_ 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     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     unescapeHTML,
  28     unsmuggle_url,
  29     update_url_query,
  30     clean_html,
  31     mimetype2ext,
  32 )
  33
  34
  35 class BrightcoveLegacyIE(InfoExtractor):
  36     IE_NAME = 'brightcove:legacy'
  37     _VALID_URL = r'(?:https?://.*brightcove\.com/(services|viewer).*?\?|brightcove:)(?P<query>.*)'
  38     _FEDERATED_URL = 'http://c.brightcove.com/services/viewer/htmlFederated'
  39
  40     _TESTS = [
  41         {
  42             # From http://www.8tv.cat/8aldia/videos/xavier-sala-i-martin-aquesta-tarda-a-8-al-dia/
  43             'url': 'http://c.brightcove.com/services/viewer/htmlFederated?playerID=1654948606001&flashID=myExperience&%40videoPlayer=2371591881001',
  44             'md5': '5423e113865d26e40624dce2e4b45d95',
  45             'note': 'Test Brightcove downloads and detection in GenericIE',
  46             'info_dict': {
  47                 'id': '2371591881001',
  48                 'ext': 'mp4',
  49                 'title': 'Xavier Sala i Martín: “Un banc que no presta és un banc zombi que no serveix per a res”',
  50                 'uploader': '8TV',
  51                 'description': 'md5:a950cc4285c43e44d763d036710cd9cd',
  52                 'timestamp': 1368213670,
  53                 'upload_date': '20130510',
  54                 'uploader_id': '1589608506001',
  55             }
  56         },
  57         {
  58             # From http://medianetwork.oracle.com/video/player/1785452137001
  59             'url': 'http://c.brightcove.com/services/viewer/htmlFederated?playerID=1217746023001&flashID=myPlayer&%40videoPlayer=1785452137001',
  60             'info_dict': {
  61                 'id': '1785452137001',
  62                 'ext': 'flv',
  63                 'title': 'JVMLS 2012: Arrays 2.0 - Opportunities and Challenges',
  64                 'description': 'John Rose speaks at the JVM Language Summit, August 1, 2012.',
  65                 'uploader': 'Oracle',
  66                 'timestamp': 1344975024,
  67                 'upload_date': '20120814',
  68                 'uploader_id': '1460825906',
  69             },
  70         },
  71         {
  72             # From http://mashable.com/2013/10/26/thermoelectric-bracelet-lets-you-control-your-body-temperature/
  73             'url': 'http://c.brightcove.com/services/viewer/federated_f9?&playerID=1265504713001&publisherID=AQ%7E%7E%2CAAABBzUwv1E%7E%2CxP-xFHVUstiMFlNYfvF4G9yFnNaqCw_9&videoID=2750934548001',
  74             'info_dict': {
  75                 'id': '2750934548001',
  76                 'ext': 'mp4',
  77                 'title': 'This Bracelet Acts as a Personal Thermostat',
  78                 'description': 'md5:547b78c64f4112766ccf4e151c20b6a0',
  79                 'uploader': 'Mashable',
  80                 'timestamp': 1382041798,
  81                 'upload_date': '20131017',
  82                 'uploader_id': '1130468786001',
  83             },
  84         },
  85         {
  86             # test that the default referer works
  87             # from http://national.ballet.ca/interact/video/Lost_in_Motion_II/
  88             'url': 'http://link.brightcove.com/services/player/bcpid756015033001?bckey=AQ~~,AAAApYJi_Ck~,GxhXCegT1Dp39ilhXuxMJxasUhVNZiil&bctid=2878862109001',
  89             'info_dict': {
  90                 'id': '2878862109001',
  91                 'ext': 'mp4',
  92                 'title': 'Lost in Motion II',
  93                 'description': 'md5:363109c02998fee92ec02211bd8000df',
  94                 'uploader': 'National Ballet of Canada',
  95             },
  96             'skip': 'Video gone',
  97         },
  98         {
  99             # test flv videos served by akamaihd.net
 100             # From http://www.redbull.com/en/bike/stories/1331655643987/replay-uci-dh-world-cup-2014-from-fort-william
 101             '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',
 102             # The md5 checksum changes on each download
 103             'info_dict': {
 104                 'id': '3750436379001',
 105                 'ext': 'flv',
 106                 'title': 'UCI MTB World Cup 2014: Fort William, UK - Downhill Finals',
