youtube_dl/extractor/brightcove.py

   1 # coding: utf-8
   2 from __future__ import unicode_literals
   3
   4 import base64
   5 import re
   6 import struct
   7
   8 from .adobepass import AdobePassIE
   9 from .common import InfoExtractor
  10 from ..compat import (
  11     compat_etree_fromstring,
  12     compat_HTTPError,
  13     compat_parse_qs,
  14     compat_urllib_parse_urlparse,
  15     compat_urlparse,
  16     compat_xml_parse_error,
  17 )
  18 from ..utils import (
  19     clean_html,
  20     extract_attributes,
  21     ExtractorError,
  22     find_xpath_attr,
  23     fix_xml_ampersands,
  24     float_or_none,
  25     int_or_none,
  26     js_to_json,
  27     mimetype2ext,
  28     parse_iso8601,
  29     smuggle_url,
  30     str_or_none,
  31     try_get,
  32     unescapeHTML,
  33     unsmuggle_url,
  34     UnsupportedError,
  35     update_url_query,
  36     url_or_none,
  37 )
  38
  39
  40 class BrightcoveLegacyIE(InfoExtractor):
  41     IE_NAME = 'brightcove:legacy'
  42     _VALID_URL = r'(?:https?://.*brightcove\.com/(services|viewer).*?\?|brightcove:)(?P<query>.*)'
  43
  44     _TESTS = [
  45         {
  46             # From http://www.8tv.cat/8aldia/videos/xavier-sala-i-martin-aquesta-tarda-a-8-al-dia/
  47             'url': 'http://c.brightcove.com/services/viewer/htmlFederated?playerID=1654948606001&flashID=myExperience&%40videoPlayer=2371591881001',
  48             'md5': '5423e113865d26e40624dce2e4b45d95',
  49             'note': 'Test Brightcove downloads and detection in GenericIE',
  50             'info_dict': {
  51                 'id': '2371591881001',
  52                 'ext': 'mp4',
  53                 'title': 'Xavier Sala i Martín: “Un banc que no presta és un banc zombi que no serveix per a res”',
  54                 'uploader': '8TV',
  55                 'description': 'md5:a950cc4285c43e44d763d036710cd9cd',
  56                 'timestamp': 1368213670,
  57                 'upload_date': '20130510',
  58                 'uploader_id': '1589608506001',
  59             },
  60             'skip': 'The player has been deactivated by the content owner',
  61         },
  62         {
  63             # From http://medianetwork.oracle.com/video/player/1785452137001
  64             'url': 'http://c.brightcove.com/services/viewer/htmlFederated?playerID=1217746023001&flashID=myPlayer&%40videoPlayer=1785452137001',
  65             'info_dict': {
  66                 'id': '1785452137001',
  67                 'ext': 'flv',
  68                 'title': 'JVMLS 2012: Arrays 2.0 - Opportunities and Challenges',
  69                 'description': 'John Rose speaks at the JVM Language Summit, August 1, 2012.',
  70                 'uploader': 'Oracle',
  71                 'timestamp': 1344975024,
  72                 'upload_date': '20120814',
  73                 'uploader_id': '1460825906',
  74             },
  75             'skip': 'video not playable',
  76         },
  77         {
  78             # From http://mashable.com/2013/10/26/thermoelectric-bracelet-lets-you-control-your-body-temperature/
  79             'url': 'http://c.brightcove.com/services/viewer/federated_f9?&playerID=1265504713001&publisherID=AQ%7E%7E%2CAAABBzUwv1E%7E%2CxP-xFHVUstiMFlNYfvF4G9yFnNaqCw_9&videoID=2750934548001',
  80             'info_dict': {
  81                 'id': '2750934548001',
  82                 'ext': 'mp4',
  83                 'title': 'This Bracelet Acts as a Personal Thermostat',
  84                 'description': 'md5:547b78c64f4112766ccf4e151c20b6a0',
  85                 # 'uploader': 'Mashable',
  86                 'timestamp': 1382041798,
  87                 'upload_date': '20131017',
  88                 'uploader_id': '1130468786001',
  89             },
  90         },
  91         {
