_ 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 .adobepass import AdobePassIE
   9 from ..compat import (
  10     compat_etree_fromstring,
  11     compat_parse_qs,
  12     compat_str,
  13     compat_urllib_parse_urlparse,
  14     compat_urlparse,
  15     compat_xml_parse_error,
  16     compat_HTTPError,
  17 )
  18 from ..utils import (
  19     determine_ext,
  20     ExtractorError,
  21     extract_attributes,
  22     find_xpath_attr,
  23     fix_xml_ampersands,
  24     float_or_none,
  25     js_to_json,
  26     int_or_none,
  27     parse_iso8601,
  28     unescapeHTML,
  29     unsmuggle_url,
  30     update_url_query,
  31     clean_html,
  32     mimetype2ext,
  33 )
  34
  35
  36 class BrightcoveLegacyIE(InfoExtractor):
  37     IE_NAME = 'brightcove:legacy'
  38     _VALID_URL = r'(?:https?://.*brightcove\.com/(services|viewer).*?\?|brightcove:)(?P<query>.*)'
  39     _FEDERATED_URL = 'http://c.brightcove.com/services/viewer/htmlFederated'
  40
  41     _TESTS = [
  42         {
  43             # From http://www.8tv.cat/8aldia/videos/xavier-sala-i-martin-aquesta-tarda-a-8-al-dia/
  44             'url': 'http://c.brightcove.com/services/viewer/htmlFederated?playerID=1654948606001&flashID=myExperience&%40videoPlayer=2371591881001',
  45             'md5': '5423e113865d26e40624dce2e4b45d95',
  46             'note': 'Test Brightcove downloads and detection in GenericIE',
  47             'info_dict': {
  48                 'id': '2371591881001',
  49                 'ext': 'mp4',
  50                 'title': 'Xavier Sala i Martín: “Un banc que no presta és un banc zombi que no serveix per a res”',
  51                 'uploader': '8TV',
  52                 'description': 'md5:a950cc4285c43e44d763d036710cd9cd',
  53                 'timestamp': 1368213670,
  54                 'upload_date': '20130510',
  55                 'uploader_id': '1589608506001',
  56             }
  57         },
  58         {
  59             # From http://medianetwork.oracle.com/video/player/1785452137001
  60             'url': 'http://c.brightcove.com/services/viewer/htmlFederated?playerID=1217746023001&flashID=myPlayer&%40videoPlayer=1785452137001',
  61             'info_dict': {
  62                 'id': '1785452137001',
  63                 'ext': 'flv',
  64                 'title': 'JVMLS 2012: Arrays 2.0 - Opportunities and Challenges',
  65                 'description': 'John Rose speaks at the JVM Language Summit, August 1, 2012.',
  66                 'uploader': 'Oracle',
  67                 'timestamp': 1344975024,
  68                 'upload_date': '20120814',
  69                 'uploader_id': '1460825906',
  70             },
  71         },
  72         {
  73             # From http://mashable.com/2013/10/26/thermoelectric-bracelet-lets-you-control-your-body-temperature/
  74             'url': 'http://c.brightcove.com/services/viewer/federated_f9?&playerID=1265504713001&publisherID=AQ%7E%7E%2CAAABBzUwv1E%7E%2CxP-xFHVUstiMFlNYfvF4G9yFnNaqCw_9&videoID=2750934548001',
  75             'info_dict': {
  76                 'id': '2750934548001',
  77                 'ext': 'mp4',
  78                 'title': 'This Bracelet Acts as a Personal Thermostat',
  79                 'description': 'md5:547b78c64f4112766ccf4e151c20b6a0',
  80                 'uploader': 'Mashable',
  81                 'timestamp': 1382041798,
  82                 'upload_date': '20131017',
  83                 'uploader_id': '1130468786001',
  84             },
  85         },
  86         {
  87             # test that the default referer works
  88             # from http://national.ballet.ca/interact/video/Lost_in_Motion_II/
  89             'url': 'http://link.brightcove.com/services/player/bcpid756015033001?bckey=AQ~~,AAAApYJi_Ck~,GxhXCegT1Dp39ilhXuxMJxasUhVNZiil&bctid=2878862109001',
  90             'info_dict': {
  91                 'id': '2878862109001',
  92                 'ext': 'mp4',
  93                 'title': 'Lost in Motion II',
  94                 'description': 'md5:363109c02998fee92ec02211bd8000df',
  95                 'uploader': 'National Ballet of Canada',
  96             },
  97             'skip': 'Video gone',
  98         },
  99         {
 100             # test flv videos served by akamaihd.net
 101             # From http://www.redbull.com/en/bike/stories/1331655643987/replay-uci-dh-world-cup-2014-from-fort-william
 102             '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',
 103             # The md5 checksum changes on each download
 104             'info_dict': {
 105                 'id': '3750436379001',
 106                 'ext': 'flv',
