Raphaël G. Git Repositories - youtubedl/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     unescapeHTML,
  29     unsmuggle_url,
  30     update_url_query,
  31     clean_html,
  32     mimetype2ext,
  33     UnsupportedError,
  34 )
  35
  36
  37 class BrightcoveLegacyIE(InfoExtractor):
  38     IE_NAME = 'brightcove:legacy'
  39     _VALID_URL = r'(?:https?://.*brightcove\.com/(services|viewer).*?\?|brightcove:)(?P<query>.*)'
  40
  41     _TESTS = [
  42         {
  43             # From http://www.8tv.cat/8aldia/videos/xavier-sala-i-martin-aquesta-tarda-a-8-al-dia/
  44             'url': 'http://c.brightcove.com/services/viewer/htmlFederated?playerID=1654948606001&flashID=myExperience&%40videoPlayer=2371591881001',
  45             'md5': '5423e113865d26e40624dce2e4b45d95',
  46             'note': 'Test Brightcove downloads and detection in GenericIE',
  47             'info_dict': {
  48                 'id': '2371591881001',
  49                 'ext': 'mp4',
  50                 'title': 'Xavier Sala i Martín: “Un banc que no presta és un banc zombi que no serveix per a res”',
  51                 'uploader': '8TV',
  52                 'description': 'md5:a950cc4285c43e44d763d036710cd9cd',
  53                 'timestamp': 1368213670,
  54                 'upload_date': '20130510',
  55                 'uploader_id': '1589608506001',
  56             },
  57             'skip': 'The player has been deactivated by the content owner',
  58         },
  59         {
  60             # From http://medianetwork.oracle.com/video/player/1785452137001
  61             'url': 'http://c.brightcove.com/services/viewer/htmlFederated?playerID=1217746023001&flashID=myPlayer&%40videoPlayer=1785452137001',
  62             'info_dict': {
  63                 'id': '1785452137001',
  64                 'ext': 'flv',
  65                 'title': 'JVMLS 2012: Arrays 2.0 - Opportunities and Challenges',
  66                 'description': 'John Rose speaks at the JVM Language Summit, August 1, 2012.',
  67                 'uploader': 'Oracle',
  68                 'timestamp': 1344975024,
  69                 'upload_date': '20120814',
  70                 'uploader_id': '1460825906',
  71             },
  72             'skip': 'video not playable',
  73         },
  74         {
  75             # From http://mashable.com/2013/10/26/thermoelectric-bracelet-lets-you-control-your-body-temperature/
  76             'url': 'http://c.brightcove.com/services/viewer/federated_f9?&playerID=1265504713001&publisherID=AQ%7E%7E%2CAAABBzUwv1E%7E%2CxP-xFHVUstiMFlNYfvF4G9yFnNaqCw_9&videoID=2750934548001',
  77             'info_dict': {
  78                 'id': '2750934548001',
  79                 'ext': 'mp4',
  80                 'title': 'This Bracelet Acts as a Personal Thermostat',
  81                 'description': 'md5:547b78c64f4112766ccf4e151c20b6a0',
  82                 # 'uploader': 'Mashable',
  83                 'timestamp': 1382041798,
  84                 'upload_date': '20131017',
  85                 'uploader_id': '1130468786001',
  86             },
  87         },
  88         {
  89             # test that the default referer works
  90             # from http://national.ballet.ca/interact/video/Lost_in_Motion_II/
  91             'url': 'http://link.brightcove.com/services/player/bcpid756015033001?bckey=AQ~~,AAAApYJi_Ck~,GxhXCegT1Dp39ilhXuxMJxasUhVNZiil&bctid=2878862109001',
  92             'info_dict': {
  93                 'id': '2878862109001',
  94                 'ext': 'mp4',
  95                 'title': 'Lost in Motion II',
  96                 'description': 'md5:363109c02998fee92ec02211bd8000df',
  97                 'uploader': 'National Ballet of Canada',
  98             },
  99             'skip': 'Video gone',
 100         },
 101         {
 102             # test flv videos served by akamaihd.net
 103             # From http://www.redbull.com/en/bike/stories/1331655643987/replay-uci-dh-world-cup-2014-from-fort-william
 104             'url': 'http://c.brightcove.com/services/viewer/htmlFederated?%40videoPlayer=ref%3Aevent-stream-356&linkBaseURL=http%3A%2F%2Fwww.redbull.com%2Fen%2Fbike%2Fvideos%2F1331655630249%2Freplay-uci-fort-william-2014-dh&playerKey=AQ%7E%7E%2CAAAApYJ7UqE%7E%2Cxqr_zXk0I-zzNndy8NlHogrCb5QdyZRf&playerID=1398061561001#__youtubedl_smuggle=%7B%22Referer%22%3A+%22http%3A%2F%2Fwww.redbull.com%2Fen%2Fbike%2Fstories%2F1331655643987%2Freplay-uci-dh-world-cup-2014-from-fort-william%22%7D',
 105             # The md5 checksum changes on each download
 106             'info_dict': {
 107                 'id': '3750436379001',
 108                 'ext': 'flv',