  92             # test that the default referer works
  93             # from http://national.ballet.ca/interact/video/Lost_in_Motion_II/
  94             'url': 'http://link.brightcove.com/services/player/bcpid756015033001?bckey=AQ~~,AAAApYJi_Ck~,GxhXCegT1Dp39ilhXuxMJxasUhVNZiil&bctid=2878862109001',
  95             'info_dict': {
  96                 'id': '2878862109001',
  97                 'ext': 'mp4',
  98                 'title': 'Lost in Motion II',
  99                 'description': 'md5:363109c02998fee92ec02211bd8000df',
 100                 'uploader': 'National Ballet of Canada',
 101             },
 102             'skip': 'Video gone',
 103         },
 104         {
 105             # test flv videos served by akamaihd.net
 106             # From http://www.redbull.com/en/bike/stories/1331655643987/replay-uci-dh-world-cup-2014-from-fort-william
 107             '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',
 108             # The md5 checksum changes on each download
 109             'info_dict': {
 110                 'id': '3750436379001',
 111                 'ext': 'flv',
 112                 'title': 'UCI MTB World Cup 2014: Fort William, UK - Downhill Finals',
 113                 'uploader': 'RBTV Old (do not use)',
 114                 'description': 'UCI MTB World Cup 2014: Fort William, UK - Downhill Finals',
 115                 'timestamp': 1409122195,
 116                 'upload_date': '20140827',
 117                 'uploader_id': '710858724001',
 118             },
 119             'skip': 'Video gone',
 120         },
 121         {
 122             # playlist with 'videoList'
 123             # from http://support.brightcove.com/en/video-cloud/docs/playlist-support-single-video-players
 124             'url': 'http://c.brightcove.com/services/viewer/htmlFederated?playerID=3550052898001&playerKey=AQ%7E%7E%2CAAABmA9XpXk%7E%2C-Kp7jNgisre1fG5OdqpAFUTcs0lP_ZoL',
 125             'info_dict': {
 126                 'title': 'Sealife',
 127                 'id': '3550319591001',
 128             },
 129             'playlist_mincount': 7,
 130             'skip': 'Unsupported URL',
 131         },
 132         {
 133             # playlist with 'playlistTab' (https://github.com/ytdl-org/youtube-dl/issues/9965)
 134             'url': 'http://c.brightcove.com/services/json/experience/runtime/?command=get_programming_for_experience&playerKey=AQ%7E%7E,AAABXlLMdok%7E,NJ4EoMlZ4rZdx9eU1rkMVd8EaYPBBUlg',
 135             'info_dict': {
 136                 'id': '1522758701001',
 137                 'title': 'Lesson 08',
 138             },
 139             'playlist_mincount': 10,
 140             'skip': 'Unsupported URL',
 141         },
 142         {
 143             # playerID inferred from bcpid
 144             # from http://www.un.org/chinese/News/story.asp?NewsID=27724
 145             'url': 'https://link.brightcove.com/services/player/bcpid1722935254001/?bctid=5360463607001&autoStart=false&secureConnections=true&width=650&height=350',
 146             'only_matching': True,  # Tested in GenericIE
 147         }
 148     ]
 149
 150     @classmethod
 151     def _build_brightcove_url(cls, object_str):
 152         """
 153         Build a Brightcove url from a xml string containing
 154         <object class="BrightcoveExperience">{params}</object>
 155         """
 156
 157         # Fix up some stupid HTML, see https://github.com/ytdl-org/youtube-dl/issues/1553
 158         object_str = re.sub(r'(<param(?:\s+[a-zA-Z0-9_]+="[^"]*")*)>',
 159                             lambda m: m.group(1) + '/>', object_str)
 160         # Fix up some stupid XML, see https://github.com/ytdl-org/youtube-dl/issues/1608
 161         object_str = object_str.replace('<--', '<!--')
 162         # remove namespace to simplify extraction