 107                 'title': 'UCI MTB World Cup 2014: Fort William, UK - Downhill Finals',
 108                 'uploader': 'RBTV Old (do not use)',
 109                 'description': 'UCI MTB World Cup 2014: Fort William, UK - Downhill Finals',
 110                 'timestamp': 1409122195,
 111                 'upload_date': '20140827',
 112                 'uploader_id': '710858724001',
 113             },
 114             'skip': 'Video gone',
 115         },
 116         {
 117             # playlist with 'videoList'
 118             # from http://support.brightcove.com/en/video-cloud/docs/playlist-support-single-video-players
 119             'url': 'http://c.brightcove.com/services/viewer/htmlFederated?playerID=3550052898001&playerKey=AQ%7E%7E%2CAAABmA9XpXk%7E%2C-Kp7jNgisre1fG5OdqpAFUTcs0lP_ZoL',
 120             'info_dict': {
 121                 'title': 'Sealife',
 122                 'id': '3550319591001',
 123             },
 124             'playlist_mincount': 7,
 125         },
 126         {
 127             # playlist with 'playlistTab' (https://github.com/rg3/youtube-dl/issues/9965)
 128             'url': 'http://c.brightcove.com/services/json/experience/runtime/?command=get_programming_for_experience&playerKey=AQ%7E%7E,AAABXlLMdok%7E,NJ4EoMlZ4rZdx9eU1rkMVd8EaYPBBUlg',
 129             'info_dict': {
 130                 'id': '1522758701001',
 131                 'title': 'Lesson 08',
 132             },
 133             'playlist_mincount': 10,
 134         },
 135         {
 136             # playerID inferred from bcpid
 137             # from http://www.un.org/chinese/News/story.asp?NewsID=27724
 138             'url': 'https://link.brightcove.com/services/player/bcpid1722935254001/?bctid=5360463607001&autoStart=false&secureConnections=true&width=650&height=350',
 139             'only_matching': True,  # Tested in GenericIE
 140         }
 141     ]
 142     FLV_VCODECS = {
 143         1: 'SORENSON',
 144         2: 'ON2',
 145         3: 'H264',
 146         4: 'VP8',
 147     }
 148
 149     @classmethod
 150     def _build_brighcove_url(cls, object_str):
 151         """
 152         Build a Brightcove url from a xml string containing
 153         <object class="BrightcoveExperience">{params}</object>
 154         """
 155
 156         # Fix up some stupid HTML, see https://github.com/rg3/youtube-dl/issues/1553
 157         object_str = re.sub(r'(<param(?:\s+[a-zA-Z0-9_]+="[^"]*")*)>',
 158                             lambda m: m.group(1) + '/>', object_str)
 159         # Fix up some stupid XML, see https://github.com/rg3/youtube-dl/issues/1608
 160         object_str = object_str.replace('<--', '<!--')
 161         # remove namespace to simplify extraction
 162         object_str = re.sub(r'(<object[^>]*)(xmlns=".*?")', r'\1', object_str)
 163         object_str = fix_xml_ampersands(object_str)
 164
 165         try:
 166             object_doc = compat_etree_fromstring(object_str.encode('utf-8'))
 167         except compat_xml_parse_error:
 168             return
 169
 170         fv_el = find_xpath_attr(object_doc, './param', 'name', 'flashVars')
 171         if fv_el is not None:
 172             flashvars = dict(
 173                 (k, v[0])
 174                 for k, v in compat_parse_qs(fv_el.attrib['value']).items())
 175         else:
 176             flashvars = {}
 177
 178         data_url = object_doc.attrib.get('data', '')
 179         data_url_params = compat_parse_qs(compat_urllib_parse_urlparse(data_url).query)
 180
 181         def find_param(name):
 182             if name in flashvars:
 183                 return flashvars[name]
 184             node = find_xpath_attr(object_doc, './param', 'name', name)
 185             if node is not None:
 186                 return node.attrib['value']
 187             return data_url_params.get(name)
 188
 189         params = {}
 190
 191         playerID = find_param('playerID') or find_param('playerId')
 192         if playerID is None:
 193             raise ExtractorError('Cannot find player ID')
 194         params['playerID'] = playerID
 195
 196         playerKey = find_param('playerKey')
 197         # Not all pages define this value
 198         if playerKey is not None:
 199             params['playerKey'] = playerKey
 200         # These fields hold the id of the video
 201         videoPlayer = find_param('@videoPlayer') or find_param('videoId') or find_param('videoID') or find_param('@videoList')
 202         if videoPlayer is not None:
 203             if isinstance(videoPlayer, list):
 204                 videoPlayer = videoPlayer[0]
 205             videoPlayer = videoPlayer.strip()
 206             # UUID is also possible for videoPlayer (e.g.
