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

   1 # coding: 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     clean_html,
  30     mimetype2ext,
  31 )
  32
  33
  34 class BrightcoveLegacyIE(InfoExtractor):
  35     IE_NAME = 'brightcove:legacy'
  36     _VALID_URL = r'(?:https?://.*brightcove\.com/(services|viewer).*?\?|brightcove:)(?P<query>.*)'
  37     _FEDERATED_URL = 'http://c.brightcove.com/services/viewer/htmlFederated'
  38
  39     _TESTS = [
  40         {
  41             # From http://www.8tv.cat/8aldia/videos/xavier-sala-i-martin-aquesta-tarda-a-8-al-dia/
  42             'url': 'http://c.brightcove.com/services/viewer/htmlFederated?playerID=1654948606001&flashID=myExperience&%40videoPlayer=2371591881001',
  43             'md5': '5423e113865d26e40624dce2e4b45d95',
  44             'note': 'Test Brightcove downloads and detection in GenericIE',
  45             'info_dict': {
  46                 'id': '2371591881001',
  47                 'ext': 'mp4',
  48                 'title': 'Xavier Sala i Martín: “Un banc que no presta és un banc zombi que no serveix per a res”',
  49                 'uploader': '8TV',
  50                 'description': 'md5:a950cc4285c43e44d763d036710cd9cd',
  51                 'timestamp': 1368213670,
  52                 'upload_date': '20130510',
  53                 'uploader_id': '1589608506001',
  54             }
  55         },
  56         {
  57             # From http://medianetwork.oracle.com/video/player/1785452137001
  58             'url': 'http://c.brightcove.com/services/viewer/htmlFederated?playerID=1217746023001&flashID=myPlayer&%40videoPlayer=1785452137001',
  59             'info_dict': {
  60                 'id': '1785452137001',
  61                 'ext': 'flv',
  62                 'title': 'JVMLS 2012: Arrays 2.0 - Opportunities and Challenges',
  63                 'description': 'John Rose speaks at the JVM Language Summit, August 1, 2012.',
  64                 'uploader': 'Oracle',
  65                 'timestamp': 1344975024,
  66                 'upload_date': '20120814',
  67                 'uploader_id': '1460825906',
  68             },
  69         },
  70         {
  71             # From http://mashable.com/2013/10/26/thermoelectric-bracelet-lets-you-control-your-body-temperature/
  72             'url': 'http://c.brightcove.com/services/viewer/federated_f9?&playerID=1265504713001&publisherID=AQ%7E%7E%2CAAABBzUwv1E%7E%2CxP-xFHVUstiMFlNYfvF4G9yFnNaqCw_9&videoID=2750934548001',
  73             'info_dict': {
  74                 'id': '2750934548001',
  75                 'ext': 'mp4',
  76                 'title': 'This Bracelet Acts as a Personal Thermostat',
  77                 'description': 'md5:547b78c64f4112766ccf4e151c20b6a0',
  78                 'uploader': 'Mashable',
  79                 'timestamp': 1382041798,
  80                 'upload_date': '20131017',
  81                 'uploader_id': '1130468786001',
  82             },
  83         },
  84         {
  85             # test that the default referer works
  86             # from http://national.ballet.ca/interact/video/Lost_in_Motion_II/
  87             'url': 'http://link.brightcove.com/services/player/bcpid756015033001?bckey=AQ~~,AAAApYJi_Ck~,GxhXCegT1Dp39ilhXuxMJxasUhVNZiil&bctid=2878862109001',
  88             'info_dict': {
  89                 'id': '2878862109001',
  90                 'ext': 'mp4',
  91                 'title': 'Lost in Motion II',
  92                 'description': 'md5:363109c02998fee92ec02211bd8000df',
  93                 'uploader': 'National Ballet of Canada',
  94             },
  95             'skip': 'Video gone',
  96         },
  97         {
  98             # test flv videos served by akamaihd.net
  99             # From http://www.redbull.com/en/bike/stories/1331655643987/replay-uci-dh-world-cup-2014-from-fort-william
 100             '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',
 101             # The md5 checksum changes on each download
 102             'info_dict': {
 103                 'id': '3750436379001',
 104                 'ext': 'flv',
 105                 'title': 'UCI MTB World Cup 2014: Fort William, UK - Downhill Finals',
