 438             # According to examples from [4] it's unclear whether video id
 439             # may be optional and what to do when it is
 440             video_id = attrs.get('data-video-id')
 441             if not video_id:
 442                 continue
 443
 444             account_id = account_id or attrs.get('data-account')
 445             if not account_id:
 446                 continue
 447
 448             player_id = player_id or attrs.get('data-player') or 'default'
 449             embed = embed or attrs.get('data-embed') or 'default'
 450
 451             bc_url = 'http://players.brightcove.net/%s/%s_%s/index.html?videoId=%s' % (
 452                 account_id, player_id, embed, video_id)
 453
 454             # Some brightcove videos may be embedded with video tag only and
 455             # without script tag or any mentioning of brightcove at all. Such
 456             # embeds are considered ambiguous since they are matched based only
 457             # on data-video-id and data-account attributes and in the wild may
 458             # not be brightcove embeds at all. Let's check reconstructed
 459             # brightcove URLs in case of such embeds and only process valid
 460             # ones. By this we ensure there is indeed a brightcove embed.
 461             if not script_tag and not ie._is_valid_url(
 462                     bc_url, video_id, 'possible brightcove video'):
 463                 continue
 464
 465             entries.append(bc_url)
 466
 467         return entries
 468
 469     def _parse_brightcove_metadata(self, json_data, video_id, headers={}):
 470         title = json_data['name'].strip()
 471
 472         formats, subtitles = [], {}
 473         sources = json_data.get('sources') or []
 474         for source in sources:
 475             container = source.get('container')
 476             ext = mimetype2ext(source.get('type'))
 477             src = source.get('src')
 478             if ext == 'm3u8' or container == 'M2TS':
 479                 if not src:
 480                     continue
 481                 fmts, subs = self._extract_m3u8_formats_and_subtitles(
 482                     src, video_id, 'mp4', 'm3u8_native', m3u8_id='hls', fatal=False)
 483                 subtitles = self._merge_subtitles(subtitles, subs)
 484             elif ext == 'mpd':
 485                 if not src:
 486                     continue
 487                 fmts, subs = self._extract_mpd_formats_and_subtitles(src, video_id, 'dash', fatal=False)
 488                 subtitles = self._merge_subtitles(subtitles, subs)
 489             else:
 490                 streaming_src = source.get('streaming_src')
 491                 stream_name, app_name = source.get('stream_name'), source.get('app_name')
 492                 if not src and not streaming_src and (not stream_name or not app_name):
 493                     continue
 494                 tbr = float_or_none(source.get('avg_bitrate'), 1000)
 495                 height = int_or_none(source.get('height'))
 496                 width = int_or_none(source.get('width'))
 497                 f = {
 498                     'tbr': tbr,
 499                     'filesize': int_or_none(source.get('size')),
 500                     'container': container,
 501                     'ext': ext or container.lower(),
 502                 }
 503                 if width == 0 and height == 0:
 504                     f.update({
 505                         'vcodec': 'none',
 506                     })
 507                 else:
 508                     f.update({
 509                         'width': width,
 510                         'height': height,
 511                         'vcodec': source.get('codec'),
 512                     })
 513
 514                 def build_format_id(kind):
 515                     format_id = kind
 516                     if tbr:
 517                         format_id += '-%dk' % int(tbr)
 518                     if height:
 519                         format_id += '-%dp' % height
 520                     return format_id
 521
 522                 if src or streaming_src:
 523                     f.update({
 524                         'url': src or streaming_src,
 525                         'format_id': build_format_id('http' if src else 'http-streaming'),
 526                         'source_preference': 0 if src else -1,
 527                     })
 528                 else:
 529                     f.update({
 530                         'url': app_name,
 531                         'play_path': stream_name,
 532                         'format_id': build_format_id('rtmp'),
 533                     })
 534                 fmts = [f]
 535
 536             # https://support.brightcove.com/playback-api-video-fields-reference#key_systems_object
 537             if container == 'WVM' or source.get('key_systems') or ext == 'ism':
 538                 for f in fmts:
 539                     f['has_drm'] = True
 540             formats.extend(fmts)
 541
 542         if not formats:
 543             errors = json_data.get('errors')
 544             if errors:
 545                 error = errors[0]
 546                 self.raise_no_formats(
 547                     error.get('message') or error.get('error_subcode') or error['error_code'], expected=True)
 548
 549         self._sort_formats(formats)
 550
 551         for f in formats:
 552             f.setdefault('http_headers', {}).update(headers)
 553
 554         for text_track in json_data.get('text_tracks', []):
 555             if text_track.get('kind') != 'captions':
 556                 continue
 557             text_track_url = url_or_none(text_track.get('src'))
 558             if not text_track_url:
 559                 continue
 560             lang = (str_or_none(text_track.get('srclang'))
 561                     or str_or_none(text_track.get('label')) or 'en').lower()
 562             subtitles.setdefault(lang, []).append({