 109                 'title': 'UCI MTB World Cup 2014: Fort William, UK - Downhill Finals',
 110                 'uploader': 'RBTV Old (do not use)',
 111                 'description': 'UCI MTB World Cup 2014: Fort William, UK - Downhill Finals',
 112                 'timestamp': 1409122195,
 113                 'upload_date': '20140827',
 114                 'uploader_id': '710858724001',
 115             },
 116             'skip': 'Video gone',
 117         },
 118         {
 119             # playlist with 'videoList'
 120             # from http://support.brightcove.com/en/video-cloud/docs/playlist-support-single-video-players
 121             'url': 'http://c.brightcove.com/services/viewer/htmlFederated?playerID=3550052898001&playerKey=AQ%7E%7E%2CAAABmA9XpXk%7E%2C-Kp7jNgisre1fG5OdqpAFUTcs0lP_ZoL',
 122             'info_dict': {
 123                 'title': 'Sealife',
 124                 'id': '3550319591001',
 125             },
 126             'playlist_mincount': 7,
 127             'skip': 'Unsupported URL',
 128         },
 129         {
 130             # playlist with 'playlistTab' (https://github.com/ytdl-org/youtube-dl/issues/9965)
 131             'url': 'http://c.brightcove.com/services/json/experience/runtime/?command=get_programming_for_experience&playerKey=AQ%7E%7E,AAABXlLMdok%7E,NJ4EoMlZ4rZdx9eU1rkMVd8EaYPBBUlg',
 132             'info_dict': {
 133                 'id': '1522758701001',
 134                 'title': 'Lesson 08',
 135             },
 136             'playlist_mincount': 10,
 137             'skip': 'Unsupported URL',
 138         },
 139         {
 140             # playerID inferred from bcpid
 141             # from http://www.un.org/chinese/News/story.asp?NewsID=27724
 142             'url': 'https://link.brightcove.com/services/player/bcpid1722935254001/?bctid=5360463607001&autoStart=false&secureConnections=true&width=650&height=350',
 143             'only_matching': True,  # Tested in GenericIE
 144         }
 145     ]
 146
 147     @classmethod
 148     def _build_brighcove_url(cls, object_str):
 149         """
 150         Build a Brightcove url from a xml string containing
 151         <object class="BrightcoveExperience">{params}</object>
 152         """
 153
 154         # Fix up some stupid HTML, see https://github.com/ytdl-org/youtube-dl/issues/1553
 155         object_str = re.sub(r'(<param(?:\s+[a-zA-Z0-9_]+="[^"]*")*)>',
 156                             lambda m: m.group(1) + '/>', object_str)
 157         # Fix up some stupid XML, see https://github.com/ytdl-org/youtube-dl/issues/1608
 158         object_str = object_str.replace('<--', '<!--')
 159         # remove namespace to simplify extraction
 160         object_str = re.sub(r'(<object[^>]*)(xmlns=".*?")', r'\1', object_str)
 161         object_str = fix_xml_ampersands(object_str)
 162
 163         try:
 164             object_doc = compat_etree_fromstring(object_str.encode('utf-8'))
 165         except compat_xml_parse_error:
 166             return
 167
 168         fv_el = find_xpath_attr(object_doc, './param', 'name', 'flashVars')
 169         if fv_el is not None:
 170             flashvars = dict(
 171                 (k, v[0])
 172                 for k, v in compat_parse_qs(fv_el.attrib['value']).items())
 173         else:
 174             flashvars = {}
 175
 176         data_url = object_doc.attrib.get('data', '')
 177         data_url_params = compat_parse_qs(compat_urllib_parse_urlparse(data_url).query)
 178
 179         def find_param(name):
 180             if name in flashvars:
 181                 return flashvars[name]
 182             node = find_xpath_attr(object_doc, './param', 'name', name)
 183             if node is not None:
 184                 return node.attrib['value']
 185             return data_url_params.get(name)
 186
 187         params = {}
 188
 189         playerID = find_param('playerID') or find_param('playerId')
 190         if playerID is None:
 191             raise ExtractorError('Cannot find player ID')
 192         params['playerID'] = playerID
 193
 194         playerKey = find_param('playerKey')
 195         # Not all pages define this value
 196         if playerKey is not None:
 197             params['playerKey'] = playerKey
 198         # These fields hold the id of the video
 199         videoPlayer = find_param('@videoPlayer') or find_param('videoId') or find_param('videoID') or find_param('@videoList')
 200         if videoPlayer is not None:
 201             if isinstance(videoPlayer, list):
 202                 videoPlayer = videoPlayer[0]
 203             videoPlayer = videoPlayer.strip()
 204             # UUID is also possible for videoPlayer (e.g.
