_ 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             # playerID inferred from bcpid
 136             # from http://www.un.org/chinese/News/story.asp?NewsID=27724
 137             'url': 'https://link.brightcove.com/services/player/bcpid1722935254001/?bctid=5360463607001&autoStart=false&secureConnections=true&width=650&height=350',
 138             'only_matching': True,  # Tested in GenericIE
 139         }
 140     ]
 141     FLV_VCODECS = {
 142         1: 'SORENSON',
 143         2: 'ON2',
 144         3: 'H264',
 145         4: 'VP8',
 146     }
 147
 148     @classmethod
 149     def _build_brighcove_url(cls, object_str):
 150         """
 151         Build a Brightcove url from a xml string containing
 152         <object class="BrightcoveExperience">{params}</object>
 153         """
 154
 155         # Fix up some stupid HTML, see https://github.com/rg3/youtube-dl/issues/1553
 156         object_str = re.sub(r'(<param(?:\s+[a-zA-Z0-9_]+="[^"]*")*)>',
 157                             lambda m: m.group(1) + '/>', object_str)
 158         # Fix up some stupid XML, see https://github.com/rg3/youtube-dl/issues/1608
 159         object_str = object_str.replace('<--', '<!--')
 160         # remove namespace to simplify extraction
 161         object_str = re.sub(r'(<object[^>]*)(xmlns=".*?")', r'\1', object_str)
 162         object_str = fix_xml_ampersands(object_str)
 163
 164         try:
 165             object_doc = compat_etree_fromstring(object_str.encode('utf-8'))
 166         except compat_xml_parse_error:
 167             return
 168
 169         fv_el = find_xpath_attr(object_doc, './param', 'name', 'flashVars')
 170         if fv_el is not None:
 171             flashvars = dict(
 172                 (k, v[0])
 173                 for k, v in compat_parse_qs(fv_el.attrib['value']).items())
 174         else:
 175             flashvars = {}
 176
 177         data_url = object_doc.attrib.get('data', '')
 178         data_url_params = compat_parse_qs(compat_urllib_parse_urlparse(data_url).query)
 179
 180         def find_param(name):
 181             if name in flashvars:
 182                 return flashvars[name]
 183             node = find_xpath_attr(object_doc, './param', 'name', name)
 184             if node is not None:
 185                 return node.attrib['value']
 186             return data_url_params.get(name)
 187
 188         params = {}
 189
 190         playerID = find_param('playerID') or find_param('playerId')
 191         if playerID is None:
 192             raise ExtractorError('Cannot find player ID')
 193         params['playerID'] = playerID
 194
 195         playerKey = find_param('playerKey')
 196         # Not all pages define this value
 197         if playerKey is not None:
 198             params['playerKey'] = playerKey
 199         # These fields hold the id of the video
 200         videoPlayer = find_param('@videoPlayer') or find_param('videoId') or find_param('videoID') or find_param('@videoList')
 201         if videoPlayer is not None:
 202             if isinstance(videoPlayer, list):
 203                 videoPlayer = videoPlayer[0]
 204             videoPlayer = videoPlayer.strip()
 205             # UUID is also possible for videoPlayer (e.g.