 106                 'uploader': 'RBTV Old (do not use)',
 107                 'description': 'UCI MTB World Cup 2014: Fort William, UK - Downhill Finals',
 108                 'timestamp': 1409122195,
 109                 'upload_date': '20140827',
 110                 'uploader_id': '710858724001',
 111             },
 112         },
 113         {
 114             # playlist with 'videoList'
 115             # from http://support.brightcove.com/en/video-cloud/docs/playlist-support-single-video-players
 116             'url': 'http://c.brightcove.com/services/viewer/htmlFederated?playerID=3550052898001&playerKey=AQ%7E%7E%2CAAABmA9XpXk%7E%2C-Kp7jNgisre1fG5OdqpAFUTcs0lP_ZoL',
 117             'info_dict': {
 118                 'title': 'Sealife',
 119                 'id': '3550319591001',
 120             },
 121             'playlist_mincount': 7,
 122         },
 123         {
 124             # playlist with 'playlistTab' (https://github.com/rg3/youtube-dl/issues/9965)
 125             'url': 'http://c.brightcove.com/services/json/experience/runtime/?command=get_programming_for_experience&playerKey=AQ%7E%7E,AAABXlLMdok%7E,NJ4EoMlZ4rZdx9eU1rkMVd8EaYPBBUlg',
 126             'info_dict': {
 127                 'id': '1522758701001',
 128                 'title': 'Lesson 08',
 129             },
 130             'playlist_mincount': 10,
 131         },
 132     ]
 133     FLV_VCODECS = {
 134         1: 'SORENSON',
 135         2: 'ON2',
 136         3: 'H264',
 137         4: 'VP8',
 138     }
 139
 140     @classmethod
 141     def _build_brighcove_url(cls, object_str):
 142         """
 143         Build a Brightcove url from a xml string containing
 144         <object class="BrightcoveExperience">{params}</object>
 145         """
 146
 147         # Fix up some stupid HTML, see https://github.com/rg3/youtube-dl/issues/1553
 148         object_str = re.sub(r'(<param(?:\s+[a-zA-Z0-9_]+="[^"]*")*)>',
 149                             lambda m: m.group(1) + '/>', object_str)
 150         # Fix up some stupid XML, see https://github.com/rg3/youtube-dl/issues/1608
 151         object_str = object_str.replace('<--', '<!--')
 152         # remove namespace to simplify extraction
 153         object_str = re.sub(r'(<object[^>]*)(xmlns=".*?")', r'\1', object_str)
 154         object_str = fix_xml_ampersands(object_str)
 155
 156         try:
 157             object_doc = compat_etree_fromstring(object_str.encode('utf-8'))
 158         except compat_xml_parse_error:
 159             return
 160
 161         fv_el = find_xpath_attr(object_doc, './param', 'name', 'flashVars')
 162         if fv_el is not None:
 163             flashvars = dict(
 164                 (k, v[0])
 165                 for k, v in compat_parse_qs(fv_el.attrib['value']).items())
 166         else:
 167             flashvars = {}
 168
 169         data_url = object_doc.attrib.get('data', '')
 170         data_url_params = compat_parse_qs(compat_urllib_parse_urlparse(data_url).query)
 171
 172         def find_param(name):
 173             if name in flashvars:
 174                 return flashvars[name]
 175             node = find_xpath_attr(object_doc, './param', 'name', name)
 176             if node is not None:
 177                 return node.attrib['value']
 178             return data_url_params.get(name)
 179
 180         params = {}
 181
 182         playerID = find_param('playerID') or find_param('playerId')
 183         if playerID is None:
 184             raise ExtractorError('Cannot find player ID')
 185         params['playerID'] = playerID
 186
 187         playerKey = find_param('playerKey')
 188         # Not all pages define this value
 189         if playerKey is not None:
 190             params['playerKey'] = playerKey
 191         # These fields hold the id of the video
 192         videoPlayer = find_param('@videoPlayer') or find_param('videoId') or find_param('videoID') or find_param('@videoList')
 193         if videoPlayer is not None:
 194             if isinstance(videoPlayer, list):
 195                 videoPlayer = videoPlayer[0]
 196             videoPlayer = videoPlayer.strip()
 197             # UUID is also possible for videoPlayer (e.g.