 205             # http://www.popcornflix.com/hoodies-vs-hooligans/7f2d2b87-bbf2-4623-acfb-ea942b4f01dd
 206             # or http://www8.hp.com/cn/zh/home.html)
 207             if not (re.match(
 208                     r'^(?:\d+|[\da-fA-F]{8}-?[\da-fA-F]{4}-?[\da-fA-F]{4}-?[\da-fA-F]{4}-?[\da-fA-F]{12})$',
 209                     videoPlayer) or videoPlayer.startswith('ref:')):
 210                 return None
 211             params['@videoPlayer'] = videoPlayer
 212         linkBase = find_param('linkBaseURL')
 213         if linkBase is not None:
 214             params['linkBaseURL'] = linkBase
 215         return cls._make_brightcove_url(params)
 216
 217     @classmethod
 218     def _build_brighcove_url_from_js(cls, object_js):
 219         # The layout of JS is as follows:
 220         # customBC.createVideo = function (width, height, playerID, playerKey, videoPlayer, VideoRandomID) {
 221         #   // build Brightcove <object /> XML
 222         # }
 223         m = re.search(
 224             r'''(?x)customBC\.createVideo\(
 225                 .*?                                                  # skipping width and height
 226                 ["\'](?P<playerID>\d+)["\']\s*,\s*                   # playerID
 227                 ["\'](?P<playerKey>AQ[^"\']{48})[^"\']*["\']\s*,\s*  # playerKey begins with AQ and is 50 characters
 228                                                                      # in length, however it's appended to itself
 229                                                                      # in places, so truncate
 230                 ["\'](?P<videoID>\d+)["\']                           # @videoPlayer
 231             ''', object_js)
 232         if m:
 233             return cls._make_brightcove_url(m.groupdict())
 234
 235     @classmethod
 236     def _make_brightcove_url(cls, params):
 237         return update_url_query(
 238             'http://c.brightcove.com/services/viewer/htmlFederated', params)
 239
 240     @classmethod
 241     def _extract_brightcove_url(cls, webpage):
 242         """Try to extract the brightcove url from the webpage, returns None
 243         if it can't be found
 244         """
 245         urls = cls._extract_brightcove_urls(webpage)
 246         return urls[0] if urls else None
 247
 248     @classmethod
 249     def _extract_brightcove_urls(cls, webpage):
 250         """Return a list of all Brightcove URLs from the webpage """
 251
 252         url_m = re.search(
 253             r'''(?x)
 254                 <meta\s+
 255                     (?:property|itemprop)=([\'"])(?:og:video|embedURL)\1[^>]+
 256                     content=([\'"])(?P<url>https?://(?:secure|c)\.brightcove.com/(?:(?!\2).)+)\2
 257             ''', webpage)
 258         if url_m:
 259             url = unescapeHTML(url_m.group('url'))
 260             # Some sites don't add it, we can't download with this url, for example:
 261             # http://www.ktvu.com/videos/news/raw-video-caltrain-releases-video-of-man-almost/vCTZdY/
 262             if 'playerKey' in url or 'videoId' in url or 'idVideo' in url:
 263                 return [url]
 264
 265         matches = re.findall(
 266             r'''(?sx)<object
 267             (?:
 268                 [^>]+?class=[\'"][^>]*?BrightcoveExperience.*?[\'"] |
 269                 [^>]*?>\s*<param\s+name="movie"\s+value="https?://[^/]*brightcove\.com/
 270             ).+?>\s*</object>''',
 271             webpage)
 272         if matches:
 273             return list(filter(None, [cls._build_brighcove_url(m) for m in matches]))
 274
 275         matches = re.findall(r'(customBC\.createVideo\(.+?\);)', webpage)
 276         if matches:
 277             return list(filter(None, [
 278                 cls._build_brighcove_url_from_js(custom_bc)
 279                 for custom_bc in matches]))
 280         return [src for _, src in re.findall(