 107                 'uploader': 'RBTV Old (do not use)',
 108                 'description': 'UCI MTB World Cup 2014: Fort William, UK - Downhill Finals',
 109                 'timestamp': 1409122195,
 110                 'upload_date': '20140827',
 111                 'uploader_id': '710858724001',
 112             },
 113             'skip': 'Video gone',
 114         },
 115         {
 116             # playlist with 'videoList'
 117             # from http://support.brightcove.com/en/video-cloud/docs/playlist-support-single-video-players
 118             'url': 'http://c.brightcove.com/services/viewer/htmlFederated?playerID=3550052898001&playerKey=AQ%7E%7E%2CAAABmA9XpXk%7E%2C-Kp7jNgisre1fG5OdqpAFUTcs0lP_ZoL',
 119             'info_dict': {
 120                 'title': 'Sealife',
 121                 'id': '3550319591001',
 122             },
 123             'playlist_mincount': 7,
 124         },
 125         {
 126             # playlist with 'playlistTab' (https://github.com/rg3/youtube-dl/issues/9965)
 127             'url': 'http://c.brightcove.com/services/json/experience/runtime/?command=get_programming_for_experience&playerKey=AQ%7E%7E,AAABXlLMdok%7E,NJ4EoMlZ4rZdx9eU1rkMVd8EaYPBBUlg',
 128             'info_dict': {
 129                 'id': '1522758701001',
 130                 'title': 'Lesson 08',
 131             },
 132             'playlist_mincount': 10,
 133         },
 134     ]
 135     FLV_VCODECS = {
 136         1: 'SORENSON',
 137         2: 'ON2',
 138         3: 'H264',
 139         4: 'VP8',
 140     }
 141
 142     @classmethod
 143     def _build_brighcove_url(cls, object_str):
 144         """
 145         Build a Brightcove url from a xml string containing
 146         <object class="BrightcoveExperience">{params}</object>
 147         """
 148
 149         # Fix up some stupid HTML, see https://github.com/rg3/youtube-dl/issues/1553
 150         object_str = re.sub(r'(<param(?:\s+[a-zA-Z0-9_]+="[^"]*")*)>',
 151                             lambda m: m.group(1) + '/>', object_str)
 152         # Fix up some stupid XML, see https://github.com/rg3/youtube-dl/issues/1608
 153         object_str = object_str.replace('<--', '<!--')
 154         # remove namespace to simplify extraction
 155         object_str = re.sub(r'(<object[^>]*)(xmlns=".*?")', r'\1', object_str)
 156         object_str = fix_xml_ampersands(object_str)
 157
 158         try:
 159             object_doc = compat_etree_fromstring(object_str.encode('utf-8'))
 160         except compat_xml_parse_error:
 161             return
 162
 163         fv_el = find_xpath_attr(object_doc, './param', 'name', 'flashVars')
 164         if fv_el is not None:
 165             flashvars = dict(
 166                 (k, v[0])
 167                 for k, v in compat_parse_qs(fv_el.attrib['value']).items())
 168         else:
 169             flashvars = {}
 170
 171         data_url = object_doc.attrib.get('data', '')
 172         data_url_params = compat_parse_qs(compat_urllib_parse_urlparse(data_url).query)
 173
 174         def find_param(name):
 175             if name in flashvars:
 176                 return flashvars[name]
 177             node = find_xpath_attr(object_doc, './param', 'name', name)
 178             if node is not None:
 179                 return node.attrib['value']
 180             return data_url_params.get(name)
 181
 182         params = {}
 183
 184         playerID = find_param('playerID') or find_param('playerId')
 185         if playerID is None:
 186             raise ExtractorError('Cannot find player ID')
 187         params['playerID'] = playerID
 188
 189         playerKey = find_param('playerKey')
 190         # Not all pages define this value
 191         if playerKey is not None:
 192             params['playerKey'] = playerKey
 193         # These fields hold the id of the video
 194         videoPlayer = find_param('@videoPlayer') or find_param('videoId') or find_param('videoID') or find_param('@videoList')
 195         if videoPlayer is not None:
 196             if isinstance(videoPlayer, list):
 197                 videoPlayer = videoPlayer[0]
 198             videoPlayer = videoPlayer.strip()
 199             # UUID is also possible for videoPlayer (e.g.