 198             # http://www.popcornflix.com/hoodies-vs-hooligans/7f2d2b87-bbf2-4623-acfb-ea942b4f01dd
 199             # or http://www8.hp.com/cn/zh/home.html)
 200             if not (re.match(
 201                     r'^(?:\d+|[\da-fA-F]{8}-?[\da-fA-F]{4}-?[\da-fA-F]{4}-?[\da-fA-F]{4}-?[\da-fA-F]{12})$',
 202                     videoPlayer) or videoPlayer.startswith('ref:')):
 203                 return None
 204             params['@videoPlayer'] = videoPlayer
 205         linkBase = find_param('linkBaseURL')
 206         if linkBase is not None:
 207             params['linkBaseURL'] = linkBase
 208         return cls._make_brightcove_url(params)
 209
 210     @classmethod
 211     def _build_brighcove_url_from_js(cls, object_js):
 212         # The layout of JS is as follows:
 213         # customBC.createVideo = function (width, height, playerID, playerKey, videoPlayer, VideoRandomID) {
 214         #   // build Brightcove <object /> XML
 215         # }
 216         m = re.search(
 217             r'''(?x)customBC\.createVideo\(
 218                 .*?                                                  # skipping width and height
 219                 ["\'](?P<playerID>\d+)["\']\s*,\s*                   # playerID
 220                 ["\'](?P<playerKey>AQ[^"\']{48})[^"\']*["\']\s*,\s*  # playerKey begins with AQ and is 50 characters
 221                                                                      # in length, however it's appended to itself
 222                                                                      # in places, so truncate
 223                 ["\'](?P<videoID>\d+)["\']                           # @videoPlayer
 224             ''', object_js)
 225         if m:
 226             return cls._make_brightcove_url(m.groupdict())
 227
 228     @classmethod
 229     def _make_brightcove_url(cls, params):
 230         return update_url_query(cls._FEDERATED_URL, params)
 231
 232     @classmethod
 233     def _extract_brightcove_url(cls, webpage):
 234         """Try to extract the brightcove url from the webpage, returns None
 235         if it can't be found
 236         """
 237         urls = cls._extract_brightcove_urls(webpage)
 238         return urls[0] if urls else None
 239
 240     @classmethod
 241     def _extract_brightcove_urls(cls, webpage):
 242         """Return a list of all Brightcove URLs from the webpage """
 243
 244         url_m = re.search(
 245             r'''(?x)
 246                 <meta\s+
 247                     (?:property|itemprop)=([\'"])(?:og:video|embedURL)\1[^>]+
 248                     content=([\'"])(?P<url>https?://(?:secure|c)\.brightcove.com/(?:(?!\2).)+)\2
 249             ''', webpage)
 250         if url_m:
 251             url = unescapeHTML(url_m.group('url'))
 252             # Some sites don't add it, we can't download with this url, for example:
 253             # http://www.ktvu.com/videos/news/raw-video-caltrain-releases-video-of-man-almost/vCTZdY/
 254             if 'playerKey' in url or 'videoId' in url or 'idVideo' in url:
 255                 return [url]
 256
 257         matches = re.findall(
 258             r'''(?sx)<object
 259             (?:
 260                 [^>]+?class=[\'"][^>]*?BrightcoveExperience.*?[\'"] |
 261                 [^>]*?>\s*<param\s+name="movie"\s+value="https?://[^/]*brightcove\.com/
 262             ).+?>\s*</object>''',
 263             webpage)
 264         if matches:
 265             return list(filter(None, [cls._build_brighcove_url(m) for m in matches]))
 266
 267         return list(filter(None, [
 268             cls._build_brighcove_url_from_js(custom_bc)
 269             for custom_bc in re.findall(r'(customBC\.createVideo\(.+?\);)', webpage)]))
 270
 271     def _real_extract(self, url):
 272         url, smuggled_data = unsmuggle_url(url, {})
 273
 274         # Change the 'videoId' and others field to '@videoPlayer'
 275         url = re.sub(r'(?<=[?&])(videoI(d|D)|idVideo|bctid)', '%40videoPlayer', url)
 276         # Change bckey (used by bcove.me urls) to playerKey
 277         url = re.sub(r'(?<=[?&])bckey', 'playerKey', url)
 278         mobj = re.match(self._VALID_URL, url)
 279         query_str = mobj.group('query')
 280         query = compat_urlparse.parse_qs(query_str)
 281
 282         videoPlayer = query.get('@videoPlayer')
 283         if videoPlayer:
 284             # We set the original url as the default 'Referer' header
 285             referer = smuggled_data.get('Referer', url)