 281             r'<iframe[^>]+src=([\'"])((?:https?:)?//link\.brightcove\.com/services/player/(?!\1).+)\1', webpage)]
 282
 283     def _real_extract(self, url):
 284         url, smuggled_data = unsmuggle_url(url, {})
 285
 286         # Change the 'videoId' and others field to '@videoPlayer'
 287         url = re.sub(r'(?<=[?&])(videoI(d|D)|idVideo|bctid)', '%40videoPlayer', url)
 288         # Change bckey (used by bcove.me urls) to playerKey
 289         url = re.sub(r'(?<=[?&])bckey', 'playerKey', url)
 290         mobj = re.match(self._VALID_URL, url)
 291         query_str = mobj.group('query')
 292         query = compat_urlparse.parse_qs(query_str)
 293
 294         videoPlayer = query.get('@videoPlayer')
 295         if videoPlayer:
 296             # We set the original url as the default 'Referer' header
 297             referer = query.get('linkBaseURL', [None])[0] or smuggled_data.get('Referer', url)
 298             video_id = videoPlayer[0]
 299             if 'playerID' not in query:
 300                 mobj = re.search(r'/bcpid(\d+)', url)
 301                 if mobj is not None:
 302                     query['playerID'] = [mobj.group(1)]
 303             publisher_id = query.get('publisherId')
 304             if publisher_id and publisher_id[0].isdigit():
 305                 publisher_id = publisher_id[0]
 306             if not publisher_id:
 307                 player_key = query.get('playerKey')
 308                 if player_key and ',' in player_key[0]:
 309                     player_key = player_key[0]
 310                 else:
 311                     player_id = query.get('playerID')
 312                     if player_id and player_id[0].isdigit():
 313                         headers = {}
 314                         if referer:
 315                             headers['Referer'] = referer
 316                         player_page = self._download_webpage(
 317                             'http://link.brightcove.com/services/player/bcpid' + player_id[0],
 318                             video_id, headers=headers, fatal=False)
 319                         if player_page:
 320                             player_key = self._search_regex(
 321                                 r'<param\s+name="playerKey"\s+value="([\w~,-]+)"',
 322                                 player_page, 'player key', fatal=False)
 323                 if player_key:
 324                     enc_pub_id = player_key.split(',')[1].replace('~', '=')
 325                     publisher_id = struct.unpack('>Q', base64.urlsafe_b64decode(enc_pub_id))[0]
 326             if publisher_id:
 327                 brightcove_new_url = 'http://players.brightcove.net/%s/default_default/index.html?videoId=%s' % (publisher_id, video_id)
 328                 if referer:
 329                     brightcove_new_url = smuggle_url(brightcove_new_url, {'referrer': referer})
 330                 return self.url_result(brightcove_new_url, BrightcoveNewIE.ie_key(), video_id)
 331         # TODO: figure out if it's possible to extract playlistId from playerKey
 332         # elif 'playerKey' in query:
 333         #     player_key = query['playerKey']
 334         #     return self._get_playlist_info(player_key[0])
 335         raise UnsupportedError(url)
 336
 337
 338 class BrightcoveNewIE(AdobePassIE):
 339     IE_NAME = 'brightcove:new'
 340     _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:[^&]+)'
 341     _TESTS = [{
 342         'url': 'http://players.brightcove.net/929656772001/e41d32dc-ec74-459e-a845-6c69f7b724ea_default/index.html?videoId=4463358922001',
 343         'md5': 'c8100925723840d4b0d243f7025703be',
 344         'info_dict': {
 345             'id': '4463358922001',
 346             'ext': 'mp4',
 347             'title': 'Meet the man behind Popcorn Time',
 348             'description': 'md5:eac376a4fe366edc70279bfb681aea16',
 349             'duration': 165.768,
 350             'timestamp': 1441391203,
 351             'upload_date': '20150904',