 207             # http://www.popcornflix.com/hoodies-vs-hooligans/7f2d2b87-bbf2-4623-acfb-ea942b4f01dd
 208             # or http://www8.hp.com/cn/zh/home.html)
 209             if not (re.match(
 210                     r'^(?:\d+|[\da-fA-F]{8}-?[\da-fA-F]{4}-?[\da-fA-F]{4}-?[\da-fA-F]{4}-?[\da-fA-F]{12})$',
 211                     videoPlayer) or videoPlayer.startswith('ref:')):
 212                 return None
 213             params['@videoPlayer'] = videoPlayer
 214         linkBase = find_param('linkBaseURL')
 215         if linkBase is not None:
 216             params['linkBaseURL'] = linkBase
 217         return cls._make_brightcove_url(params)
 218
 219     @classmethod
 220     def _build_brighcove_url_from_js(cls, object_js):
 221         # The layout of JS is as follows:
 222         # customBC.createVideo = function (width, height, playerID, playerKey, videoPlayer, VideoRandomID) {
 223         #   // build Brightcove <object /> XML
 224         # }
 225         m = re.search(
 226             r'''(?x)customBC\.createVideo\(
 227                 .*?                                                  # skipping width and height
 228                 ["\'](?P<playerID>\d+)["\']\s*,\s*                   # playerID
 229                 ["\'](?P<playerKey>AQ[^"\']{48})[^"\']*["\']\s*,\s*  # playerKey begins with AQ and is 50 characters
 230                                                                      # in length, however it's appended to itself
 231                                                                      # in places, so truncate
 232                 ["\'](?P<videoID>\d+)["\']                           # @videoPlayer
 233             ''', object_js)
 234         if m:
 235             return cls._make_brightcove_url(m.groupdict())
 236
 237     @classmethod
 238     def _make_brightcove_url(cls, params):
 239         return update_url_query(cls._FEDERATED_URL, params)
 240
 241     @classmethod
 242     def _extract_brightcove_url(cls, webpage):
 243         """Try to extract the brightcove url from the webpage, returns None
 244         if it can't be found
 245         """
 246         urls = cls._extract_brightcove_urls(webpage)
 247         return urls[0] if urls else None
 248
 249     @classmethod
 250     def _extract_brightcove_urls(cls, webpage):
 251         """Return a list of all Brightcove URLs from the webpage """
 252
 253         url_m = re.search(
 254             r'''(?x)
 255                 <meta\s+
 256                     (?:property|itemprop)=([\'"])(?:og:video|embedURL)\1[^>]+
 257                     content=([\'"])(?P<url>https?://(?:secure|c)\.brightcove.com/(?:(?!\2).)+)\2
 258             ''', webpage)
 259         if url_m:
 260             url = unescapeHTML(url_m.group('url'))
 261             # Some sites don't add it, we can't download with this url, for example:
 262             # http://www.ktvu.com/videos/news/raw-video-caltrain-releases-video-of-man-almost/vCTZdY/
 263             if 'playerKey' in url or 'videoId' in url or 'idVideo' in url:
 264                 return [url]
 265
 266         matches = re.findall(
 267             r'''(?sx)<object
 268             (?:
 269                 [^>]+?class=[\'"][^>]*?BrightcoveExperience.*?[\'"] |
 270                 [^>]*?>\s*<param\s+name="movie"\s+value="https?://[^/]*brightcove\.com/
 271             ).+?>\s*</object>''',
 272             webpage)
 273         if matches:
 274             return list(filter(None, [cls._build_brighcove_url(m) for m in matches]))
 275
 276         matches = re.findall(r'(customBC\.createVideo\(.+?\);)', webpage)
 277         if matches:
 278             return list(filter(None, [
 279                 cls._build_brighcove_url_from_js(custom_bc)
 280                 for custom_bc in matches]))
 281         return [src for _, src in re.findall(
 282             r'<iframe[^>]+src=([\'"])((?:https?:)?//link\.brightcove\.com/services/player/(?!\1).+)\1', webpage)]
 283
 284     def _real_extract(self, url):
 285         url, smuggled_data = unsmuggle_url(url, {})
 286
 287         # Change the 'videoId' and others field to '@videoPlayer'
 288         url = re.sub(r'(?<=[?&])(videoI(d|D)|idVideo|bctid)', '%40videoPlayer', url)
 289         # Change bckey (used by bcove.me urls) to playerKey