 163         object_str = re.sub(r'(<object[^>]*)(xmlns=".*?")', r'\1', object_str)
 164         object_str = fix_xml_ampersands(object_str)
 165
 166         try:
 167             object_doc = compat_etree_fromstring(object_str.encode('utf-8'))
 168         except compat_xml_parse_error:
 169             return
 170
 171         fv_el = find_xpath_attr(object_doc, './param', 'name', 'flashVars')
 172         if fv_el is not None:
 173             flashvars = dict(
 174                 (k, v[0])
 175                 for k, v in compat_parse_qs(fv_el.attrib['value']).items())
 176         else:
 177             flashvars = {}
 178
 179         data_url = object_doc.attrib.get('data', '')
 180         data_url_params = compat_parse_qs(compat_urllib_parse_urlparse(data_url).query)
 181
 182         def find_param(name):
 183             if name in flashvars:
 184                 return flashvars[name]
 185             node = find_xpath_attr(object_doc, './param', 'name', name)
 186             if node is not None:
 187                 return node.attrib['value']
 188             return data_url_params.get(name)
 189
 190         params = {}
 191
 192         playerID = find_param('playerID') or find_param('playerId')
 193         if playerID is None:
 194             raise ExtractorError('Cannot find player ID')
 195         params['playerID'] = playerID
 196
 197         playerKey = find_param('playerKey')
 198         # Not all pages define this value
 199         if playerKey is not None:
 200             params['playerKey'] = playerKey
 201         # These fields hold the id of the video
 202         videoPlayer = find_param('@videoPlayer') or find_param('videoId') or find_param('videoID') or find_param('@videoList')
 203         if videoPlayer is not None:
 204             if isinstance(videoPlayer, list):
 205                 videoPlayer = videoPlayer[0]
 206             videoPlayer = videoPlayer.strip()
 207             # UUID is also possible for videoPlayer (e.g.
 208             # http://www.popcornflix.com/hoodies-vs-hooligans/7f2d2b87-bbf2-4623-acfb-ea942b4f01dd
 209             # or http://www8.hp.com/cn/zh/home.html)
 210             if not (re.match(
 211                     r'^(?:\d+|[\da-fA-F]{8}-?[\da-fA-F]{4}-?[\da-fA-F]{4}-?[\da-fA-F]{4}-?[\da-fA-F]{12})$',
 212                     videoPlayer) or videoPlayer.startswith('ref:')):
 213                 return None
 214             params['@videoPlayer'] = videoPlayer
 215         linkBase = find_param('linkBaseURL')
 216         if linkBase is not None:
 217             params['linkBaseURL'] = linkBase
 218         return cls._make_brightcove_url(params)
 219
 220     @classmethod
 221     def _build_brightcove_url_from_js(cls, object_js):
 222         # The layout of JS is as follows:
 223         # customBC.createVideo = function (width, height, playerID, playerKey, videoPlayer, VideoRandomID) {
 224         #   // build Brightcove <object /> XML
 225         # }
 226         m = re.search(
 227             r'''(?x)customBC\.createVideo\(
 228                 .*?                                                  # skipping width and height
 229                 ["\'](?P<playerID>\d+)["\']\s*,\s*                   # playerID
 230                 ["\'](?P<playerKey>AQ[^"\']{48})[^"\']*["\']\s*,\s*  # playerKey begins with AQ and is 50 characters
 231                                                                      # in length, however it's appended to itself
 232                                                                      # in places, so truncate
 233                 ["\'](?P<videoID>\d+)["\']                           # @videoPlayer
 234             ''', object_js)
 235         if m:
 236             return cls._make_brightcove_url(m.groupdict())
 237
 238     @classmethod