 200             # http://www.popcornflix.com/hoodies-vs-hooligans/7f2d2b87-bbf2-4623-acfb-ea942b4f01dd
 201             # or http://www8.hp.com/cn/zh/home.html)
 202             if not (re.match(
 203                     r'^(?:\d+|[\da-fA-F]{8}-?[\da-fA-F]{4}-?[\da-fA-F]{4}-?[\da-fA-F]{4}-?[\da-fA-F]{12})$',
 204                     videoPlayer) or videoPlayer.startswith('ref:')):
 205                 return None
 206             params['@videoPlayer'] = videoPlayer
 207         linkBase = find_param('linkBaseURL')
 208         if linkBase is not None:
 209             params['linkBaseURL'] = linkBase
 210         return cls._make_brightcove_url(params)
 211
 212     @classmethod
 213     def _build_brighcove_url_from_js(cls, object_js):
 214         # The layout of JS is as follows:
 215         # customBC.createVideo = function (width, height, playerID, playerKey, videoPlayer, VideoRandomID) {
 216         #   // build Brightcove <object /> XML
 217         # }
 218         m = re.search(
 219             r'''(?x)customBC\.createVideo\(
 220                 .*?                                                  # skipping width and height
 221                 ["\'](?P<playerID>\d+)["\']\s*,\s*                   # playerID
 222                 ["\'](?P<playerKey>AQ[^"\']{48})[^"\']*["\']\s*,\s*  # playerKey begins with AQ and is 50 characters
 223                                                                      # in length, however it's appended to itself
 224                                                                      # in places, so truncate
 225                 ["\'](?P<videoID>\d+)["\']                           # @videoPlayer
 226             ''', object_js)
 227         if m:
 228             return cls._make_brightcove_url(m.groupdict())
 229
 230     @classmethod
 231     def _make_brightcove_url(cls, params):
 232         return update_url_query(cls._FEDERATED_URL, params)
 233
 234     @classmethod
 235     def _extract_brightcove_url(cls, webpage):
 236         """Try to extract the brightcove url from the webpage, returns None
 237         if it can't be found
 238         """
 239         urls = cls._extract_brightcove_urls(webpage)
 240         return urls[0] if urls else None
 241
 242     @classmethod
 243     def _extract_brightcove_urls(cls, webpage):
 244         """Return a list of all Brightcove URLs from the webpage """
 245
 246         url_m = re.search(
 247             r'''(?x)
 248                 <meta\s+
 249                     (?:property|itemprop)=([\'"])(?:og:video|embedURL)\1[^>]+
 250                     content=([\'"])(?P<url>https?://(?:secure|c)\.brightcove.com/(?:(?!\2).)+)\2
 251             ''', webpage)
 252         if url_m:
 253             url = unescapeHTML(url_m.group('url'))
 254             # Some sites don't add it, we can't download with this url, for example:
 255             # http://www.ktvu.com/videos/news/raw-video-caltrain-releases-video-of-man-almost/vCTZdY/
 256             if 'playerKey' in url or 'videoId' in url or 'idVideo' in url:
 257                 return [url]
 258
 259         matches = re.findall(
 260             r'''(?sx)<object
 261             (?:
 262                 [^>]+?class=[\'"][^>]*?BrightcoveExperience.*?[\'"] |
 263                 [^>]*?>\s*<param\s+name="movie"\s+value="https?://[^/]*brightcove\.com/
 264             ).+?>\s*</object>''',
 265             webpage)
 266         if matches:
 267             return list(filter(None, [cls._build_brighcove_url(m) for m in matches]))
 268
 269         return list(filter(None, [
 270             cls._build_brighcove_url_from_js(custom_bc)
 271             for custom_bc in re.findall(r'(customBC\.createVideo\(.+?\);)', webpage)]))
 272
 273     def _real_extract(self, url):
 274         url, smuggled_data = unsmuggle_url(url, {})
 275
 276         # Change the 'videoId' and others field to '@videoPlayer'
 277         url = re.sub(r'(?<=[?&])(videoI(d|D)|idVideo|bctid)', '%40videoPlayer', url)
 278         # Change bckey (used by bcove.me urls) to playerKey
 279         url = re.sub(r'(?<=[?&])bckey', 'playerKey', url)
 280         mobj = re.match(self._VALID_URL, url)
 281         query_str = mobj.group('query')
 282         query = compat_urlparse.parse_qs(query_str)
 283
 284         videoPlayer = query.get('@videoPlayer')
 285         if videoPlayer:
 286             # We set the original url as the default 'Referer' header
 287             referer = smuggled_data.get('Referer', url)
 288             return self._get_video_info(
 289                 videoPlayer[0], query, referer=referer)
 290         elif 'playerKey' in query:
