_ Git - youtube-dl/blob - youtube_dl/extractor/brightcove.py

   1 # encoding: utf-8
   2 from __future__ import unicode_literals
   3
   4 import re
   5 import json
   6
   7 from .common import InfoExtractor
   8 from ..compat import (
   9     compat_etree_fromstring,
  10     compat_parse_qs,
  11     compat_str,
  12     compat_urllib_parse_urlparse,
  13     compat_urlparse,
  14     compat_xml_parse_error,
  15     compat_HTTPError,
  16 )
  17 from ..utils import (
  18     determine_ext,
  19     ExtractorError,
  20     find_xpath_attr,
  21     fix_xml_ampersands,
  22     float_or_none,
  23     js_to_json,
  24     int_or_none,
  25     parse_iso8601,
  26     unescapeHTML,
  27     unsmuggle_url,
  28     update_url_query,
  29 )
  30
  31
  32 class BrightcoveLegacyIE(InfoExtractor):
  33     IE_NAME = 'brightcove:legacy'
  34     _VALID_URL = r'(?:https?://.*brightcove\.com/(services|viewer).*?\?|brightcove:)(?P<query>.*)'
  35     _FEDERATED_URL = 'http://c.brightcove.com/services/viewer/htmlFederated'
  36
  37     _TESTS = [
  38         {
  39             # From http://www.8tv.cat/8aldia/videos/xavier-sala-i-martin-aquesta-tarda-a-8-al-dia/
  40             'url': 'http://c.brightcove.com/services/viewer/htmlFederated?playerID=1654948606001&flashID=myExperience&%40videoPlayer=2371591881001',
  41             'md5': '5423e113865d26e40624dce2e4b45d95',
  42             'note': 'Test Brightcove downloads and detection in GenericIE',
  43             'info_dict': {
  44                 'id': '2371591881001',
  45                 'ext': 'mp4',
  46                 'title': 'Xavier Sala i Martín: “Un banc que no presta és un banc zombi que no serveix per a res”',
  47                 'uploader': '8TV',
  48                 'description': 'md5:a950cc4285c43e44d763d036710cd9cd',
  49                 'timestamp': 1368213670,
  50                 'upload_date': '20130510',
  51                 'uploader_id': '1589608506001',
  52             }
  53         },
  54         {
  55             # From http://medianetwork.oracle.com/video/player/1785452137001
  56             'url': 'http://c.brightcove.com/services/viewer/htmlFederated?playerID=1217746023001&flashID=myPlayer&%40videoPlayer=1785452137001',
  57             'info_dict': {
  58                 'id': '1785452137001',
  59                 'ext': 'flv',
  60                 'title': 'JVMLS 2012: Arrays 2.0 - Opportunities and Challenges',
  61                 'description': 'John Rose speaks at the JVM Language Summit, August 1, 2012.',
  62                 'uploader': 'Oracle',
  63                 'timestamp': 1344975024,
  64                 'upload_date': '20120814',
  65                 'uploader_id': '1460825906',
  66             },
  67         },
  68         {
  69             # From http://mashable.com/2013/10/26/thermoelectric-bracelet-lets-you-control-your-body-temperature/
  70             'url': 'http://c.brightcove.com/services/viewer/federated_f9?&playerID=1265504713001&publisherID=AQ%7E%7E%2CAAABBzUwv1E%7E%2CxP-xFHVUstiMFlNYfvF4G9yFnNaqCw_9&videoID=2750934548001',
  71             'info_dict': {
  72                 'id': '2750934548001',
  73                 'ext': 'mp4',
  74                 'title': 'This Bracelet Acts as a Personal Thermostat',
  75                 'description': 'md5:547b78c64f4112766ccf4e151c20b6a0',
  76                 'uploader': 'Mashable',
  77                 'timestamp': 1382041798,
  78                 'upload_date': '20131017',
  79                 'uploader_id': '1130468786001',
  80             },
  81         },
  82         {
  83             # test that the default referer works
  84             # from http://national.ballet.ca/interact/video/Lost_in_Motion_II/
  85             'url': 'http://link.brightcove.com/services/player/bcpid756015033001?bckey=AQ~~,AAAApYJi_Ck~,GxhXCegT1Dp39ilhXuxMJxasUhVNZiil&bctid=2878862109001',
  86             'info_dict': {
  87                 'id': '2878862109001',
  88                 'ext': 'mp4',