 286             return self._get_video_info(
 287                 videoPlayer[0], query, referer=referer)
 288         elif 'playerKey' in query:
 289             player_key = query['playerKey']
 290             return self._get_playlist_info(player_key[0])
 291         else:
 292             raise ExtractorError(
 293                 'Cannot find playerKey= variable. Did you forget quotes in a shell invocation?',
 294                 expected=True)
 295
 296     def _get_video_info(self, video_id, query, referer=None):
 297         headers = {}
 298         linkBase = query.get('linkBaseURL')
 299         if linkBase is not None:
 300             referer = linkBase[0]
 301         if referer is not None:
 302             headers['Referer'] = referer
 303         webpage = self._download_webpage(self._FEDERATED_URL, video_id, headers=headers, query=query)
 304
 305         error_msg = self._html_search_regex(
 306             r"<h1>We're sorry.</h1>([\s\n]*<p>.*?</p>)+", webpage,
 307             'error message', default=None)
 308         if error_msg is not None:
 309             raise ExtractorError(
 310                 'brightcove said: %s' % error_msg, expected=True)
 311
 312         self.report_extraction(video_id)
 313         info = self._search_regex(r'var experienceJSON = ({.*});', webpage, 'json')
 314         info = json.loads(info)['data']
 315         video_info = info['programmedContent']['videoPlayer']['mediaDTO']
 316         video_info['_youtubedl_adServerURL'] = info.get('adServerURL')
 317
 318         return self._extract_video_info(video_info)
 319
 320     def _get_playlist_info(self, player_key):
 321         info_url = 'http://c.brightcove.com/services/json/experience/runtime/?command=get_programming_for_experience&playerKey=%s' % player_key
 322         playlist_info = self._download_webpage(
 323             info_url, player_key, 'Downloading playlist information')
 324
 325         json_data = json.loads(playlist_info)
 326         if 'videoList' in json_data:
 327             playlist_info = json_data['videoList']
 328             playlist_dto = playlist_info['mediaCollectionDTO']
 329         elif 'playlistTabs' in json_data:
 330             playlist_info = json_data['playlistTabs']
 331             playlist_dto = playlist_info['lineupListDTO']['playlistDTOs'][0]
 332         else:
 333             raise ExtractorError('Empty playlist')
 334
 335         videos = [self._extract_video_info(video_info) for video_info in playlist_dto['videoDTOs']]
 336
 337         return self.playlist_result(videos, playlist_id='%s' % playlist_info['id'],
 338                                     playlist_title=playlist_dto['displayName'])
 339
 340     def _extract_video_info(self, video_info):
 341         video_id = compat_str(video_info['id'])
 342         publisher_id = video_info.get('publisherId')
 343         info = {
 344             'id': video_id,
 345             'title': video_info['displayName'].strip(),
 346             'description': video_info.get('shortDescription'),
 347             'thumbnail': video_info.get('videoStillURL') or video_info.get('thumbnailURL'),
 348             'uploader': video_info.get('publisherName'),
 349             'uploader_id': compat_str(publisher_id) if publisher_id else None,
 350             'duration': float_or_none(video_info.get('length'), 1000),
 351             'timestamp': int_or_none(video_info.get('creationDate'), 1000),
 352         }
 353
 354         renditions = video_info.get('renditions', []) + video_info.get('IOSRenditions', [])
 355         if renditions:
 356             formats = []
 357             for rend in renditions:
 358                 url = rend['defaultURL']
 359                 if not url:
 360                     continue
 361                 ext = None
 362                 if rend['remote']:
 363                     url_comp = compat_urllib_parse_urlparse(url)
 364                     if url_comp.path.endswith('.m3u8'):
 365                         formats.extend(
 366                             self._extract_m3u8_formats(
 367                                 url, video_id, 'mp4', 'm3u8_native', m3u8_id='hls', fatal=False))
 368                         continue
 369                     elif 'akamaihd.net' in url_comp.netloc:
 370                         # This type of renditions are served through
 371                         # akamaihd.net, but they don't use f4m manifests
 372                         url = url.replace('control/', '') + '?&v=3.3.0&fp=13&r=FEEFJ&g=RTSJIMBMPFPB'