 239     def _make_brightcove_url(cls, params):
 240         return update_url_query(
 241             'http://c.brightcove.com/services/viewer/htmlFederated', params)
 242
 243     @classmethod
 244     def _extract_brightcove_url(cls, webpage):
 245         """Try to extract the brightcove url from the webpage, returns None
 246         if it can't be found
 247         """
 248         urls = cls._extract_brightcove_urls(webpage)
 249         return urls[0] if urls else None
 250
 251     @classmethod
 252     def _extract_brightcove_urls(cls, webpage):
 253         """Return a list of all Brightcove URLs from the webpage """
 254
 255         url_m = re.search(
 256             r'''(?x)
 257                 <meta\s+
 258                     (?:property|itemprop)=([\'"])(?:og:video|embedURL)\1[^>]+
 259                     content=([\'"])(?P<url>https?://(?:secure|c)\.brightcove.com/(?:(?!\2).)+)\2
 260             ''', webpage)
 261         if url_m:
 262             url = unescapeHTML(url_m.group('url'))
 263             # Some sites don't add it, we can't download with this url, for example:
 264             # http://www.ktvu.com/videos/news/raw-video-caltrain-releases-video-of-man-almost/vCTZdY/
 265             if 'playerKey' in url or 'videoId' in url or 'idVideo' in url:
 266                 return [url]
 267
 268         matches = re.findall(
 269             r'''(?sx)<object
 270             (?:
 271                 [^>]+?class=[\'"][^>]*?BrightcoveExperience.*?[\'"] |
 272                 [^>]*?>\s*<param\s+name="movie"\s+value="https?://[^/]*brightcove\.com/
 273             ).+?>\s*</object>''',
 274             webpage)
 275         if matches:
 276             return list(filter(None, [cls._build_brightcove_url(m) for m in matches]))
 277
 278         matches = re.findall(r'(customBC\.createVideo\(.+?\);)', webpage)
 279         if matches:
 280             return list(filter(None, [
 281                 cls._build_brightcove_url_from_js(custom_bc)
 282                 for custom_bc in matches]))
 283         return [src for _, src in re.findall(
 284             r'<iframe[^>]+src=([\'"])((?:https?:)?//link\.brightcove\.com/services/player/(?!\1).+)\1', webpage)]
 285
 286     def _real_extract(self, url):
 287         url, smuggled_data = unsmuggle_url(url, {})
 288
 289         # Change the 'videoId' and others field to '@videoPlayer'
 290         url = re.sub(r'(?<=[?&])(videoI(d|D)|idVideo|bctid)', '%40videoPlayer', url)
 291         # Change bckey (used by bcove.me urls) to playerKey
 292         url = re.sub(r'(?<=[?&])bckey', 'playerKey', url)
 293         mobj = re.match(self._VALID_URL, url)
 294         query_str = mobj.group('query')
 295         query = compat_urlparse.parse_qs(query_str)
 296
 297         videoPlayer = query.get('@videoPlayer')
 298         if videoPlayer:
 299             # We set the original url as the default 'Referer' header
 300             referer = query.get('linkBaseURL', [None])[0] or smuggled_data.get('Referer', url)
 301             video_id = videoPlayer[0]
 302             if 'playerID' not in query:
 303                 mobj = re.search(r'/bcpid(\d+)', url)
 304                 if mobj is not None:
 305                     query['playerID'] = [mobj.group(1)]
 306             publisher_id = query.get('publisherId')
 307             if publisher_id and publisher_id[0].isdigit():
 308                 publisher_id = publisher_id[0]
 309             if not publisher_id:
 310                 player_key = query.get('playerKey')
 311                 if player_key and ',' in player_key[0]:
 312                     player_key = player_key[0]
 313                 else:
 314                     player_id = query.get('playerID')
 315                     if player_id and player_id[0].isdigit():