  89                 'title': 'Lost in Motion II',
  90                 'description': 'md5:363109c02998fee92ec02211bd8000df',
  91                 'uploader': 'National Ballet of Canada',
  92             },
  93             'skip': 'Video gone',
  94         },
  95         {
  96             # test flv videos served by akamaihd.net
  97             # From http://www.redbull.com/en/bike/stories/1331655643987/replay-uci-dh-world-cup-2014-from-fort-william
  98             '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',
  99             # The md5 checksum changes on each download
 100             'info_dict': {
 101                 'id': '3750436379001',
 102                 'ext': 'flv',
 103                 'title': 'UCI MTB World Cup 2014: Fort William, UK - Downhill Finals',
 104                 'uploader': 'RBTV Old (do not use)',
 105                 'description': 'UCI MTB World Cup 2014: Fort William, UK - Downhill Finals',
 106                 'timestamp': 1409122195,
 107                 'upload_date': '20140827',
 108                 'uploader_id': '710858724001',
 109             },
 110         },
 111         {
 112             # playlist with 'videoList'
 113             # from http://support.brightcove.com/en/video-cloud/docs/playlist-support-single-video-players
 114             'url': 'http://c.brightcove.com/services/viewer/htmlFederated?playerID=3550052898001&playerKey=AQ%7E%7E%2CAAABmA9XpXk%7E%2C-Kp7jNgisre1fG5OdqpAFUTcs0lP_ZoL',
 115             'info_dict': {
 116                 'title': 'Sealife',
 117                 'id': '3550319591001',
 118             },
 119             'playlist_mincount': 7,
 120         },
 121         {
 122             # playlist with 'playlistTab' (https://github.com/rg3/youtube-dl/issues/9965)
 123             'url': 'http://c.brightcove.com/services/json/experience/runtime/?command=get_programming_for_experience&playerKey=AQ%7E%7E,AAABXlLMdok%7E,NJ4EoMlZ4rZdx9eU1rkMVd8EaYPBBUlg',
 124             'info_dict': {
 125                 'id': '1522758701001',
 126                 'title': 'Lesson 08',
 127             },
 128             'playlist_mincount': 10,
 129         },
 130     ]
 131     FLV_VCODECS = {
 132         1: 'SORENSON',
 133         2: 'ON2',
 134         3: 'H264',
 135         4: 'VP8',
 136     }
 137
 138     @classmethod
 139     def _build_brighcove_url(cls, object_str):
 140         """
 141         Build a Brightcove url from a xml string containing
 142         <object class="BrightcoveExperience">{params}</object>
 143         """
 144
 145         # Fix up some stupid HTML, see https://github.com/rg3/youtube-dl/issues/1553
 146         object_str = re.sub(r'(<param(?:\s+[a-zA-Z0-9_]+="[^"]*")*)>',
 147                             lambda m: m.group(1) + '/>', object_str)
 148         # Fix up some stupid XML, see https://github.com/rg3/youtube-dl/issues/1608
 149         object_str = object_str.replace('<--', '<!--')
 150         # remove namespace to simplify extraction
 151         object_str = re.sub(r'(<object[^>]*)(xmlns=".*?")', r'\1', object_str)
 152         object_str = fix_xml_ampersands(object_str)
 153
 154         try:
 155             object_doc = compat_etree_fromstring(object_str.encode('utf-8'))
 156         except compat_xml_parse_error:
 157             return
 158
 159         fv_el = find_xpath_attr(object_doc, './param', 'name', 'flashVars')
 160         if fv_el is not None:
 161             flashvars = dict(
 162                 (k, v[0])
 163                 for k, v in compat_parse_qs(fv_el.attrib['value']).items())
 164         else:
 165             flashvars = {}
 166
 167         data_url = object_doc.attrib.get('data', '')
 168         data_url_params = compat_parse_qs(compat_urllib_parse_urlparse(data_url).query)
 169
 170         def find_param(name):
 171             if name in flashvars:
 172                 return flashvars[name]
 173             node = find_xpath_attr(object_doc, './param', 'name', name)