 373                         ext = 'flv'
 374                 if ext is None:
 375                     ext = determine_ext(url)
 376                 tbr = int_or_none(rend.get('encodingRate'), 1000)
 377                 a_format = {
 378                     'format_id': 'http%s' % ('-%s' % tbr if tbr else ''),
 379                     'url': url,
 380                     'ext': ext,
 381                     'filesize': int_or_none(rend.get('size')) or None,
 382                     'tbr': tbr,
 383                 }
 384                 if rend.get('audioOnly'):
 385                     a_format.update({
 386                         'vcodec': 'none',
 387                     })
 388                 else:
 389                     a_format.update({
 390                         'height': int_or_none(rend.get('frameHeight')),
 391                         'width': int_or_none(rend.get('frameWidth')),
 392                         'vcodec': rend.get('videoCodec'),
 393                     })
 394
 395                 # m3u8 manifests with remote == false are media playlists
 396                 # Not calling _extract_m3u8_formats here to save network traffic
 397                 if ext == 'm3u8':
 398                     a_format.update({
 399                         'format_id': 'hls%s' % ('-%s' % tbr if tbr else ''),
 400                         'ext': 'mp4',
 401                         'protocol': 'm3u8_native',
 402                     })
 403
 404                 formats.append(a_format)
 405             self._sort_formats(formats)
 406             info['formats'] = formats
 407         elif video_info.get('FLVFullLengthURL') is not None:
 408             info.update({
 409                 'url': video_info['FLVFullLengthURL'],
 410                 'vcodec': self.FLV_VCODECS.get(video_info.get('FLVFullCodec')),
 411                 'filesize': int_or_none(video_info.get('FLVFullSize')),
 412             })
 413
 414         if self._downloader.params.get('include_ads', False):
 415             adServerURL = video_info.get('_youtubedl_adServerURL')
 416             if adServerURL:
 417                 ad_info = {
 418                     '_type': 'url',
 419                     'url': adServerURL,
 420                 }
 421                 if 'url' in info:
 422                     return {
 423                         '_type': 'playlist',
 424                         'title': info['title'],
 425                         'entries': [ad_info, info],
 426                     }
 427                 else:
 428                     return ad_info
 429
 430         if 'url' not in info and not info.get('formats'):
 431             raise ExtractorError('Unable to extract video url for %s' % video_id)
 432         return info
 433
 434
 435 class BrightcoveNewIE(InfoExtractor):
 436     IE_NAME = 'brightcove:new'
 437     _VALID_URL = r'https?://players\.brightcove\.net/(?P<account_id>\d+)/(?P<player_id>[^/]+)_(?P<embed>[^/]+)/index\.html\?.*videoId=(?P<video_id>\d+|ref:[^&]+)'
 438     _TESTS = [{
 439         'url': 'http://players.brightcove.net/929656772001/e41d32dc-ec74-459e-a845-6c69f7b724ea_default/index.html?videoId=4463358922001',
 440         'md5': 'c8100925723840d4b0d243f7025703be',
 441         'info_dict': {
 442             'id': '4463358922001',
 443             'ext': 'mp4',
 444             'title': 'Meet the man behind Popcorn Time',
 445             'description': 'md5:eac376a4fe366edc70279bfb681aea16',
 446             'duration': 165.768,
 447             'timestamp': 1441391203,
 448             'upload_date': '20150904',
 449             'uploader_id': '929656772001',
 450             'formats': 'mincount:22',
 451         },
 452     }, {
 453         # with rtmp streams
 454         'url': 'http://players.brightcove.net/4036320279001/5d112ed9-283f-485f-a7f9-33f42e8bc042_default/index.html?videoId=4279049078001',
 455         'info_dict': {
 456             'id': '4279049078001',
 457             'ext': 'mp4',
 458             'title': 'Titansgrave: Chapter 0',
 459             'description': 'Titansgrave: Chapter 0',
 460             'duration': 1242.058,
 461             'timestamp': 1433556729,
 462             'upload_date': '20150606',
 463             'uploader_id': '4036320279001',
 464             'formats': 'mincount:41',
 465         },
 466         'params': {
 467             # m3u8 download
 468             'skip_download': True,
 469         }
 470     }, {