 352             'uploader_id': '929656772001',
 353             'formats': 'mincount:20',
 354         },
 355     }, {
 356         # with rtmp streams
 357         'url': 'http://players.brightcove.net/4036320279001/5d112ed9-283f-485f-a7f9-33f42e8bc042_default/index.html?videoId=4279049078001',
 358         'info_dict': {
 359             'id': '4279049078001',
 360             'ext': 'mp4',
 361             'title': 'Titansgrave: Chapter 0',
 362             'description': 'Titansgrave: Chapter 0',
 363             'duration': 1242.058,
 364             'timestamp': 1433556729,
 365             'upload_date': '20150606',
 366             'uploader_id': '4036320279001',
 367             'formats': 'mincount:39',
 368         },
 369         'params': {
 370             # m3u8 download
 371             'skip_download': True,
 372         }
 373     }, {
 374         # playlist stream
 375         'url': 'https://players.brightcove.net/1752604059001/S13cJdUBz_default/index.html?playlistId=5718313430001',
 376         'info_dict': {
 377             'id': '5718313430001',
 378             'title': 'No Audio Playlist',
 379         },
 380         'playlist_count': 7,
 381         'params': {
 382             # m3u8 download
 383             'skip_download': True,
 384         }
 385     }, {
 386         'url': 'http://players.brightcove.net/5690807595001/HyZNerRl7_default/index.html?playlistId=5743160747001',
 387         'only_matching': True,
 388     }, {
 389         # ref: prefixed video id
 390         'url': 'http://players.brightcove.net/3910869709001/21519b5c-4b3b-4363-accb-bdc8f358f823_default/index.html?videoId=ref:7069442',
 391         'only_matching': True,
 392     }, {
 393         # non numeric ref: prefixed video id
 394         'url': 'http://players.brightcove.net/710858724001/default_default/index.html?videoId=ref:event-stream-356',
 395         'only_matching': True,
 396     }, {
 397         # unavailable video without message but with error_code
 398         'url': 'http://players.brightcove.net/1305187701/c832abfb-641b-44eb-9da0-2fe76786505f_default/index.html?videoId=4377407326001',
 399         'only_matching': True,
 400     }]
 401
 402     @staticmethod
 403     def _extract_url(ie, webpage):
 404         urls = BrightcoveNewIE._extract_urls(ie, webpage)
 405         return urls[0] if urls else None
 406
 407     @staticmethod
 408     def _extract_urls(ie, webpage):
 409         # Reference:
 410         # 1. http://docs.brightcove.com/en/video-cloud/brightcove-player/guides/publish-video.html#setvideoiniframe
 411         # 2. http://docs.brightcove.com/en/video-cloud/brightcove-player/guides/publish-video.html#tag
 412         # 3. http://docs.brightcove.com/en/video-cloud/brightcove-player/guides/publish-video.html#setvideousingjavascript
 413         # 4. http://docs.brightcove.com/en/video-cloud/brightcove-player/guides/in-page-embed-player-implementation.html
 414         # 5. https://support.brightcove.com/en/video-cloud/docs/dynamically-assigning-videos-player
 415
 416         entries = []
 417
 418         # Look for iframe embeds [1]
 419         for _, url in re.findall(
 420                 r'<iframe[^>]+src=(["\'])((?:https?:)?//players\.brightcove\.net/\d+/[^/]+/index\.html.+?)\1', webpage):
 421             entries.append(url if url.startswith('http') else 'http:' + url)
 422
 423         # Look for <video> tags [2] and embed_in_page embeds [3]
 424         # [2] looks like:
 425         for video, script_tag, account_id, player_id, embed in re.findall(
 426                 r'''(?isx)
 427                     (<video\s+[^>]*\bdata-video-id\s*=\s*['"]?[^>]+>)
 428                     (?:.*?
 429                         (<script[^>]+
 430                             src=["\'](?:https?:)?//players\.brightcove\.net/
 431                             (\d+)/([^/]+)_([^/]+)/index(?:\.min)?\.js
 432                         )
 433                     )?