 174             if node is not None:
 175                 return node.attrib['value']
 176             return data_url_params.get(name)
 177
 178         params = {}
 179
 180         playerID = find_param('playerID')
 181         if playerID is None:
 182             raise ExtractorError('Cannot find player ID')
 183         params['playerID'] = playerID
 184
 185         playerKey = find_param('playerKey')
 186         # Not all pages define this value
 187         if playerKey is not None:
 188             params['playerKey'] = playerKey
 189         # These fields hold the id of the video
 190         videoPlayer = find_param('@videoPlayer') or find_param('videoId') or find_param('videoID') or find_param('@videoList')
 191         if videoPlayer is not None:
 192             params['@videoPlayer'] = videoPlayer
 193         linkBase = find_param('linkBaseURL')
 194         if linkBase is not None:
 195             params['linkBaseURL'] = linkBase
 196         return cls._make_brightcove_url(params)
 197
 198     @classmethod
 199     def _build_brighcove_url_from_js(cls, object_js):
 200         # The layout of JS is as follows:
 201         # customBC.createVideo = function (width, height, playerID, playerKey, videoPlayer, VideoRandomID) {
 202         #   // build Brightcove <object /> XML
 203         # }
 204         m = re.search(
 205             r'''(?x)customBC.\createVideo\(
 206                 .*?                                                  # skipping width and height
 207                 ["\'](?P<playerID>\d+)["\']\s*,\s*                   # playerID
 208                 ["\'](?P<playerKey>AQ[^"\']{48})[^"\']*["\']\s*,\s*  # playerKey begins with AQ and is 50 characters
 209                                                                      # in length, however it's appended to itself
 210                                                                      # in places, so truncate
 211                 ["\'](?P<videoID>\d+)["\']                           # @videoPlayer
 212             ''', object_js)
 213         if m:
 214             return cls._make_brightcove_url(m.groupdict())
 215
 216     @classmethod
 217     def _make_brightcove_url(cls, params):
 218         return update_url_query(cls._FEDERATED_URL, params)
 219
 220     @classmethod
 221     def _extract_brightcove_url(cls, webpage):
 222         """Try to extract the brightcove url from the webpage, returns None
 223         if it can't be found
 224         """
 225         urls = cls._extract_brightcove_urls(webpage)
 226         return urls[0] if urls else None
 227
 228     @classmethod
 229     def _extract_brightcove_urls(cls, webpage):
 230         """Return a list of all Brightcove URLs from the webpage """
 231
 232         url_m = re.search(
 233             r'<meta\s+property=[\'"]og:video[\'"]\s+content=[\'"](https?://(?:secure|c)\.brightcove.com/[^\'"]+)[\'"]',
 234             webpage)
 235         if url_m:
 236             url = unescapeHTML(url_m.group(1))
 237             # Some sites don't add it, we can't download with this url, for example:
 238             # http://www.ktvu.com/videos/news/raw-video-caltrain-releases-video-of-man-almost/vCTZdY/
 239             if 'playerKey' in url or 'videoId' in url:
 240                 return [url]
 241
 242         matches = re.findall(
 243             r'''(?sx)<object
 244             (?:
 245                 [^>]+?class=[\'"][^>]*?BrightcoveExperience.*?[\'"] |
 246                 [^>]*?>\s*<param\s+name="movie"\s+value="https?://[^/]*brightcove\.com/
 247             ).+?>\s*</object>''',
 248             webpage)
 249         if matches:
 250             return list(filter(None, [cls._build_brighcove_url(m) for m in matches]))
 251
 252         return list(filter(None, [
 253             cls._build_brighcove_url_from_js(custom_bc)
 254             for custom_bc in re.findall(r'(customBC\.createVideo\(.+?\);)', webpage)]))
 255
 256     def _real_extract(self, url):