 563                 'url': text_track_url,
 564             })
 565
 566         is_live = False
 567         duration = float_or_none(json_data.get('duration'), 1000)
 568         if duration is not None and duration <= 0:
 569             is_live = True
 570
 571         common_res = [(160, 90), (320, 180), (480, 720), (640, 360), (768, 432), (1024, 576), (1280, 720), (1366, 768), (1920, 1080)]
 572         thumb_base_url = dict_get(json_data, ('poster', 'thumbnail'))
 573         thumbnails = [{
 574             'url': re.sub(r'\d+x\d+', f'{w}x{h}', thumb_base_url),
 575             'width': w,
 576             'height': h,
 577         } for w, h in common_res] if thumb_base_url else None
 578
 579         return {
 580             'id': video_id,
 581             'title': title,
 582             'description': clean_html(json_data.get('description')),
 583             'thumbnails': thumbnails,
 584             'duration': duration,
 585             'timestamp': parse_iso8601(json_data.get('published_at')),
 586             'uploader_id': json_data.get('account_id'),
 587             'formats': formats,
 588             'subtitles': subtitles,
 589             'tags': json_data.get('tags', []),
 590             'is_live': is_live,
 591         }
 592
 593     def _real_extract(self, url):
 594         url, smuggled_data = unsmuggle_url(url, {})
 595         self._initialize_geo_bypass({
 596             'countries': smuggled_data.get('geo_countries'),
 597             'ip_blocks': smuggled_data.get('geo_ip_blocks'),
 598         })
 599
 600         account_id, player_id, embed, content_type, video_id = self._match_valid_url(url).groups()
 601
 602         policy_key_id = '%s_%s' % (account_id, player_id)
 603         policy_key = self.cache.load('brightcove', policy_key_id)
 604         policy_key_extracted = False
 605         store_pk = lambda x: self.cache.store('brightcove', policy_key_id, x)
 606
 607         def extract_policy_key():
 608             base_url = 'http://players.brightcove.net/%s/%s_%s/' % (account_id, player_id, embed)
 609             config = self._download_json(
 610                 base_url + 'config.json', video_id, fatal=False) or {}
 611             policy_key = try_get(
 612                 config, lambda x: x['video_cloud']['policy_key'])
 613             if not policy_key:
 614                 webpage = self._download_webpage(
 615                     base_url + 'index.min.js', video_id)
 616
 617                 catalog = self._search_regex(
 618                     r'catalog\(({.+?})\);', webpage, 'catalog', default=None)
 619                 if catalog:
 620                     catalog = self._parse_json(
 621                         js_to_json(catalog), video_id, fatal=False)
 622                     if catalog:
 623                         policy_key = catalog.get('policyKey')
 624
 625                 if not policy_key:
 626                     policy_key = self._search_regex(
 627                         r'policyKey\s*:\s*(["\'])(?P<pk>.+?)\1',
 628                         webpage, 'policy key', group='pk')
 629
 630             store_pk(policy_key)
 631             return policy_key
 632
 633         api_url = 'https://edge.api.brightcove.com/playback/v1/accounts/%s/%ss/%s' % (account_id, content_type, video_id)
 634         headers = {}
 635         referrer = smuggled_data.get('referrer')
 636         if referrer:
 637             headers.update({
 638                 'Referer': referrer,
 639                 'Origin': re.search(r'https?://[^/]+', referrer).group(0),
 640             })
 641
 642         for _ in range(2):
 643             if not policy_key:
 644                 policy_key = extract_policy_key()
 645                 policy_key_extracted = True
 646             headers['Accept'] = 'application/json;pk=%s' % policy_key
 647             try:
 648                 json_data = self._download_json(api_url, video_id, headers=headers)
 649                 break
 650             except ExtractorError as e:
 651                 if isinstance(e.cause, compat_HTTPError) and e.cause.code in (401, 403):
 652                     json_data = self._parse_json(e.cause.read().decode(), video_id)[0]
 653                     message = json_data.get('message') or json_data['error_code']
 654                     if json_data.get('error_subcode') == 'CLIENT_GEO':
 655                         self.raise_geo_restricted(msg=message)
 656                     elif json_data.get('error_code') == 'INVALID_POLICY_KEY' and not policy_key_extracted:
 657                         policy_key = None
 658                         store_pk(None)
 659                         continue
 660                     raise ExtractorError(message, expected=True)
 661                 raise
 662
 663         errors = json_data.get('errors')
 664         if errors and errors[0].get('error_subcode') == 'TVE_AUTH':
 665             custom_fields = json_data['custom_fields']
 666             tve_token = self._extract_mvpd_auth(
 667                 smuggled_data['source_url'], video_id,
 668                 custom_fields['bcadobepassrequestorid'],
 669                 custom_fields['bcadobepassresourceid'])
 670             json_data = self._download_json(
 671                 api_url, video_id, headers={
 672                     'Accept': 'application/json;pk=%s' % policy_key
 673                 }, query={
 674                     'tveToken': tve_token,
 675                 })
 676
 677         if content_type == 'playlist':
 678             return self.playlist_result(
 679                 [self._parse_brightcove_metadata(vid, vid.get('id'), headers)
 680                  for vid in json_data.get('videos', []) if vid.get('id')],
 681                 json_data.get('id'), json_data.get('name'),
 682                 json_data.get('description'))
 683
 684         return self._parse_brightcove_metadata(
 685             json_data, video_id, headers=headers)