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