 206             # http://www.popcornflix.com/hoodies-vs-hooligans/7f2d2b87-bbf2-4623-acfb-ea942b4f01dd
 207             # or http://www8.hp.com/cn/zh/home.html)
 208             if not (re.match(
 209                     r'^(?:\d+|[\da-fA-F]{8}-?[\da-fA-F]{4}-?[\da-fA-F]{4}-?[\da-fA-F]{4}-?[\da-fA-F]{12})$',
 210                     videoPlayer) or videoPlayer.startswith('ref:')):
 211                 return None
 212             params['@videoPlayer'] = videoPlayer
 213         linkBase = find_param('linkBaseURL')
 214         if linkBase is not None:
 215             params['linkBaseURL'] = linkBase
 216         return cls._make_brightcove_url(params)
 217
 218     @classmethod
 219     def _build_brighcove_url_from_js(cls, object_js):
 220         # The layout of JS is as follows:
 221         # customBC.createVideo = function (width, height, playerID, playerKey, videoPlayer, VideoRandomID) {
 222         #   // build Brightcove <object /> XML
 223         # }
 224         m = re.search(
 225             r'''(?x)customBC\.createVideo\(
 226                 .*?                                                  # skipping width and height
 227                 ["\'](?P<playerID>\d+)["\']\s*,\s*                   # playerID
 228                 ["\'](?P<playerKey>AQ[^"\']{48})[^"\']*["\']\s*,\s*  # playerKey begins with AQ and is 50 characters
 229                                                                      # in length, however it's appended to itself
 230                                                                      # in places, so truncate
 231                 ["\'](?P<videoID>\d+)["\']                           # @videoPlayer
 232             ''', object_js)
 233         if m:
 234             return cls._make_brightcove_url(m.groupdict())
 235
 236     @classmethod
 237     def _make_brightcove_url(cls, params):
 238         return update_url_query(cls._FEDERATED_URL, params)
 239
 240     @classmethod
 241     def _extract_brightcove_url(cls, webpage):
 242         """Try to extract the brightcove url from the webpage, returns None
 243         if it can't be found
 244         """
 245         urls = cls._extract_brightcove_urls(webpage)
 246         return urls[0] if urls else None
 247
 248     @classmethod
 249     def _extract_brightcove_urls(cls, webpage):
 250         """Return a list of all Brightcove URLs from the webpage """
 251
 252         url_m = re.search(
 253             r'''(?x)
 254                 <meta\s+
 255                     (?:property|itemprop)=([\'"])(?:og:video|embedURL)\1[^>]+
 256                     content=([\'"])(?P<url>https?://(?:secure|c)\.brightcove.com/(?:(?!\2).)+)\2
 257             ''', webpage)
 258         if url_m:
 259             url = unescapeHTML(url_m.group('url'))
 260             # Some sites don't add it, we can't download with this url, for example:
 261             # http://www.ktvu.com/videos/news/raw-video-caltrain-releases-video-of-man-almost/vCTZdY/
 262             if 'playerKey' in url or 'videoId' in url or 'idVideo' in url:
 263                 return [url]
 264
 265         matches = re.findall(
 266             r'''(?sx)<object
 267             (?:
 268                 [^>]+?class=[\'"][^>]*?BrightcoveExperience.*?[\'"] |
 269                 [^>]*?>\s*<param\s+name="movie"\s+value="https?://[^/]*brightcove\.com/
 270             ).+?>\s*</object>''',
 271             webpage)
 272         if matches:
 273             return list(filter(None, [cls._build_brighcove_url(m) for m in matches]))
 274
 275         matches = re.findall(r'(customBC\.createVideo\(.+?\);)', webpage)
 276         if matches:
 277             return list(filter(None, [
 278                 cls._build_brighcove_url_from_js(custom_bc)
 279                 for custom_bc in matches]))
 280         return [src for _, src in re.findall(
 281             r'<iframe[^>]+src=([\'"])((?:https?:)?//link\.brightcove\.com/services/player/(?!\1).+)\1', webpage)]
 282
 283     def _real_extract(self, url):
 284         url, smuggled_data = unsmuggle_url(url, {})
 285
 286         # Change the 'videoId' and others field to '@videoPlayer'
 287         url = re.sub(r'(?<=[?&])(videoI(d|D)|idVideo|bctid)', '%40videoPlayer', url)
 288         # Change bckey (used by bcove.me urls) to playerKey
 289         url = re.sub(r'(?<=[?&])bckey', 'playerKey', url)
 290         mobj = re.match(self._VALID_URL, url)
 291         query_str = mobj.group('query')