 290         url = re.sub(r'(?<=[?&])bckey', 'playerKey', url)
 291         mobj = re.match(self._VALID_URL, url)
 292         query_str = mobj.group('query')
 293         query = compat_urlparse.parse_qs(query_str)
 294
 295         videoPlayer = query.get('@videoPlayer')
 296         if videoPlayer:
 297             # We set the original url as the default 'Referer' header
 298             referer = smuggled_data.get('Referer', url)
 299             if 'playerID' not in query:
 300                 mobj = re.search(r'/bcpid(\d+)', url)
 301                 if mobj is not None:
 302                     query['playerID'] = [mobj.group(1)]
 303             return self._get_video_info(
 304                 videoPlayer[0], query, referer=referer)
 305         elif 'playerKey' in query:
 306             player_key = query['playerKey']
 307             return self._get_playlist_info(player_key[0])
 308         else:
 309             raise ExtractorError(
 310                 'Cannot find playerKey= variable. Did you forget quotes in a shell invocation?',
 311                 expected=True)
 312
 313     def _get_video_info(self, video_id, query, referer=None):
 314         headers = {}
 315         linkBase = query.get('linkBaseURL')
 316         if linkBase is not None:
 317             referer = linkBase[0]
 318         if referer is not None:
 319             headers['Referer'] = referer
 320         webpage = self._download_webpage(self._FEDERATED_URL, video_id, headers=headers, query=query)
 321
 322         error_msg = self._html_search_regex(
 323             r"<h1>We're sorry.</h1>([\s\n]*<p>.*?</p>)+", webpage,
 324             'error message', default=None)
 325         if error_msg is not None:
 326             raise ExtractorError(
 327                 'brightcove said: %s' % error_msg, expected=True)
 328
 329         self.report_extraction(video_id)
 330         info = self._search_regex(r'var experienceJSON = ({.*});', webpage, 'json')
 331         info = json.loads(info)['data']
 332         video_info = info['programmedContent']['videoPlayer']['mediaDTO']
 333         video_info['_youtubedl_adServerURL'] = info.get('adServerURL')
 334
 335         return self._extract_video_info(video_info)
 336
 337     def _get_playlist_info(self, player_key):
 338         info_url = 'http://c.brightcove.com/services/json/experience/runtime/?command=get_programming_for_experience&playerKey=%s' % player_key
 339         playlist_info = self._download_webpage(
 340             info_url, player_key, 'Downloading playlist information')
 341
 342         json_data = json.loads(playlist_info)
 343         if 'videoList' in json_data:
 344             playlist_info = json_data['videoList']
 345             playlist_dto = playlist_info['mediaCollectionDTO']
 346         elif 'playlistTabs' in json_data:
 347             playlist_info = json_data['playlistTabs']
 348             playlist_dto = playlist_info['lineupListDTO']['playlistDTOs'][0]
 349         else:
 350             raise ExtractorError('Empty playlist')
 351
 352         videos = [self._extract_video_info(video_info) for video_info in playlist_dto['videoDTOs']]
 353
 354         return self.playlist_result(videos, playlist_id='%s' % playlist_info['id'],
 355                                     playlist_title=playlist_dto['displayName'])
 356
 357     def _extract_video_info(self, video_info):
 358         video_id = compat_str(video_info['id'])
 359         publisher_id = video_info.get('publisherId')
 360         info = {
 361             'id': video_id,
 362             'title': video_info['displayName'].strip(),
 363             'description': video_info.get('shortDescription'),
 364             'thumbnail': video_info.get('videoStillURL') or video_info.get('thumbnailURL'),
 365             'uploader': video_info.get('publisherName'),
 366             'uploader_id': compat_str(publisher_id) if publisher_id else None,
 367             'duration': float_or_none(video_info.get('length'), 1000),
 368             'timestamp': int_or_none(video_info.get('creationDate'), 1000),
 369         }
 370
 371         renditions = video_info.get('renditions', []) + video_info.get('IOSRenditions', [])
 372         if renditions:
 373             formats = []