 257         url, smuggled_data = unsmuggle_url(url, {})
 258
 259         # Change the 'videoId' and others field to '@videoPlayer'
 260         url = re.sub(r'(?<=[?&])(videoI(d|D)|bctid)', '%40videoPlayer', url)
 261         # Change bckey (used by bcove.me urls) to playerKey
 262         url = re.sub(r'(?<=[?&])bckey', 'playerKey', url)
 263         mobj = re.match(self._VALID_URL, url)
 264         query_str = mobj.group('query')
 265         query = compat_urlparse.parse_qs(query_str)
 266
 267         videoPlayer = query.get('@videoPlayer')
 268         if videoPlayer:
 269             # We set the original url as the default 'Referer' header
 270             referer = smuggled_data.get('Referer', url)
 271             return self._get_video_info(
 272                 videoPlayer[0], query, referer=referer)
 273         elif 'playerKey' in query:
 274             player_key = query['playerKey']
 275             return self._get_playlist_info(player_key[0])
 276         else:
 277             raise ExtractorError(
 278                 'Cannot find playerKey= variable. Did you forget quotes in a shell invocation?',
 279                 expected=True)
 280
 281     def _get_video_info(self, video_id, query, referer=None):
 282         headers = {}
 283         linkBase = query.get('linkBaseURL')
 284         if linkBase is not None:
 285             referer = linkBase[0]
 286         if referer is not None:
 287             headers['Referer'] = referer
 288         webpage = self._download_webpage(self._FEDERATED_URL, video_id, headers=headers, query=query)
 289
 290         error_msg = self._html_search_regex(
 291             r"<h1>We're sorry.</h1>([\s\n]*<p>.*?</p>)+", webpage,
 292             'error message', default=None)
 293         if error_msg is not None:
 294             raise ExtractorError(
 295                 'brightcove said: %s' % error_msg, expected=True)
 296
 297         self.report_extraction(video_id)
 298         info = self._search_regex(r'var experienceJSON = ({.*});', webpage, 'json')
 299         info = json.loads(info)['data']
 300         video_info = info['programmedContent']['videoPlayer']['mediaDTO']
 301         video_info['_youtubedl_adServerURL'] = info.get('adServerURL')
 302
 303         return self._extract_video_info(video_info)
 304
 305     def _get_playlist_info(self, player_key):
 306         info_url = 'http://c.brightcove.com/services/json/experience/runtime/?command=get_programming_for_experience&playerKey=%s' % player_key
 307         playlist_info = self._download_webpage(
 308             info_url, player_key, 'Downloading playlist information')
 309
 310         json_data = json.loads(playlist_info)
 311         if 'videoList' in json_data:
 312             playlist_info = json_data['videoList']
 313             playlist_dto = playlist_info['mediaCollectionDTO']
 314         elif 'playlistTabs' in json_data:
 315             playlist_info = json_data['playlistTabs']
 316             playlist_dto = playlist_info['lineupListDTO']['playlistDTOs'][0]
 317         else:
 318             raise ExtractorError('Empty playlist')
 319
 320         videos = [self._extract_video_info(video_info) for video_info in playlist_dto['videoDTOs']]
 321
 322         return self.playlist_result(videos, playlist_id='%s' % playlist_info['id'],
 323                                     playlist_title=playlist_dto['displayName'])
 324
 325     def _extract_video_info(self, video_info):
 326         video_id = compat_str(video_info['id'])
 327         publisher_id = video_info.get('publisherId')
 328         info = {
 329             'id': video_id,
 330             'title': video_info['displayName'].strip(),
 331             'description': video_info.get('shortDescription'),
 332             'thumbnail': video_info.get('videoStillURL') or video_info.get('thumbnailURL'),
 333             'uploader': video_info.get('publisherName'),
 334             'uploader_id': compat_str(publisher_id) if publisher_id else None,