 316                         headers = {}
 317                         if referer:
 318                             headers['Referer'] = referer
 319                         player_page = self._download_webpage(
 320                             'http://link.brightcove.com/services/player/bcpid' + player_id[0],
 321                             video_id, headers=headers, fatal=False)
 322                         if player_page:
 323                             player_key = self._search_regex(
 324                                 r'<param\s+name="playerKey"\s+value="([\w~,-]+)"',
 325                                 player_page, 'player key', fatal=False)
 326                 if player_key:
 327                     enc_pub_id = player_key.split(',')[1].replace('~', '=')
 328                     publisher_id = struct.unpack('>Q', base64.urlsafe_b64decode(enc_pub_id))[0]
 329             if publisher_id:
 330                 brightcove_new_url = 'http://players.brightcove.net/%s/default_default/index.html?videoId=%s' % (publisher_id, video_id)
 331                 if referer:
 332                     brightcove_new_url = smuggle_url(brightcove_new_url, {'referrer': referer})
 333                 return self.url_result(brightcove_new_url, BrightcoveNewIE.ie_key(), video_id)
 334         # TODO: figure out if it's possible to extract playlistId from playerKey
 335         # elif 'playerKey' in query:
 336         #     player_key = query['playerKey']
 337         #     return self._get_playlist_info(player_key[0])
 338         raise UnsupportedError(url)
 339
 340
 341 class BrightcoveNewIE(AdobePassIE):
 342     IE_NAME = 'brightcove:new'
 343     _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:[^&]+)'
 344     _TESTS = [{
 345         'url': 'http://players.brightcove.net/929656772001/e41d32dc-ec74-459e-a845-6c69f7b724ea_default/index.html?videoId=4463358922001',
 346         'md5': 'c8100925723840d4b0d243f7025703be',
 347         'info_dict': {
 348             'id': '4463358922001',
 349             'ext': 'mp4',
 350             'title': 'Meet the man behind Popcorn Time',
 351             'description': 'md5:eac376a4fe366edc70279bfb681aea16',
 352             'duration': 165.768,
 353             'timestamp': 1441391203,
 354             'upload_date': '20150904',
 355             'uploader_id': '929656772001',
 356             'formats': 'mincount:20',
 357         },
 358     }, {
 359         # with rtmp streams
 360         'url': 'http://players.brightcove.net/4036320279001/5d112ed9-283f-485f-a7f9-33f42e8bc042_default/index.html?videoId=4279049078001',
 361         'info_dict': {
 362             'id': '4279049078001',
 363             'ext': 'mp4',
 364             'title': 'Titansgrave: Chapter 0',
 365             'description': 'Titansgrave: Chapter 0',
 366             'duration': 1242.058,
 367             'timestamp': 1433556729,
 368             'upload_date': '20150606',
 369             'uploader_id': '4036320279001',
 370             'formats': 'mincount:39',
 371         },
 372         'params': {
 373             # m3u8 download
 374             'skip_download': True,
 375         }
 376     }, {
 377         # playlist stream
 378         'url': 'https://players.brightcove.net/1752604059001/S13cJdUBz_default/index.html?playlistId=5718313430001',
 379         'info_dict': {
 380             'id': '5718313430001',
 381             'title': 'No Audio Playlist',
 382         },
 383         'playlist_count': 7,
 384         'params': {
 385             # m3u8 download
 386             'skip_download': True,
 387         }
 388     }, {
 389         'url': 'http://players.brightcove.net/5690807595001/HyZNerRl7_default/index.html?playlistId=5743160747001',
 390         'only_matching': True,
 391     }, {
 392         # ref: prefixed video id