 374             for rend in renditions:
 375                 url = rend['defaultURL']
 376                 if not url:
 377                     continue
 378                 ext = None
 379                 if rend['remote']:
 380                     url_comp = compat_urllib_parse_urlparse(url)
 381                     if url_comp.path.endswith('.m3u8'):
 382                         formats.extend(
 383                             self._extract_m3u8_formats(
 384                                 url, video_id, 'mp4', 'm3u8_native', m3u8_id='hls', fatal=False))
 385                         continue
 386                     elif 'akamaihd.net' in url_comp.netloc:
 387                         # This type of renditions are served through
 388                         # akamaihd.net, but they don't use f4m manifests
 389                         url = url.replace('control/', '') + '?&v=3.3.0&fp=13&r=FEEFJ&g=RTSJIMBMPFPB'
 390                         ext = 'flv'
 391                 if ext is None:
 392                     ext = determine_ext(url)
 393                 tbr = int_or_none(rend.get('encodingRate'), 1000)
 394                 a_format = {
 395                     'format_id': 'http%s' % ('-%s' % tbr if tbr else ''),
 396                     'url': url,
 397                     'ext': ext,
 398                     'filesize': int_or_none(rend.get('size')) or None,
 399                     'tbr': tbr,
 400                 }
 401                 if rend.get('audioOnly'):
 402                     a_format.update({
 403                         'vcodec': 'none',
 404                     })
 405                 else:
 406                     a_format.update({
 407                         'height': int_or_none(rend.get('frameHeight')),
 408                         'width': int_or_none(rend.get('frameWidth')),
 409                         'vcodec': rend.get('videoCodec'),
 410                     })
 411
 412                 # m3u8 manifests with remote == false are media playlists
 413                 # Not calling _extract_m3u8_formats here to save network traffic
 414                 if ext == 'm3u8':
 415                     a_format.update({
 416                         'format_id': 'hls%s' % ('-%s' % tbr if tbr else ''),
 417                         'ext': 'mp4',
 418                         'protocol': 'm3u8_native',
 419                     })
 420
 421                 formats.append(a_format)
 422             self._sort_formats(formats)
 423             info['formats'] = formats
 424         elif video_info.get('FLVFullLengthURL') is not None:
 425             info.update({
 426                 'url': video_info['FLVFullLengthURL'],
 427                 'vcodec': self.FLV_VCODECS.get(video_info.get('FLVFullCodec')),
 428                 'filesize': int_or_none(video_info.get('FLVFullSize')),
 429             })
 430
 431         if self._downloader.params.get('include_ads', False):
 432             adServerURL = video_info.get('_youtubedl_adServerURL')
 433             if adServerURL:
 434                 ad_info = {
 435                     '_type': 'url',
 436                     'url': adServerURL,
 437                 }
 438                 if 'url' in info:
 439                     return {
 440                         '_type': 'playlist',
 441                         'title': info['title'],
 442                         'entries': [ad_info, info],
 443                     }
 444                 else:
 445                     return ad_info
 446
 447         if 'url' not in info and not info.get('formats'):
 448             raise ExtractorError('Unable to extract video url for %s' % video_id)
 449         return info
 450
 451
 452 class BrightcoveNewIE(AdobePassIE):
 453     IE_NAME = 'brightcove:new'
 454     _VALID_URL = r'https?://players\.brightcove\.net/(?P<account_id>\d+)/(?P<player_id>[^/]+)_(?P<embed>[^/]+)/index\.html\?.*videoId=(?P<video_id>\d+|ref:[^&]+)'
 455     _TESTS = [{
 456         'url': 'http://players.brightcove.net/929656772001/e41d32dc-ec74-459e-a845-6c69f7b724ea_default/index.html?videoId=4463358922001',
 457         'md5': 'c8100925723840d4b0d243f7025703be',
 458         'info_dict': {
 459             'id': '4463358922001',
 460             'ext': 'mp4',
 461             'title': 'Meet the man behind Popcorn Time',