 291             player_key = query['playerKey']
 292             return self._get_playlist_info(player_key[0])
 293         else:
 294             raise ExtractorError(
 295                 'Cannot find playerKey= variable. Did you forget quotes in a shell invocation?',
 296                 expected=True)
 297
 298     def _get_video_info(self, video_id, query, referer=None):
 299         headers = {}
 300         linkBase = query.get('linkBaseURL')
 301         if linkBase is not None:
 302             referer = linkBase[0]
 303         if referer is not None:
 304             headers['Referer'] = referer
 305         webpage = self._download_webpage(self._FEDERATED_URL, video_id, headers=headers, query=query)
 306
 307         error_msg = self._html_search_regex(
 308             r"<h1>We're sorry.</h1>([\s\n]*<p>.*?</p>)+", webpage,
 309             'error message', default=None)
 310         if error_msg is not None:
 311             raise ExtractorError(
 312                 'brightcove said: %s' % error_msg, expected=True)
 313
 314         self.report_extraction(video_id)
 315         info = self._search_regex(r'var experienceJSON = ({.*});', webpage, 'json')
 316         info = json.loads(info)['data']
 317         video_info = info['programmedContent']['videoPlayer']['mediaDTO']
 318         video_info['_youtubedl_adServerURL'] = info.get('adServerURL')
 319
 320         return self._extract_video_info(video_info)
 321
 322     def _get_playlist_info(self, player_key):
 323         info_url = 'http://c.brightcove.com/services/json/experience/runtime/?command=get_programming_for_experience&playerKey=%s' % player_key
 324         playlist_info = self._download_webpage(
 325             info_url, player_key, 'Downloading playlist information')
 326
 327         json_data = json.loads(playlist_info)
 328         if 'videoList' in json_data:
 329             playlist_info = json_data['videoList']
 330             playlist_dto = playlist_info['mediaCollectionDTO']
 331         elif 'playlistTabs' in json_data:
 332             playlist_info = json_data['playlistTabs']
 333             playlist_dto = playlist_info['lineupListDTO']['playlistDTOs'][0]
 334         else:
 335             raise ExtractorError('Empty playlist')
 336
 337         videos = [self._extract_video_info(video_info) for video_info in playlist_dto['videoDTOs']]
 338
 339         return self.playlist_result(videos, playlist_id='%s' % playlist_info['id'],
 340                                     playlist_title=playlist_dto['displayName'])
 341
 342     def _extract_video_info(self, video_info):
 343         video_id = compat_str(video_info['id'])
 344         publisher_id = video_info.get('publisherId')
 345         info = {
 346             'id': video_id,
 347             'title': video_info['displayName'].strip(),
 348             'description': video_info.get('shortDescription'),
 349             'thumbnail': video_info.get('videoStillURL') or video_info.get('thumbnailURL'),
 350             'uploader': video_info.get('publisherName'),
 351             'uploader_id': compat_str(publisher_id) if publisher_id else None,
 352             'duration': float_or_none(video_info.get('length'), 1000),
 353             'timestamp': int_or_none(video_info.get('creationDate'), 1000),
 354         }
 355
 356         renditions = video_info.get('renditions', []) + video_info.get('IOSRenditions', [])
 357         if renditions:
 358             formats = []
 359             for rend in renditions:
 360                 url = rend['defaultURL']
 361                 if not url:
 362                     continue
 363                 ext = None
 364                 if rend['remote']:
 365                     url_comp = compat_urllib_parse_urlparse(url)
 366                     if url_comp.path.endswith('.m3u8'):
 367                         formats.extend(
 368                             self._extract_m3u8_formats(
 369                                 url, video_id, 'mp4', 'm3u8_native', m3u8_id='hls', fatal=False))
 370                         continue
 371                     elif 'akamaihd.net' in url_comp.netloc:
 372                         # This type of renditions are served through
 373                         # akamaihd.net, but they don't use f4m manifests
 374                         url = url.replace('control/', '') + '?&v=3.3.0&fp=13&r=FEEFJ&g=RTSJIMBMPFPB'
 375                         ext = 'flv'
 376                 if ext is None:
 377                     ext = determine_ext(url)
 378                 tbr = int_or_none(rend.get('encodingRate'), 1000)
 379                 a_format = {
 380                     'format_id': 'http%s' % ('-%s' % tbr if tbr else ''),