 393         'url': 'http://players.brightcove.net/3910869709001/21519b5c-4b3b-4363-accb-bdc8f358f823_default/index.html?videoId=ref:7069442',
 394         'only_matching': True,
 395     }, {
 396         # non numeric ref: prefixed video id
 397         'url': 'http://players.brightcove.net/710858724001/default_default/index.html?videoId=ref:event-stream-356',
 398         'only_matching': True,
 399     }, {
 400         # unavailable video without message but with error_code
 401         'url': 'http://players.brightcove.net/1305187701/c832abfb-641b-44eb-9da0-2fe76786505f_default/index.html?videoId=4377407326001',
 402         'only_matching': True,
 403     }]
 404
 405     @staticmethod
 406     def _extract_url(ie, webpage):
 407         urls = BrightcoveNewIE._extract_urls(ie, webpage)
 408         return urls[0] if urls else None
 409
 410     @staticmethod
 411     def _extract_urls(ie, webpage):
 412         # Reference:
 413         # 1. http://docs.brightcove.com/en/video-cloud/brightcove-player/guides/publish-video.html#setvideoiniframe
 414         # 2. http://docs.brightcove.com/en/video-cloud/brightcove-player/guides/publish-video.html#tag
 415         # 3. http://docs.brightcove.com/en/video-cloud/brightcove-player/guides/publish-video.html#setvideousingjavascript
 416         # 4. http://docs.brightcove.com/en/video-cloud/brightcove-player/guides/in-page-embed-player-implementation.html
 417         # 5. https://support.brightcove.com/en/video-cloud/docs/dynamically-assigning-videos-player
 418
 419         entries = []
 420
 421         # Look for iframe embeds [1]
 422         for _, url in re.findall(
 423                 r'<iframe[^>]+src=(["\'])((?:https?:)?//players\.brightcove\.net/\d+/[^/]+/index\.html.+?)\1', webpage):
 424             entries.append(url if url.startswith('http') else 'http:' + url)
 425
 426         # Look for <video> tags [2] and embed_in_page embeds [3]
 427         # [2] looks like:
 428         for video, script_tag, account_id, player_id, embed in re.findall(
 429                 r'''(?isx)
 430                     (<video(?:-js)?\s+[^>]*\bdata-video-id\s*=\s*['"]?[^>]+>)
 431                     (?:.*?
 432                         (<script[^>]+
 433                             src=["\'](?:https?:)?//players\.brightcove\.net/
 434                             (\d+)/([^/]+)_([^/]+)/index(?:\.min)?\.js
 435                         )
 436                     )?
 437                 ''', webpage):
 438             attrs = extract_attributes(video)
 439
 440             # According to examples from [4] it's unclear whether video id
 441             # may be optional and what to do when it is
 442             video_id = attrs.get('data-video-id')
 443             if not video_id:
 444                 continue
 445
 446             account_id = account_id or attrs.get('data-account')
 447             if not account_id:
 448                 continue
 449
 450             player_id = player_id or attrs.get('data-player') or 'default'
 451             embed = embed or attrs.get('data-embed') or 'default'
 452
 453             bc_url = 'http://players.brightcove.net/%s/%s_%s/index.html?videoId=%s' % (
 454                 account_id, player_id, embed, video_id)
 455
 456             # Some brightcove videos may be embedded with video tag only and
 457             # without script tag or any mentioning of brightcove at all. Such
 458             # embeds are considered ambiguous since they are matched based only
 459             # on data-video-id and data-account attributes and in the wild may
 460             # not be brightcove embeds at all. Let's check reconstructed
 461             # brightcove URLs in case of such embeds and only process valid
 462             # ones. By this we ensure there is indeed a brightcove embed.