 292         query = compat_urlparse.parse_qs(query_str)
 293
 294         videoPlayer = query.get('@videoPlayer')
 295         if videoPlayer:
 296             # We set the original url as the default 'Referer' header
 297             referer = smuggled_data.get('Referer', url)
 298             if 'playerID' not in query:
 299                 mobj = re.search(r'/bcpid(\d+)', url)
 300                 if mobj is not None:
 301                     query['playerID'] = [mobj.group(1)]
 302             return self._get_video_info(
 303                 videoPlayer[0], query, referer=referer)
 304         elif 'playerKey' in query:
 305             player_key = query['playerKey']
 306             return self._get_playlist_info(player_key[0])
 307         else:
 308             raise ExtractorError(
 309                 'Cannot find playerKey= variable. Did you forget quotes in a shell invocation?',
 310                 expected=True)
 311
 312     def _get_video_info(self, video_id, query, referer=None):
 313         headers = {}
 314         linkBase = query.get('linkBaseURL')
 315         if linkBase is not None:
 316             referer = linkBase[0]
 317         if referer is not None:
 318             headers['Referer'] = referer
 319         webpage = self._download_webpage(self._FEDERATED_URL, video_id, headers=headers, query=query)
 320
 321         error_msg = self._html_search_regex(
 322             r"<h1>We're sorry.</h1>([\s\n]*<p>.*?</p>)+", webpage,
 323             'error message', default=None)
 324         if error_msg is not None:
 325             raise ExtractorError(
 326                 'brightcove said: %s' % error_msg, expected=True)
 327
 328         self.report_extraction(video_id)
 329         info = self._search_regex(r'var experienceJSON = ({.*});', webpage, 'json')
 330         info = json.loads(info)['data']
 331         video_info = info['programmedContent']['videoPlayer']['mediaDTO']
 332         video_info['_youtubedl_adServerURL'] = info.get('adServerURL')
 333
 334         return self._extract_video_info(video_info)
 335
 336     def _get_playlist_info(self, player_key):
 337         info_url = 'http://c.brightcove.com/services/json/experience/runtime/?command=get_programming_for_experience&playerKey=%s' % player_key
 338         playlist_info = self._download_webpage(
 339             info_url, player_key, 'Downloading playlist information')
 340
 341         json_data = json.loads(playlist_info)
 342         if 'videoList' in json_data:
 343             playlist_info = json_data['videoList']
 344             playlist_dto = playlist_info['mediaCollectionDTO']
 345         elif 'playlistTabs' in json_data:
 346             playlist_info = json_data['playlistTabs']
 347             playlist_dto = playlist_info['lineupListDTO']['playlistDTOs'][0]
 348         else:
 349             raise ExtractorError('Empty playlist')
 350
 351         videos = [self._extract_video_info(video_info) for video_info in playlist_dto['videoDTOs']]
 352
 353         return self.playlist_result(videos, playlist_id='%s' % playlist_info['id'],
 354                                     playlist_title=playlist_dto['displayName'])
 355
 356     def _extract_video_info(self, video_info):
 357         video_id = compat_str(video_info['id'])
 358         publisher_id = video_info.get('publisherId')
 359         info = {
 360             'id': video_id,
 361             'title': video_info['displayName'].strip(),
 362             'description': video_info.get('shortDescription'),
 363             'thumbnail': video_info.get('videoStillURL') or video_info.get('thumbnailURL'),
 364             'uploader': video_info.get('publisherName'),
 365             'uploader_id': compat_str(publisher_id) if publisher_id else None,
 366             'duration': float_or_none(video_info.get('length'), 1000),
 367             'timestamp': int_or_none(video_info.get('creationDate'), 1000),
 368         }
 369
 370         renditions = video_info.get('renditions', []) + video_info.get('IOSRenditions', [])
 371         if renditions:
 372             formats = []
 373             for rend in renditions:
 374                 url = rend['defaultURL']
 375                 if not url:
 376                     continue
 377                 ext = None
 378                 if rend['remote']:
 379                     url_comp = compat_urllib_parse_urlparse(url)
 380                     if url_comp.path.endswith('.m3u8'):