 471         # ref: prefixed video id
 472         'url': 'http://players.brightcove.net/3910869709001/21519b5c-4b3b-4363-accb-bdc8f358f823_default/index.html?videoId=ref:7069442',
 473         'only_matching': True,
 474     }, {
 475         # non numeric ref: prefixed video id
 476         'url': 'http://players.brightcove.net/710858724001/default_default/index.html?videoId=ref:event-stream-356',
 477         'only_matching': True,
 478     }, {
 479         # unavailable video without message but with error_code
 480         'url': 'http://players.brightcove.net/1305187701/c832abfb-641b-44eb-9da0-2fe76786505f_default/index.html?videoId=4377407326001',
 481         'only_matching': True,
 482     }]
 483
 484     @staticmethod
 485     def _extract_url(webpage):
 486         urls = BrightcoveNewIE._extract_urls(webpage)
 487         return urls[0] if urls else None
 488
 489     @staticmethod
 490     def _extract_urls(webpage):
 491         # Reference:
 492         # 1. http://docs.brightcove.com/en/video-cloud/brightcove-player/guides/publish-video.html#setvideoiniframe
 493         # 2. http://docs.brightcove.com/en/video-cloud/brightcove-player/guides/publish-video.html#setvideousingjavascript
 494         # 3. http://docs.brightcove.com/en/video-cloud/brightcove-player/guides/embed-in-page.html
 495         # 4. https://support.brightcove.com/en/video-cloud/docs/dynamically-assigning-videos-player
 496
 497         entries = []
 498
 499         # Look for iframe embeds [1]
 500         for _, url in re.findall(
 501                 r'<iframe[^>]+src=(["\'])((?:https?:)?//players\.brightcove\.net/\d+/[^/]+/index\.html.+?)\1', webpage):
 502             entries.append(url if url.startswith('http') else 'http:' + url)
 503
 504         # Look for embed_in_page embeds [2]
 505         for video_id, account_id, player_id, embed in re.findall(
 506                 # According to examples from [3] it's unclear whether video id
 507                 # may be optional and what to do when it is
 508                 # According to [4] data-video-id may be prefixed with ref:
 509                 r'''(?sx)
 510                     <video[^>]+
 511                         data-video-id=["\'](\d+|ref:[^"\']+)["\'][^>]*>.*?
 512                     </video>.*?
 513                     <script[^>]+
 514                         src=["\'](?:https?:)?//players\.brightcove\.net/
 515                         (\d+)/([^/]+)_([^/]+)/index(?:\.min)?\.js
 516                 ''', webpage):
 517             entries.append(
 518                 'http://players.brightcove.net/%s/%s_%s/index.html?videoId=%s'
 519                 % (account_id, player_id, embed, video_id))
 520
 521         return entries
 522
 523     def _real_extract(self, url):
 524         url, smuggled_data = unsmuggle_url(url, {})
 525         self._initialize_geo_bypass(smuggled_data.get('geo_countries'))
 526
 527         account_id, player_id, embed, video_id = re.match(self._VALID_URL, url).groups()
 528
 529         webpage = self._download_webpage(
 530             'http://players.brightcove.net/%s/%s_%s/index.min.js'
 531             % (account_id, player_id, embed), video_id)
 532
 533         policy_key = None
 534
 535         catalog = self._search_regex(
 536             r'catalog\(({.+?})\);', webpage, 'catalog', default=None)
 537         if catalog:
 538             catalog = self._parse_json(
 539                 js_to_json(catalog), video_id, fatal=False)
 540             if catalog:
 541                 policy_key = catalog.get('policyKey')
 542
 543         if not policy_key:
 544             policy_key = self._search_regex(
 545                 r'policyKey\s*:\s*(["\'])(?P<pk>.+?)\1',
 546                 webpage, 'policy key', group='pk')
 547
 548         api_url = 'https://edge.api.brightcove.com/playback/v1/accounts/%s/videos/%s' % (account_id, video_id)
 549         try:
 550             json_data = self._download_json(api_url, video_id, headers={
 551                 'Accept': 'application/json;pk=%s' % policy_key
 552             })
 553         except ExtractorError as e:
 554             if isinstance(e.cause, compat_HTTPError) and e.cause.code == 403:
 555                 json_data = self._parse_json(e.cause.read().decode(), video_id)[0]
 556                 message = json_data.get('message') or json_data['error_code']
 557                 if json_data.get('error_subcode') == 'CLIENT_GEO':
 558                     self.raise_geo_restricted(msg=message)