 335             'duration': float_or_none(video_info.get('length'), 1000),
 336             'timestamp': int_or_none(video_info.get('creationDate'), 1000),
 337         }
 338
 339         renditions = video_info.get('renditions', []) + video_info.get('IOSRenditions', [])
 340         if renditions:
 341             formats = []
 342             for rend in renditions:
 343                 url = rend['defaultURL']
 344                 if not url:
 345                     continue
 346                 ext = None
 347                 if rend['remote']:
 348                     url_comp = compat_urllib_parse_urlparse(url)
 349                     if url_comp.path.endswith('.m3u8'):
 350                         formats.extend(
 351                             self._extract_m3u8_formats(
 352                                 url, video_id, 'mp4', 'm3u8_native', m3u8_id='hls', fatal=False))
 353                         continue
 354                     elif 'akamaihd.net' in url_comp.netloc:
 355                         # This type of renditions are served through
 356                         # akamaihd.net, but they don't use f4m manifests
 357                         url = url.replace('control/', '') + '?&v=3.3.0&fp=13&r=FEEFJ&g=RTSJIMBMPFPB'
 358                         ext = 'flv'
 359                 if ext is None:
 360                     ext = determine_ext(url)
 361                 tbr = int_or_none(rend.get('encodingRate'), 1000)
 362                 a_format = {
 363                     'format_id': 'http%s' % ('-%s' % tbr if tbr else ''),
 364                     'url': url,
 365                     'ext': ext,
 366                     'filesize': int_or_none(rend.get('size')) or None,
 367                     'tbr': tbr,
 368                 }
 369                 if rend.get('audioOnly'):
 370                     a_format.update({
 371                         'vcodec': 'none',
 372                     })
 373                 else:
 374                     a_format.update({
 375                         'height': int_or_none(rend.get('frameHeight')),
 376                         'width': int_or_none(rend.get('frameWidth')),
 377                         'vcodec': rend.get('videoCodec'),
 378                     })
 379
 380                 # m3u8 manifests with remote == false are media playlists
 381                 # Not calling _extract_m3u8_formats here to save network traffic
 382                 if ext == 'm3u8':
 383                     a_format.update({
 384                         'format_id': 'hls%s' % ('-%s' % tbr if tbr else ''),
 385                         'ext': 'mp4',
 386                         'protocol': 'm3u8_native',
 387                     })
 388
 389                 formats.append(a_format)
 390             self._sort_formats(formats)
 391             info['formats'] = formats
 392         elif video_info.get('FLVFullLengthURL') is not None:
 393             info.update({
 394                 'url': video_info['FLVFullLengthURL'],
 395                 'vcodec': self.FLV_VCODECS.get(video_info.get('FLVFullCodec')),
 396                 'filesize': int_or_none(video_info.get('FLVFullSize')),
 397             })
 398
 399         if self._downloader.params.get('include_ads', False):
 400             adServerURL = video_info.get('_youtubedl_adServerURL')
 401             if adServerURL:
 402                 ad_info = {
 403                     '_type': 'url',
 404                     'url': adServerURL,
 405                 }
 406                 if 'url' in info:
 407                     return {
 408                         '_type': 'playlist',
 409                         'title': info['title'],
 410                         'entries': [ad_info, info],
 411                     }
 412                 else:
 413                     return ad_info
 414
 415         if 'url' not in info and not info.get('formats'):
 416             raise ExtractorError('Unable to extract video url for %s' % video_id)