 381                         formats.extend(
 382                             self._extract_m3u8_formats(
 383                                 url, video_id, 'mp4', 'm3u8_native', m3u8_id='hls', fatal=False))
 384                         continue
 385                     elif 'akamaihd.net' in url_comp.netloc:
 386                         # This type of renditions are served through
 387                         # akamaihd.net, but they don't use f4m manifests
 388                         url = url.replace('control/', '') + '?&v=3.3.0&fp=13&r=FEEFJ&g=RTSJIMBMPFPB'
 389                         ext = 'flv'
 390                 if ext is None:
 391                     ext = determine_ext(url)
 392                 tbr = int_or_none(rend.get('encodingRate'), 1000)
 393                 a_format = {
 394                     'format_id': 'http%s' % ('-%s' % tbr if tbr else ''),
 395                     'url': url,
 396                     'ext': ext,
 397                     'filesize': int_or_none(rend.get('size')) or None,
 398                     'tbr': tbr,
 399                 }
 400                 if rend.get('audioOnly'):
 401                     a_format.update({
 402                         'vcodec': 'none',
 403                     })
 404                 else:
 405                     a_format.update({
 406                         'height': int_or_none(rend.get('frameHeight')),
 407                         'width': int_or_none(rend.get('frameWidth')),
 408                         'vcodec': rend.get('videoCodec'),
 409                     })
 410
 411                 # m3u8 manifests with remote == false are media playlists
 412                 # Not calling _extract_m3u8_formats here to save network traffic
 413                 if ext == 'm3u8':
 414                     a_format.update({
 415                         'format_id': 'hls%s' % ('-%s' % tbr if tbr else ''),
 416                         'ext': 'mp4',
 417                         'protocol': 'm3u8_native',
 418                     })
 419
 420                 formats.append(a_format)
 421             self._sort_formats(formats)
 422             info['formats'] = formats
 423         elif video_info.get('FLVFullLengthURL') is not None:
 424             info.update({
 425                 'url': video_info['FLVFullLengthURL'],
 426                 'vcodec': self.FLV_VCODECS.get(video_info.get('FLVFullCodec')),
 427                 'filesize': int_or_none(video_info.get('FLVFullSize')),
 428             })
 429
 430         if self._downloader.params.get('include_ads', False):
 431             adServerURL = video_info.get('_youtubedl_adServerURL')
 432             if adServerURL:
 433                 ad_info = {
 434                     '_type': 'url',
 435                     'url': adServerURL,
 436                 }
 437                 if 'url' in info:
 438                     return {
 439                         '_type': 'playlist',
 440                         'title': info['title'],
 441                         'entries': [ad_info, info],
 442                     }
 443                 else:
 444                     return ad_info
 445
 446         if 'url' not in info and not info.get('formats'):
 447             raise ExtractorError('Unable to extract video url for %s' % video_id)
 448         return info
 449
 450
 451 class BrightcoveNewIE(InfoExtractor):
 452     IE_NAME = 'brightcove:new'
 453     _VALID_URL = r'https?://players\.brightcove\.net/(?P<account_id>\d+)/(?P<player_id>[^/]+)_(?P<embed>[^/]+)/index\.html\?.*videoId=(?P<video_id>\d+|ref:[^&]+)'
 454     _TESTS = [{
 455         'url': 'http://players.brightcove.net/929656772001/e41d32dc-ec74-459e-a845-6c69f7b724ea_default/index.html?videoId=4463358922001',
 456         'md5': 'c8100925723840d4b0d243f7025703be',
 457         'info_dict': {
 458             'id': '4463358922001',
 459             'ext': 'mp4',
 460             'title': 'Meet the man behind Popcorn Time',
 461             'description': 'md5:eac376a4fe366edc70279bfb681aea16',
 462             'duration': 165.768,
 463             'timestamp': 1441391203,
 464             'upload_date': '20150904',
 465             'uploader_id': '929656772001',
 466             'formats': 'mincount:22',
 467         },
 468     }, {
 469         # with rtmp streams
 470         'url': 'http://players.brightcove.net/4036320279001/5d112ed9-283f-485f-a7f9-33f42e8bc042_default/index.html?videoId=4279049078001',