 559                 raise ExtractorError(message, expected=True)
 560             raise
 561
 562         title = json_data['name'].strip()
 563
 564         formats = []
 565         for source in json_data.get('sources', []):
 566             container = source.get('container')
 567             ext = mimetype2ext(source.get('type'))
 568             src = source.get('src')
 569             if ext == 'ism' or container == 'WVM':
 570                 continue
 571             elif ext == 'm3u8' or container == 'M2TS':
 572                 if not src:
 573                     continue
 574                 formats.extend(self._extract_m3u8_formats(
 575                     src, video_id, 'mp4', 'm3u8_native', m3u8_id='hls', fatal=False))
 576             elif ext == 'mpd':
 577                 if not src:
 578                     continue
 579                 formats.extend(self._extract_mpd_formats(src, video_id, 'dash', fatal=False))
 580             else:
 581                 streaming_src = source.get('streaming_src')
 582                 stream_name, app_name = source.get('stream_name'), source.get('app_name')
 583                 if not src and not streaming_src and (not stream_name or not app_name):
 584                     continue
 585                 tbr = float_or_none(source.get('avg_bitrate'), 1000)
 586                 height = int_or_none(source.get('height'))
 587                 width = int_or_none(source.get('width'))
 588                 f = {
 589                     'tbr': tbr,
 590                     'filesize': int_or_none(source.get('size')),
 591                     'container': container,
 592                     'ext': ext or container.lower(),
 593                 }
 594                 if width == 0 and height == 0:
 595                     f.update({
 596                         'vcodec': 'none',
 597                     })
 598                 else:
 599                     f.update({
 600                         'width': width,
 601                         'height': height,
 602                         'vcodec': source.get('codec'),
 603                     })
 604
 605                 def build_format_id(kind):
 606                     format_id = kind
 607                     if tbr:
 608                         format_id += '-%dk' % int(tbr)
 609                     if height:
 610                         format_id += '-%dp' % height
 611                     return format_id
 612
 613                 if src or streaming_src:
 614                     f.update({
 615                         'url': src or streaming_src,
 616                         'format_id': build_format_id('http' if src else 'http-streaming'),
 617                         'source_preference': 0 if src else -1,
 618                     })
 619                 else:
 620                     f.update({
 621                         'url': app_name,
 622                         'play_path': stream_name,
 623                         'format_id': build_format_id('rtmp'),
 624                     })
 625                 formats.append(f)
 626
 627         errors = json_data.get('errors')
 628         if not formats and errors:
 629             error = errors[0]
 630             raise ExtractorError(
 631                 error.get('message') or error.get('error_subcode') or error['error_code'], expected=True)
 632
 633         self._sort_formats(formats)
 634
 635         subtitles = {}
 636         for text_track in json_data.get('text_tracks', []):
 637             if text_track.get('src'):
 638                 subtitles.setdefault(text_track.get('srclang'), []).append({
 639                     'url': text_track['src'],
 640                 })
 641
 642         is_live = False
 643         duration = float_or_none(json_data.get('duration'), 1000)
 644         if duration and duration < 0:
 645             is_live = True
 646
 647         return {
 648             'id': video_id,
 649             'title': self._live_title(title) if is_live else title,
 650             'description': clean_html(json_data.get('description')),
 651             'thumbnail': json_data.get('thumbnail') or json_data.get('poster'),
 652             'duration': duration,
 653             'timestamp': parse_iso8601(json_data.get('published_at')),
 654             'uploader_id': account_id,
 655             'formats': formats,
 656             'subtitles': subtitles,
 657             'tags': json_data.get('tags', []),
 658             'is_live': is_live,
 659         }