 463             if not script_tag and not ie._is_valid_url(
 464                     bc_url, video_id, 'possible brightcove video'):
 465                 continue
 466
 467             entries.append(bc_url)
 468
 469         return entries
 470
 471     def _parse_brightcove_metadata(self, json_data, video_id, headers={}):
 472         title = json_data['name'].strip()
 473
 474         formats = []
 475         for source in json_data.get('sources', []):
 476             container = source.get('container')
 477             ext = mimetype2ext(source.get('type'))
 478             src = source.get('src')
 479             # https://support.brightcove.com/playback-api-video-fields-reference#key_systems_object
 480             if ext == 'ism' or container == 'WVM' or source.get('key_systems'):
 481                 continue
 482             elif ext == 'm3u8' or container == 'M2TS':
 483                 if not src:
 484                     continue
 485                 formats.extend(self._extract_m3u8_formats(
 486                     src, video_id, 'mp4', 'm3u8_native', m3u8_id='hls', fatal=False))
 487             elif ext == 'mpd':
 488                 if not src:
 489                     continue
 490                 formats.extend(self._extract_mpd_formats(src, video_id, 'dash', fatal=False))
 491             else:
 492                 streaming_src = source.get('streaming_src')
 493                 stream_name, app_name = source.get('stream_name'), source.get('app_name')
 494                 if not src and not streaming_src and (not stream_name or not app_name):
 495                     continue
 496                 tbr = float_or_none(source.get('avg_bitrate'), 1000)
 497                 height = int_or_none(source.get('height'))
 498                 width = int_or_none(source.get('width'))
 499                 f = {
 500                     'tbr': tbr,
 501                     'filesize': int_or_none(source.get('size')),
 502                     'container': container,
 503                     'ext': ext or container.lower(),
 504                 }
 505                 if width == 0 and height == 0:
 506                     f.update({
 507                         'vcodec': 'none',
 508                     })
 509                 else:
 510                     f.update({
 511                         'width': width,
 512                         'height': height,
 513                         'vcodec': source.get('codec'),
 514                     })
 515
 516                 def build_format_id(kind):
 517                     format_id = kind
 518                     if tbr:
 519                         format_id += '-%dk' % int(tbr)
 520                     if height:
 521                         format_id += '-%dp' % height
 522                     return format_id
 523
 524                 if src or streaming_src:
 525                     f.update({
 526                         'url': src or streaming_src,
 527                         'format_id': build_format_id('http' if src else 'http-streaming'),
 528                         'source_preference': 0 if src else -1,
 529                     })
 530                 else:
 531                     f.update({
 532                         'url': app_name,
 533                         'play_path': stream_name,
 534                         'format_id': build_format_id('rtmp'),
 535                     })
 536                 formats.append(f)
 537         if not formats:
 538             # for sonyliv.com DRM protected videos
 539             s3_source_url = json_data.get('custom_fields', {}).get('s3sourceurl')
 540             if s3_source_url:
 541                 formats.append({
 542                     'url': s3_source_url,
 543                     'format_id': 'source',
 544                 })
 545
 546         errors = json_data.get('errors')
 547         if not formats and errors:
 548             error = errors[0]
 549             raise ExtractorError(
 550                 error.get('message') or error.get('error_subcode') or error['error_code'], expected=True)
 551
 552         self._sort_formats(formats)
 553
 554         for f in formats:
 555             f.setdefault('http_headers', {}).update(headers)
 556
 557         subtitles = {}
 558         for text_track in json_data.get('text_tracks', []):
 559             if text_track.get('kind') != 'captions':
 560                 continue
 561             text_track_url = url_or_none(text_track.get('src'))
 562             if not text_track_url:
 563                 continue
 564             lang = (str_or_none(text_track.get('srclang'))
 565                     or str_or_none(text_track.get('label')) or 'en').lower()
 566             subtitles.setdefault(lang, []).append({
 567                 'url': text_track_url,
 568             })
 569
 570         is_live = False
 571         duration = float_or_none(json_data.get('duration'), 1000)
 572         if duration is not None and duration <= 0:
 573             is_live = True
 574
 575         return {
 576             'id': video_id,
 577             'title': self._live_title(title) if is_live else title,
 578             'description': clean_html(json_data.get('description')),