 471         'info_dict': {
 472             'id': '4279049078001',
 473             'ext': 'mp4',
 474             'title': 'Titansgrave: Chapter 0',
 475             'description': 'Titansgrave: Chapter 0',
 476             'duration': 1242.058,
 477             'timestamp': 1433556729,
 478             'upload_date': '20150606',
 479             'uploader_id': '4036320279001',
 480             'formats': 'mincount:41',
 481         },
 482         'params': {
 483             # m3u8 download
 484             'skip_download': True,
 485         }
 486     }, {
 487         # ref: prefixed video id
 488         'url': 'http://players.brightcove.net/3910869709001/21519b5c-4b3b-4363-accb-bdc8f358f823_default/index.html?videoId=ref:7069442',
 489         'only_matching': True,
 490     }, {
 491         # non numeric ref: prefixed video id
 492         'url': 'http://players.brightcove.net/710858724001/default_default/index.html?videoId=ref:event-stream-356',
 493         'only_matching': True,
 494     }, {
 495         # unavailable video without message but with error_code
 496         'url': 'http://players.brightcove.net/1305187701/c832abfb-641b-44eb-9da0-2fe76786505f_default/index.html?videoId=4377407326001',
 497         'only_matching': True,
 498     }]
 499
 500     @staticmethod
 501     def _extract_url(ie, webpage):
 502         urls = BrightcoveNewIE._extract_urls(ie, webpage)
 503         return urls[0] if urls else None
 504
 505     @staticmethod
 506     def _extract_urls(ie, webpage):
 507         # Reference:
 508         # 1. http://docs.brightcove.com/en/video-cloud/brightcove-player/guides/publish-video.html#setvideoiniframe
 509         # 2. http://docs.brightcove.com/en/video-cloud/brightcove-player/guides/publish-video.html#tag
 510         # 3. http://docs.brightcove.com/en/video-cloud/brightcove-player/guides/publish-video.html#setvideousingjavascript
 511         # 4. http://docs.brightcove.com/en/video-cloud/brightcove-player/guides/in-page-embed-player-implementation.html
 512         # 5. https://support.brightcove.com/en/video-cloud/docs/dynamically-assigning-videos-player
 513
 514         entries = []
 515
 516         # Look for iframe embeds [1]
 517         for _, url in re.findall(
 518                 r'<iframe[^>]+src=(["\'])((?:https?:)?//players\.brightcove\.net/\d+/[^/]+/index\.html.+?)\1', webpage):
 519             entries.append(url if url.startswith('http') else 'http:' + url)
 520
 521         # Look for <video> tags [2] and embed_in_page embeds [3]
 522         # [2] looks like:
 523         for video, script_tag, account_id, player_id, embed in re.findall(
 524                 r'''(?isx)
 525                     (<video\s+[^>]*data-video-id=['"]?[^>]+>)
 526                     (?:.*?
 527                         (<script[^>]+
 528                             src=["\'](?:https?:)?//players\.brightcove\.net/
 529                             (\d+)/([^/]+)_([^/]+)/index(?:\.min)?\.js
 530                         )
 531                     )?
 532                 ''', webpage):
 533             attrs = extract_attributes(video)
 534
 535             # According to examples from [4] it's unclear whether video id
 536             # may be optional and what to do when it is
 537             video_id = attrs.get('data-video-id')
 538             if not video_id:
 539                 continue
 540
 541             account_id = account_id or attrs.get('data-account')
 542             if not account_id:
 543                 continue
 544
 545             player_id = player_id or attrs.get('data-player') or 'default'
 546             embed = embed or attrs.get('data-embed') or 'default'
 547
 548             bc_url = 'http://players.brightcove.net/%s/%s_%s/index.html?videoId=%s' % (
 549                 account_id, player_id, embed, video_id)
 550
 551             # Some brightcove videos may be embedded with video tag only and
 552             # without script tag or any mentioning of brightcove at all. Such
 553             # embeds are considered ambiguous since they are matched based only
 554             # on data-video-id and data-account attributes and in the wild may
 555             # not be brightcove embeds at all. Let's check reconstructed
 556             # brightcove URLs in case of such embeds and only process valid
 557             # ones. By this we ensure there is indeed a brightcove embed.