 381                     'url': url,
 382                     'ext': ext,
 383                     'filesize': int_or_none(rend.get('size')) or None,
 384                     'tbr': tbr,
 385                 }
 386                 if rend.get('audioOnly'):
 387                     a_format.update({
 388                         'vcodec': 'none',
 389                     })
 390                 else:
 391                     a_format.update({
 392                         'height': int_or_none(rend.get('frameHeight')),
 393                         'width': int_or_none(rend.get('frameWidth')),
 394                         'vcodec': rend.get('videoCodec'),
 395                     })
 396
 397                 # m3u8 manifests with remote == false are media playlists
 398                 # Not calling _extract_m3u8_formats here to save network traffic
 399                 if ext == 'm3u8':
 400                     a_format.update({
 401                         'format_id': 'hls%s' % ('-%s' % tbr if tbr else ''),
 402                         'ext': 'mp4',
 403                         'protocol': 'm3u8_native',
 404                     })
 405
 406                 formats.append(a_format)
 407             self._sort_formats(formats)
 408             info['formats'] = formats
 409         elif video_info.get('FLVFullLengthURL') is not None:
 410             info.update({
 411                 'url': video_info['FLVFullLengthURL'],
 412                 'vcodec': self.FLV_VCODECS.get(video_info.get('FLVFullCodec')),
 413                 'filesize': int_or_none(video_info.get('FLVFullSize')),
 414             })
 415
 416         if self._downloader.params.get('include_ads', False):
 417             adServerURL = video_info.get('_youtubedl_adServerURL')
 418             if adServerURL:
 419                 ad_info = {
 420                     '_type': 'url',
 421                     'url': adServerURL,
 422                 }
 423                 if 'url' in info:
 424                     return {
 425                         '_type': 'playlist',
 426                         'title': info['title'],
 427                         'entries': [ad_info, info],
 428                     }
 429                 else:
 430                     return ad_info
 431
 432         if 'url' not in info and not info.get('formats'):
 433             raise ExtractorError('Unable to extract video url for %s' % video_id)
 434         return info
 435
 436
 437 class BrightcoveNewIE(InfoExtractor):
 438     IE_NAME = 'brightcove:new'
 439     _VALID_URL = r'https?://players\.brightcove\.net/(?P<account_id>\d+)/(?P<player_id>[^/]+)_(?P<embed>[^/]+)/index\.html\?.*videoId=(?P<video_id>\d+|ref:[^&]+)'
 440     _TESTS = [{
 441         'url': 'http://players.brightcove.net/929656772001/e41d32dc-ec74-459e-a845-6c69f7b724ea_default/index.html?videoId=4463358922001',
 442         'md5': 'c8100925723840d4b0d243f7025703be',
 443         'info_dict': {
 444             'id': '4463358922001',
 445             'ext': 'mp4',
 446             'title': 'Meet the man behind Popcorn Time',
 447             'description': 'md5:eac376a4fe366edc70279bfb681aea16',
 448             'duration': 165.768,
 449             'timestamp': 1441391203,
 450             'upload_date': '20150904',
 451             'uploader_id': '929656772001',
 452             'formats': 'mincount:22',
 453         },
 454     }, {
 455         # with rtmp streams
 456         'url': 'http://players.brightcove.net/4036320279001/5d112ed9-283f-485f-a7f9-33f42e8bc042_default/index.html?videoId=4279049078001',
 457         'info_dict': {
 458             'id': '4279049078001',
 459             'ext': 'mp4',
 460             'title': 'Titansgrave: Chapter 0',
 461             'description': 'Titansgrave: Chapter 0',
 462             'duration': 1242.058,
 463             'timestamp': 1433556729,
 464             'upload_date': '20150606',
 465             'uploader_id': '4036320279001',
 466             'formats': 'mincount:41',
 467         },
 468         'params': {
 469             # m3u8 download
 470             'skip_download': True,
 471         }
 472     }, {
 473         # ref: prefixed video id
 474         'url': 'http://players.brightcove.net/3910869709001/21519b5c-4b3b-4363-accb-bdc8f358f823_default/index.html?videoId=ref:7069442',
 475         'only_matching': True,
 476     }, {
 477         # non numeric ref: prefixed video id
 478         'url': 'http://players.brightcove.net/710858724001/default_default/index.html?videoId=ref:event-stream-356',
 479         'only_matching': True,
 480     }, {
 481         # unavailable video without message but with error_code