 434                 ''', webpage):
 435             attrs = extract_attributes(video)
 436
 437             # According to examples from [4] it's unclear whether video id
 438             # may be optional and what to do when it is
 439             video_id = attrs.get('data-video-id')
 440             if not video_id:
 441                 continue
 442
 443             account_id = account_id or attrs.get('data-account')
 444             if not account_id:
 445                 continue
 446
 447             player_id = player_id or attrs.get('data-player') or 'default'
 448             embed = embed or attrs.get('data-embed') or 'default'
 449
 450             bc_url = 'http://players.brightcove.net/%s/%s_%s/index.html?videoId=%s' % (
 451                 account_id, player_id, embed, video_id)
 452
 453             # Some brightcove videos may be embedded with video tag only and
 454             # without script tag or any mentioning of brightcove at all. Such
 455             # embeds are considered ambiguous since they are matched based only
 456             # on data-video-id and data-account attributes and in the wild may
 457             # not be brightcove embeds at all. Let's check reconstructed
 458             # brightcove URLs in case of such embeds and only process valid
 459             # ones. By this we ensure there is indeed a brightcove embed.
 460             if not script_tag and not ie._is_valid_url(
 461                     bc_url, video_id, 'possible brightcove video'):
 462                 continue
 463
 464             entries.append(bc_url)
 465
 466         return entries
 467
 468     def _parse_brightcove_metadata(self, json_data, video_id, headers={}):
 469         title = json_data['name'].strip()
 470
 471         formats = []
 472         for source in json_data.get('sources', []):
 473             container = source.get('container')
 474             ext = mimetype2ext(source.get('type'))
 475             src = source.get('src')
 476             # https://support.brightcove.com/playback-api-video-fields-reference#key_systems_object
 477             if ext == 'ism' or container == 'WVM' or source.get('key_systems'):
 478                 continue
 479             elif ext == 'm3u8' or container == 'M2TS':
 480                 if not src:
 481                     continue
 482                 formats.extend(self._extract_m3u8_formats(
 483                     src, video_id, 'mp4', 'm3u8_native', m3u8_id='hls', fatal=False))
 484             elif ext == 'mpd':
 485                 if not src:
 486                     continue
 487                 formats.extend(self._extract_mpd_formats(src, video_id, 'dash', fatal=False))
 488             else:
 489                 streaming_src = source.get('streaming_src')
 490                 stream_name, app_name = source.get('stream_name'), source.get('app_name')
 491                 if not src and not streaming_src and (not stream_name or not app_name):
 492                     continue
 493                 tbr = float_or_none(source.get('avg_bitrate'), 1000)
 494                 height = int_or_none(source.get('height'))
 495                 width = int_or_none(source.get('width'))
 496                 f = {
 497                     'tbr': tbr,
 498                     'filesize': int_or_none(source.get('size')),
 499                     'container': container,
 500                     'ext': ext or container.lower(),
 501                 }
 502                 if width == 0 and height == 0:
 503                     f.update({
 504                         'vcodec': 'none',
 505                     })
 506                 else:
 507                     f.update({
 508                         'width': width,
 509                         'height': height,
 510                         'vcodec': source.get('codec'),
 511                     })
 512
 513                 def build_format_id(kind):
 514                     format_id = kind
 515                     if tbr:
 516                         format_id += '-%dk' % int(tbr)
 517                     if height:
 518                         format_id += '-%dp' % height
 519                     return format_id
 520
 521                 if src or streaming_src:
 522                     f.update({
 523                         'url': src or streaming_src,
 524                         'format_id': build_format_id('http' if src else 'http-streaming'),
 525                         'source_preference': 0 if src else -1,
 526                     })
 527                 else:
 528                     f.update({
 529                         'url': app_name,
 530                         'play_path': stream_name,
 531                         'format_id': build_format_id('rtmp'),
 532                     })
 533                 formats.append(f)
 534         if not formats:
 535             # for sonyliv.com DRM protected videos
 536             s3_source_url = json_data.get('custom_fields', {}).get('s3sourceurl')
 537             if s3_source_url:
 538                 formats.append({
 539                     'url': s3_source_url,
 540                     'format_id': 'source',
 541                 })
 542
 543         errors = json_data.get('errors')
 544         if not formats and errors:
 545             error = errors[0]
 546             raise ExtractorError(
 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         subtitles = {}
 555         for text_track in json_data.get('text_tracks', []):