 558             if not script_tag and not ie._is_valid_url(
 559                     bc_url, video_id, 'possible brightcove video'):
 560                 continue
 561
 562             entries.append(bc_url)
 563
 564         return entries
 565
 566     def _real_extract(self, url):
 567         url, smuggled_data = unsmuggle_url(url, {})
 568         self._initialize_geo_bypass(smuggled_data.get('geo_countries'))
 569
 570         account_id, player_id, embed, video_id = re.match(self._VALID_URL, url).groups()
 571
 572         webpage = self._download_webpage(
 573             'http://players.brightcove.net/%s/%s_%s/index.min.js'
 574             % (account_id, player_id, embed), video_id)
 575
 576         policy_key = None
 577
 578         catalog = self._search_regex(
 579             r'catalog\(({.+?})\);', webpage, 'catalog', default=None)
 580         if catalog:
 581             catalog = self._parse_json(
 582                 js_to_json(catalog), video_id, fatal=False)
 583             if catalog:
 584                 policy_key = catalog.get('policyKey')
 585
 586         if not policy_key:
 587             policy_key = self._search_regex(
 588                 r'policyKey\s*:\s*(["\'])(?P<pk>.+?)\1',
 589                 webpage, 'policy key', group='pk')
 590
 591         api_url = 'https://edge.api.brightcove.com/playback/v1/accounts/%s/videos/%s' % (account_id, video_id)
 592         try:
 593             json_data = self._download_json(api_url, video_id, headers={
 594                 'Accept': 'application/json;pk=%s' % policy_key
 595             })
 596         except ExtractorError as e:
 597             if isinstance(e.cause, compat_HTTPError) and e.cause.code == 403:
 598                 json_data = self._parse_json(e.cause.read().decode(), video_id)[0]
 599                 message = json_data.get('message') or json_data['error_code']
 600                 if json_data.get('error_subcode') == 'CLIENT_GEO':
 601                     self.raise_geo_restricted(msg=message)
 602                 raise ExtractorError(message, expected=True)
 603             raise
 604
 605         title = json_data['name'].strip()
 606
 607         formats = []
 608         for source in json_data.get('sources', []):
 609             container = source.get('container')
 610             ext = mimetype2ext(source.get('type'))
 611             src = source.get('src')
 612             if ext == 'ism' or container == 'WVM':
 613                 continue
 614             elif ext == 'm3u8' or container == 'M2TS':
 615                 if not src:
 616                     continue
 617                 formats.extend(self._extract_m3u8_formats(
 618                     src, video_id, 'mp4', 'm3u8_native', m3u8_id='hls', fatal=False))
 619             elif ext == 'mpd':
 620                 if not src:
 621                     continue
 622                 formats.extend(self._extract_mpd_formats(src, video_id, 'dash', fatal=False))
 623             else:
 624                 streaming_src = source.get('streaming_src')
 625                 stream_name, app_name = source.get('stream_name'), source.get('app_name')
 626                 if not src and not streaming_src and (not stream_name or not app_name):
 627                     continue
 628                 tbr = float_or_none(source.get('avg_bitrate'), 1000)
 629                 height = int_or_none(source.get('height'))
 630                 width = int_or_none(source.get('width'))
 631                 f = {
 632                     'tbr': tbr,
 633                     'filesize': int_or_none(source.get('size')),
 634                     'container': container,
 635                     'ext': ext or container.lower(),
 636                 }
 637                 if width == 0 and height == 0:
 638                     f.update({
 639                         'vcodec': 'none',
 640                     })
 641                 else:
 642                     f.update({
 643                         'width': width,
 644                         'height': height,
 645                         'vcodec': source.get('codec'),
 646                     })
 647
 648                 def build_format_id(kind):
 649                     format_id = kind
 650                     if tbr:
 651                         format_id += '-%dk' % int(tbr)
 652                     if height:
 653                         format_id += '-%dp' % height
 654                     return format_id
 655
 656                 if src or streaming_src:
 657                     f.update({
 658                         'url': src or streaming_src,
 659                         'format_id': build_format_id('http' if src else 'http-streaming'),
 660                         'source_preference': 0 if src else -1,
 661                     })
 662                 else:
 663                     f.update({
 664                         'url': app_name,
 665                         'play_path': stream_name,
 666                         'format_id': build_format_id('rtmp'),
 667                     })
 668                 formats.append(f)
 669
 670         errors = json_data.get('errors')
 671         if not formats and errors:
 672             error = errors[0]
 673             raise ExtractorError(
 674                 error.get('message') or error.get('error_subcode') or error['error_code'], expected=True)
 675
 676         self._sort_formats(formats)
 677
 678         subtitles = {}
 679         for text_track in json_data.get('text_tracks', []):
 680             if text_track.get('src'):
 681                 subtitles.setdefault(text_track.get('srclang'), []).append({
 682                     'url': text_track['src'],
 683                 })
 684
 685         is_live = False
 686         duration = float_or_none(json_data.get('duration'), 1000)
 687         if duration and duration < 0:
 688             is_live = True
 689
 690         return {
 691             'id': video_id,
 692             'title': self._live_title(title) if is_live else title,
 693             'description': clean_html(json_data.get('description')),
 694             'thumbnail': json_data.get('thumbnail') or json_data.get('poster'),
 695             'duration': duration,
 696             'timestamp': parse_iso8601(json_data.get('published_at')),
 697             'uploader_id': account_id,
 698             'formats': formats,
 699             'subtitles': subtitles,
 700             'tags': json_data.get('tags', []),
 701             'is_live': is_live,
 702         }