 482         'url': 'http://players.brightcove.net/1305187701/c832abfb-641b-44eb-9da0-2fe76786505f_default/index.html?videoId=4377407326001',
 483         'only_matching': True,
 484     }]
 485
 486     @staticmethod
 487     def _extract_url(webpage):
 488         urls = BrightcoveNewIE._extract_urls(webpage)
 489         return urls[0] if urls else None
 490
 491     @staticmethod
 492     def _extract_urls(webpage):
 493         # Reference:
 494         # 1. http://docs.brightcove.com/en/video-cloud/brightcove-player/guides/publish-video.html#setvideoiniframe
 495         # 2. http://docs.brightcove.com/en/video-cloud/brightcove-player/guides/publish-video.html#tag
 496         # 3. http://docs.brightcove.com/en/video-cloud/brightcove-player/guides/publish-video.html#setvideousingjavascript
 497         # 4. http://docs.brightcove.com/en/video-cloud/brightcove-player/guides/in-page-embed-player-implementation.html
 498         # 5. https://support.brightcove.com/en/video-cloud/docs/dynamically-assigning-videos-player
 499
 500         entries = []
 501
 502         # Look for iframe embeds [1]
 503         for _, url in re.findall(
 504                 r'<iframe[^>]+src=(["\'])((?:https?:)?//players\.brightcove\.net/\d+/[^/]+/index\.html.+?)\1', webpage):
 505             entries.append(url if url.startswith('http') else 'http:' + url)
 506
 507         # Look for <video> tags [2] and embed_in_page embeds [3]
 508         # [2] looks like:
 509         # <video data-video-id="5320421710001" data-account="245991542" data-player="SJWAiyYWg" data-embed="default" class="video-js" controls itemscope itemtype="http://schema.org/VideoObject">
 510
 511         for video, script_tag, account_id, player_id, embed in re.findall(
 512             r'''(?isx)
 513                 (<video[^>]+>)
 514                 (?:.*?
 515                   (<script[^>]+
 516                     src=["\'](?:https?:)?//players\.brightcove\.net/
 517                     (\d+)/([^/]+)_([^/]+)/index(?:\.min)?\.js
 518                   )
 519                 )?
 520             ''', webpage
 521         ):
 522             attrs = extract_attributes(video)
 523
 524             # According to examples from [4] it's unclear whether video id
 525             # may be optional and what to do when it is
 526             video_id = attrs.get('data-video-id')
 527             # See PR#12099/bostonglobe.py for 'data-brightcove-video-id' variant
 528
 529             if not account_id:
 530                 account_id = attrs.get('data-account')
 531             if not player_id:
 532                 player_id = attrs.get('data-player')
 533             if not embed:
 534                 embed = attrs.get('data-embed')
 535
 536             if video_id and account_id and player_id and embed:
 537                 entries.append(
 538                     'http://players.brightcove.net/%s/%s_%s/index.html?videoId=%s'
 539                     % (account_id, player_id, embed, video_id))
 540
 541         return entries
 542
 543     def _real_extract(self, url):
 544         url, smuggled_data = unsmuggle_url(url, {})
 545         self._initialize_geo_bypass(smuggled_data.get('geo_countries'))
 546
 547         account_id, player_id, embed, video_id = re.match(self._VALID_URL, url).groups()
 548
 549         webpage = self._download_webpage(
 550             'http://players.brightcove.net/%s/%s_%s/index.min.js'
 551             % (account_id, player_id, embed), video_id)
 552
 553         policy_key = None
 554
 555         catalog = self._search_regex(
 556             r'catalog\(({.+?})\);', webpage, 'catalog', default=None)
 557         if catalog:
 558             catalog = self._parse_json(
 559                 js_to_json(catalog), video_id, fatal=False)
 560             if catalog:
 561                 policy_key = catalog.get('policyKey')
 562
 563         if not policy_key:
 564             policy_key = self._search_regex(
 565                 r'policyKey\s*:\s*(["\'])(?P<pk>.+?)\1',
 566                 webpage, 'policy key', group='pk')
 567
 568         api_url = 'https://edge.api.brightcove.com/playback/v1/accounts/%s/videos/%s' % (account_id, video_id)
 569         try:
 570             json_data = self._download_json(api_url, video_id, headers={
 571                 'Accept': 'application/json;pk=%s' % policy_key
 572             })
 573         except ExtractorError as e:
 574             if isinstance(e.cause, compat_HTTPError) and e.cause.code == 403:
 575                 json_data = self._parse_json(e.cause.read().decode(), video_id)[0]
 576                 message = json_data.get('message') or json_data['error_code']