 556             if text_track.get('src'):
 557                 subtitles.setdefault(text_track.get('srclang'), []).append({
 558                     'url': text_track['src'],
 559                 })
 560
 561         is_live = False
 562         duration = float_or_none(json_data.get('duration'), 1000)
 563         if duration is not None and duration <= 0:
 564             is_live = True
 565
 566         return {
 567             'id': video_id,
 568             'title': self._live_title(title) if is_live else title,
 569             'description': clean_html(json_data.get('description')),
 570             'thumbnail': json_data.get('thumbnail') or json_data.get('poster'),
 571             'duration': duration,
 572             'timestamp': parse_iso8601(json_data.get('published_at')),
 573             'uploader_id': json_data.get('account_id'),
 574             'formats': formats,
 575             'subtitles': subtitles,
 576             'tags': json_data.get('tags', []),
 577             'is_live': is_live,
 578         }
 579
 580     def _real_extract(self, url):
 581         url, smuggled_data = unsmuggle_url(url, {})
 582         self._initialize_geo_bypass({
 583             'countries': smuggled_data.get('geo_countries'),
 584             'ip_blocks': smuggled_data.get('geo_ip_blocks'),
 585         })
 586
 587         account_id, player_id, embed, content_type, video_id = re.match(self._VALID_URL, url).groups()
 588
 589         policy_key_id = '%s_%s' % (account_id, player_id)
 590         policy_key = self._downloader.cache.load('brightcove', policy_key_id)
 591         policy_key_extracted = False
 592         store_pk = lambda x: self._downloader.cache.store('brightcove', policy_key_id, x)
 593
 594         def extract_policy_key():
 595             webpage = self._download_webpage(
 596                 'http://players.brightcove.net/%s/%s_%s/index.min.js'
 597                 % (account_id, player_id, embed), video_id)
 598
 599             policy_key = None
 600
 601             catalog = self._search_regex(
 602                 r'catalog\(({.+?})\);', webpage, 'catalog', default=None)
 603             if catalog:
 604                 catalog = self._parse_json(
 605                     js_to_json(catalog), video_id, fatal=False)
 606                 if catalog:
 607                     policy_key = catalog.get('policyKey')
 608
 609             if not policy_key:
 610                 policy_key = self._search_regex(
 611                     r'policyKey\s*:\s*(["\'])(?P<pk>.+?)\1',
 612                     webpage, 'policy key', group='pk')
 613
 614             store_pk(policy_key)
 615             return policy_key
 616
 617         api_url = 'https://edge.api.brightcove.com/playback/v1/accounts/%s/%ss/%s' % (account_id, content_type, video_id)
 618         headers = {}
 619         referrer = smuggled_data.get('referrer')
 620         if referrer:
 621             headers.update({
 622                 'Referer': referrer,
 623                 'Origin': re.search(r'https?://[^/]+', referrer).group(0),
 624             })
 625
 626         for _ in range(2):
 627             if not policy_key:
 628                 policy_key = extract_policy_key()
 629                 policy_key_extracted = True
 630             headers['Accept'] = 'application/json;pk=%s' % policy_key
 631             try:
 632                 json_data = self._download_json(api_url, video_id, headers=headers)
 633                 break
 634             except ExtractorError as e:
 635                 if isinstance(e.cause, compat_HTTPError) and e.cause.code in (401, 403):
 636                     json_data = self._parse_json(e.cause.read().decode(), video_id)[0]
 637                     message = json_data.get('message') or json_data['error_code']
 638                     if json_data.get('error_subcode') == 'CLIENT_GEO':
 639                         self.raise_geo_restricted(msg=message)
 640                     elif json_data.get('error_code') == 'INVALID_POLICY_KEY' and not policy_key_extracted:
 641                         policy_key = None
 642                         store_pk(None)
 643                         continue
 644                     raise ExtractorError(message, expected=True)
 645                 raise
 646
 647         errors = json_data.get('errors')
 648         if errors and errors[0].get('error_subcode') == 'TVE_AUTH':
 649             custom_fields = json_data['custom_fields']
 650             tve_token = self._extract_mvpd_auth(
 651                 smuggled_data['source_url'], video_id,
 652                 custom_fields['bcadobepassrequestorid'],
 653                 custom_fields['bcadobepassresourceid'])
 654             json_data = self._download_json(
 655                 api_url, video_id, headers={
 656                     'Accept': 'application/json;pk=%s' % policy_key
 657                 }, query={
 658                     'tveToken': tve_token,
 659                 })
 660
 661         if content_type == 'playlist':
 662             return self.playlist_result(
 663                 [self._parse_brightcove_metadata(vid, vid.get('id'), headers)
 664                  for vid in json_data.get('videos', []) if vid.get('id')],
 665                 json_data.get('id'), json_data.get('name'),
 666                 json_data.get('description'))
 667
 668         return self._parse_brightcove_metadata(
 669             json_data, video_id, headers=headers)