 579             'thumbnail': json_data.get('thumbnail') or json_data.get('poster'),
 580             'duration': duration,
 581             'timestamp': parse_iso8601(json_data.get('published_at')),
 582             'uploader_id': json_data.get('account_id'),
 583             'formats': formats,
 584             'subtitles': subtitles,
 585             'tags': json_data.get('tags', []),
 586             'is_live': is_live,
 587         }
 588
 589     def _real_extract(self, url):
 590         url, smuggled_data = unsmuggle_url(url, {})
 591         self._initialize_geo_bypass({
 592             'countries': smuggled_data.get('geo_countries'),
 593             'ip_blocks': smuggled_data.get('geo_ip_blocks'),
 594         })
 595
 596         account_id, player_id, embed, content_type, video_id = re.match(self._VALID_URL, url).groups()
 597
 598         policy_key_id = '%s_%s' % (account_id, player_id)
 599         policy_key = self._downloader.cache.load('brightcove', policy_key_id)
 600         policy_key_extracted = False
 601         store_pk = lambda x: self._downloader.cache.store('brightcove', policy_key_id, x)
 602
 603         def extract_policy_key():
 604             base_url = 'http://players.brightcove.net/%s/%s_%s/' % (account_id, player_id, embed)
 605             config = self._download_json(
 606                 base_url + 'config.json', video_id, fatal=False) or {}
 607             policy_key = try_get(
 608                 config, lambda x: x['video_cloud']['policy_key'])
 609             if not policy_key:
 610                 webpage = self._download_webpage(
 611                     base_url + 'index.min.js', video_id)
 612
 613                 catalog = self._search_regex(
 614                     r'catalog\(({.+?})\);', webpage, 'catalog', default=None)
 615                 if catalog:
 616                     catalog = self._parse_json(
 617                         js_to_json(catalog), video_id, fatal=False)
 618                     if catalog:
 619                         policy_key = catalog.get('policyKey')
 620
 621                 if not policy_key:
 622                     policy_key = self._search_regex(
 623                         r'policyKey\s*:\s*(["\'])(?P<pk>.+?)\1',
 624                         webpage, 'policy key', group='pk')
 625
 626             store_pk(policy_key)
 627             return policy_key
 628
 629         api_url = 'https://edge.api.brightcove.com/playback/v1/accounts/%s/%ss/%s' % (account_id, content_type, video_id)
 630         headers = {}
 631         referrer = smuggled_data.get('referrer')
 632         if referrer:
 633             headers.update({
 634                 'Referer': referrer,
 635                 'Origin': re.search(r'https?://[^/]+', referrer).group(0),
 636             })
 637
 638         for _ in range(2):
 639             if not policy_key:
 640                 policy_key = extract_policy_key()
 641                 policy_key_extracted = True
 642             headers['Accept'] = 'application/json;pk=%s' % policy_key
 643             try:
 644                 json_data = self._download_json(api_url, video_id, headers=headers)
 645                 break
 646             except ExtractorError as e:
 647                 if isinstance(e.cause, compat_HTTPError) and e.cause.code in (401, 403):
 648                     json_data = self._parse_json(e.cause.read().decode(), video_id)[0]
 649                     message = json_data.get('message') or json_data['error_code']
 650                     if json_data.get('error_subcode') == 'CLIENT_GEO':
 651                         self.raise_geo_restricted(msg=message)
 652                     elif json_data.get('error_code') == 'INVALID_POLICY_KEY' and not policy_key_extracted:
 653                         policy_key = None
 654                         store_pk(None)
 655                         continue
 656                     raise ExtractorError(message, expected=True)
 657                 raise
 658
 659         errors = json_data.get('errors')
 660         if errors and errors[0].get('error_subcode') == 'TVE_AUTH':
 661             custom_fields = json_data['custom_fields']
 662             tve_token = self._extract_mvpd_auth(
 663                 smuggled_data['source_url'], video_id,
 664                 custom_fields['bcadobepassrequestorid'],
 665                 custom_fields['bcadobepassresourceid'])
 666             json_data = self._download_json(
 667                 api_url, video_id, headers={
 668                     'Accept': 'application/json;pk=%s' % policy_key
 669                 }, query={
 670                     'tveToken': tve_token,
 671                 })
 672
 673         if content_type == 'playlist':
 674             return self.playlist_result(
 675                 [self._parse_brightcove_metadata(vid, vid.get('id'), headers)
 676                  for vid in json_data.get('videos', []) if vid.get('id')],
 677                 json_data.get('id'), json_data.get('name'),
 678                 json_data.get('description'))
 679
 680         return self._parse_brightcove_metadata(
 681             json_data, video_id, headers=headers)