 577                 if json_data.get('error_subcode') == 'CLIENT_GEO':
 578                     self.raise_geo_restricted(msg=message)
 579                 raise ExtractorError(message, expected=True)
 580             raise
 581
 582         title = json_data['name'].strip()
 583
 584         formats = []
 585         for source in json_data.get('sources', []):
 586             container = source.get('container')
 587             ext = mimetype2ext(source.get('type'))
 588             src = source.get('src')
 589             if ext == 'ism' or container == 'WVM':
 590                 continue
 591             elif ext == 'm3u8' or container == 'M2TS':
 592                 if not src:
 593                     continue
 594                 formats.extend(self._extract_m3u8_formats(
 595                     src, video_id, 'mp4', 'm3u8_native', m3u8_id='hls', fatal=False))
 596             elif ext == 'mpd':
 597                 if not src:
 598                     continue
 599                 formats.extend(self._extract_mpd_formats(src, video_id, 'dash', fatal=False))
 600             else:
 601                 streaming_src = source.get('streaming_src')
 602                 stream_name, app_name = source.get('stream_name'), source.get('app_name')
 603                 if not src and not streaming_src and (not stream_name or not app_name):
 604                     continue
 605                 tbr = float_or_none(source.get('avg_bitrate'), 1000)
 606                 height = int_or_none(source.get('height'))
 607                 width = int_or_none(source.get('width'))
 608                 f = {
 609                     'tbr': tbr,
 610                     'filesize': int_or_none(source.get('size')),
 611                     'container': container,
 612                     'ext': ext or container.lower(),
 613                 }
 614                 if width == 0 and height == 0:
 615                     f.update({
 616                         'vcodec': 'none',
 617                     })
 618                 else:
 619                     f.update({
 620                         'width': width,
 621                         'height': height,
 622                         'vcodec': source.get('codec'),
 623                     })
 624
 625                 def build_format_id(kind):
 626                     format_id = kind
 627                     if tbr:
 628                         format_id += '-%dk' % int(tbr)
 629                     if height:
 630                         format_id += '-%dp' % height
 631                     return format_id
 632
 633                 if src or streaming_src:
 634                     f.update({
 635                         'url': src or streaming_src,
 636                         'format_id': build_format_id('http' if src else 'http-streaming'),
 637                         'source_preference': 0 if src else -1,
 638                     })
 639                 else:
 640                     f.update({
 641                         'url': app_name,
 642                         'play_path': stream_name,
 643                         'format_id': build_format_id('rtmp'),
 644                     })
 645                 formats.append(f)
 646
 647         errors = json_data.get('errors')
 648         if not formats and errors:
 649             error = errors[0]
 650             raise ExtractorError(
 651                 error.get('message') or error.get('error_subcode') or error['error_code'], expected=True)
 652
 653         self._sort_formats(formats)
 654
 655         subtitles = {}
 656         for text_track in json_data.get('text_tracks', []):
 657             if text_track.get('src'):
 658                 subtitles.setdefault(text_track.get('srclang'), []).append({
 659                     'url': text_track['src'],
 660                 })
 661
 662         is_live = False
 663         duration = float_or_none(json_data.get('duration'), 1000)
 664         if duration and duration < 0:
 665             is_live = True
 666
 667         return {
 668             'id': video_id,
 669             'title': self._live_title(title) if is_live else title,
 670             'description': clean_html(json_data.get('description')),
 671             'thumbnail': json_data.get('thumbnail') or json_data.get('poster'),
 672             'duration': duration,
 673             'timestamp': parse_iso8601(json_data.get('published_at')),
 674             'uploader_id': account_id,
 675             'formats': formats,
 676             'subtitles': subtitles,
 677             'tags': json_data.get('tags', []),
 678             'is_live': is_live,
 679         }