 417         return info
 418
 419
 420 class BrightcoveNewIE(InfoExtractor):
 421     IE_NAME = 'brightcove:new'
 422     _VALID_URL = r'https?://players\.brightcove\.net/(?P<account_id>\d+)/(?P<player_id>[^/]+)_(?P<embed>[^/]+)/index\.html\?.*videoId=(?P<video_id>\d+|ref:[^&]+)'
 423     _TESTS = [{
 424         'url': 'http://players.brightcove.net/929656772001/e41d32dc-ec74-459e-a845-6c69f7b724ea_default/index.html?videoId=4463358922001',
 425         'md5': 'c8100925723840d4b0d243f7025703be',
 426         'info_dict': {
 427             'id': '4463358922001',
 428             'ext': 'mp4',
 429             'title': 'Meet the man behind Popcorn Time',
 430             'description': 'md5:eac376a4fe366edc70279bfb681aea16',
 431             'duration': 165.768,
 432             'timestamp': 1441391203,
 433             'upload_date': '20150904',
 434             'uploader_id': '929656772001',
 435             'formats': 'mincount:22',
 436         },
 437     }, {
 438         # with rtmp streams
 439         'url': 'http://players.brightcove.net/4036320279001/5d112ed9-283f-485f-a7f9-33f42e8bc042_default/index.html?videoId=4279049078001',
 440         'info_dict': {
 441             'id': '4279049078001',
 442             'ext': 'mp4',
 443             'title': 'Titansgrave: Chapter 0',
 444             'description': 'Titansgrave: Chapter 0',
 445             'duration': 1242.058,
 446             'timestamp': 1433556729,
 447             'upload_date': '20150606',
 448             'uploader_id': '4036320279001',
 449             'formats': 'mincount:41',
 450         },
 451         'params': {
 452             # m3u8 download
 453             'skip_download': True,
 454         }
 455     }, {
 456         # ref: prefixed video id
 457         'url': 'http://players.brightcove.net/3910869709001/21519b5c-4b3b-4363-accb-bdc8f358f823_default/index.html?videoId=ref:7069442',
 458         'only_matching': True,
 459     }, {
 460         # non numeric ref: prefixed video id
 461         'url': 'http://players.brightcove.net/710858724001/default_default/index.html?videoId=ref:event-stream-356',
 462         'only_matching': True,
 463     }, {
 464         # unavailable video without message but with error_code
 465         'url': 'http://players.brightcove.net/1305187701/c832abfb-641b-44eb-9da0-2fe76786505f_default/index.html?videoId=4377407326001',
 466         'only_matching': True,
 467     }]
 468
 469     @staticmethod
 470     def _extract_url(webpage):
 471         urls = BrightcoveNewIE._extract_urls(webpage)
 472         return urls[0] if urls else None
 473
 474     @staticmethod
 475     def _extract_urls(webpage):
 476         # Reference:
 477         # 1. http://docs.brightcove.com/en/video-cloud/brightcove-player/guides/publish-video.html#setvideoiniframe
 478         # 2. http://docs.brightcove.com/en/video-cloud/brightcove-player/guides/publish-video.html#setvideousingjavascript
 479         # 3. http://docs.brightcove.com/en/video-cloud/brightcove-player/guides/embed-in-page.html
 480         # 4. https://support.brightcove.com/en/video-cloud/docs/dynamically-assigning-videos-player
 481
 482         entries = []
 483
 484         # Look for iframe embeds [1]
 485         for _, url in re.findall(
 486                 r'<iframe[^>]+src=(["\'])((?:https?:)?//players\.brightcove\.net/\d+/[^/]+/index\.html.+?)\1', webpage):
 487             entries.append(url if url.startswith('http') else 'http:' + url)
 488
 489         # Look for embed_in_page embeds [2]
 490         for video_id, account_id, player_id, embed in re.findall(
 491                 # According to examples from [3] it's unclear whether video id
 492                 # may be optional and what to do when it is
 493                 # According to [4] data-video-id may be prefixed with ref:
 494                 r'''(?sx)
 495                     <video[^>]+
 496                         data-video-id=["\'](\d+|ref:[^"\']+)["\'][^>]*>.*?
 497                     </video>.*?
 498                     <script[^>]+
 499                         src=["\'](?:https?:)?//players\.brightcove\.net/
 500                         (\d+)/([^/]+)_([^/]+)/index(?:\.min)?\.js
 501                 ''', webpage):
 502             entries.append(
 503                 'http://players.brightcove.net/%s/%s_%s/index.html?videoId=%s'
 504                 % (account_id, player_id, embed, video_id))
 505
 506         return entries
 507
 508     def _real_extract(self, url):
 509         account_id, player_id, embed, video_id = re.match(self._VALID_URL, url).groups()
 510
 511         webpage = self._download_webpage(
 512             'http://players.brightcove.net/%s/%s_%s/index.min.js'
 513             % (account_id, player_id, embed), video_id)
 514
 515         policy_key = None
 516
 517         catalog = self._search_regex(
 518             r'catalog\(({.+?})\);', webpage, 'catalog', default=None)
 519         if catalog:
 520             catalog = self._parse_json(
 521                 js_to_json(catalog), video_id, fatal=False)
 522             if catalog:
 523                 policy_key = catalog.get('policyKey')
 524
 525         if not policy_key:
 526             policy_key = self._search_regex(
 527                 r'policyKey\s*:\s*(["\'])(?P<pk>.+?)\1',
 528                 webpage, 'policy key', group='pk')
 529
 530         api_url = 'https://edge.api.brightcove.com/playback/v1/accounts/%s/videos/%s' % (account_id, video_id)
 531         try:
 532             json_data = self._download_json(api_url, video_id, headers={
 533                 'Accept': 'application/json;pk=%s' % policy_key
 534             })
 535         except ExtractorError as e:
 536             if isinstance(e.cause, compat_HTTPError) and e.cause.code == 403:
 537                 json_data = self._parse_json(e.cause.read().decode(), video_id)[0]
 538                 raise ExtractorError(
 539                     json_data.get('message') or json_data['error_code'], expected=True)
 540             raise
 541
 542         title = json_data['name'].strip()
 543
 544         formats = []
 545         for source in json_data.get('sources', []):
 546             container = source.get('container')
 547             source_type = source.get('type')
 548             src = source.get('src')
 549             if source_type == 'application/x-mpegURL' or container == 'M2TS':
 550                 if not src:
 551                     continue
 552                 formats.extend(self._extract_m3u8_formats(
 553                     src, video_id, 'mp4', 'm3u8_native', m3u8_id='hls', fatal=False))
 554             elif source_type == 'application/dash+xml':
 555                 if not src:
 556                     continue
 557                 formats.extend(self._extract_mpd_formats(src, video_id, 'dash', fatal=False))
 558             else:
 559                 streaming_src = source.get('streaming_src')
 560                 stream_name, app_name = source.get('stream_name'), source.get('app_name')
 561                 if not src and not streaming_src and (not stream_name or not app_name):
 562                     continue
 563                 tbr = float_or_none(source.get('avg_bitrate'), 1000)
 564                 height = int_or_none(source.get('height'))
 565                 width = int_or_none(source.get('width'))
 566                 f = {
 567                     'tbr': tbr,
 568                     'filesize': int_or_none(source.get('size')),
 569                     'container': container,
 570                     'ext': container.lower(),
 571                 }
 572                 if width == 0 and height == 0:
 573                     f.update({
 574                         'vcodec': 'none',
 575                     })
 576                 else:
 577                     f.update({
 578                         'width': width,
 579                         'height': height,
 580                         'vcodec': source.get('codec'),
 581                     })
 582
 583                 def build_format_id(kind):
 584                     format_id = kind
 585                     if tbr:
 586                         format_id += '-%dk' % int(tbr)
 587                     if height:
 588                         format_id += '-%dp' % height
 589                     return format_id
 590
 591                 if src or streaming_src:
 592                     f.update({
 593                         'url': src or streaming_src,
 594                         'format_id': build_format_id('http' if src else 'http-streaming'),
 595                         'source_preference': 0 if src else -1,
 596                     })
 597                 else:
 598                     f.update({
 599                         'url': app_name,
 600                         'play_path': stream_name,
 601                         'format_id': build_format_id('rtmp'),
 602                     })
 603                 formats.append(f)
 604
 605         errors = json_data.get('errors')
 606         if not formats and errors:
 607             error = errors[0]
 608             raise ExtractorError(
 609                 error.get('message') or error.get('error_subcode') or error['error_code'], expected=True)
 610
 611         self._sort_formats(formats)
 612
 613         subtitles = {}
 614         for text_track in json_data.get('text_tracks', []):
 615             if text_track.get('src'):
 616                 subtitles.setdefault(text_track.get('srclang'), []).append({
 617                     'url': text_track['src'],
 618                 })
 619
 620         return {
 621             'id': video_id,
 622             'title': title,
 623             'description': json_data.get('description'),
 624             'thumbnail': json_data.get('thumbnail') or json_data.get('poster'),
 625             'duration': float_or_none(json_data.get('duration'), 1000),
 626             'timestamp': parse_iso8601(json_data.get('published_at')),
 627             'uploader_id': account_id,
 628             'formats': formats,
 629             'subtitles': subtitles,
 630             'tags': json_data.get('tags', []),
 631         }