 462             'description': 'md5:eac376a4fe366edc70279bfb681aea16',
 463             'duration': 165.768,
 464             'timestamp': 1441391203,
 465             'upload_date': '20150904',
 466             'uploader_id': '929656772001',
 467             'formats': 'mincount:20',
 468         },
 469     }, {
 470         # with rtmp streams
 471         'url': 'http://players.brightcove.net/4036320279001/5d112ed9-283f-485f-a7f9-33f42e8bc042_default/index.html?videoId=4279049078001',
 472         'info_dict': {
 473             'id': '4279049078001',
 474             'ext': 'mp4',
 475             'title': 'Titansgrave: Chapter 0',
 476             'description': 'Titansgrave: Chapter 0',
 477             'duration': 1242.058,
 478             'timestamp': 1433556729,
 479             'upload_date': '20150606',
 480             'uploader_id': '4036320279001',
 481             'formats': 'mincount:39',
 482         },
 483         'params': {
 484             # m3u8 download
 485             'skip_download': True,
 486         }
 487     }, {
 488         # ref: prefixed video id
 489         'url': 'http://players.brightcove.net/3910869709001/21519b5c-4b3b-4363-accb-bdc8f358f823_default/index.html?videoId=ref:7069442',
 490         'only_matching': True,
 491     }, {
 492         # non numeric ref: prefixed video id
 493         'url': 'http://players.brightcove.net/710858724001/default_default/index.html?videoId=ref:event-stream-356',
 494         'only_matching': True,
 495     }, {
 496         # unavailable video without message but with error_code
 497         'url': 'http://players.brightcove.net/1305187701/c832abfb-641b-44eb-9da0-2fe76786505f_default/index.html?videoId=4377407326001',
 498         'only_matching': True,
 499     }]
 500
 501     @staticmethod
 502     def _extract_url(ie, webpage):
 503         urls = BrightcoveNewIE._extract_urls(ie, webpage)
 504         return urls[0] if urls else None
 505
 506     @staticmethod
 507     def _extract_urls(ie, webpage):
 508         # Reference:
 509         # 1. http://docs.brightcove.com/en/video-cloud/brightcove-player/guides/publish-video.html#setvideoiniframe
 510         # 2. http://docs.brightcove.com/en/video-cloud/brightcove-player/guides/publish-video.html#tag
 511         # 3. http://docs.brightcove.com/en/video-cloud/brightcove-player/guides/publish-video.html#setvideousingjavascript
 512         # 4. http://docs.brightcove.com/en/video-cloud/brightcove-player/guides/in-page-embed-player-implementation.html
 513         # 5. https://support.brightcove.com/en/video-cloud/docs/dynamically-assigning-videos-player
 514
 515         entries = []
 516
 517         # Look for iframe embeds [1]
 518         for _, url in re.findall(
 519                 r'<iframe[^>]+src=(["\'])((?:https?:)?//players\.brightcove\.net/\d+/[^/]+/index\.html.+?)\1', webpage):
 520             entries.append(url if url.startswith('http') else 'http:' + url)
 521
 522         # Look for <video> tags [2] and embed_in_page embeds [3]
 523         # [2] looks like:
 524         for video, script_tag, account_id, player_id, embed in re.findall(
 525                 r'''(?isx)
 526                     (<video\s+[^>]*\bdata-video-id\s*=\s*['"]?[^>]+>)
 527                     (?:.*?
 528                         (<script[^>]+
 529                             src=["\'](?:https?:)?//players\.brightcove\.net/
 530                             (\d+)/([^/]+)_([^/]+)/index(?:\.min)?\.js
 531                         )
 532                     )?
 533                 ''', webpage):
 534             attrs = extract_attributes(video)
 535
 536             # According to examples from [4] it's unclear whether video id
 537             # may be optional and what to do when it is
 538             video_id = attrs.get('data-video-id')
 539             if not video_id:
 540                 continue
 541
 542             account_id = account_id or attrs.get('data-account')
 543             if not account_id:
 544                 continue
 545
 546             player_id = player_id or attrs.get('data-player') or 'default'
 547             embed = embed or attrs.get('data-embed') or 'default'
 548
 549             bc_url = 'http://players.brightcove.net/%s/%s_%s/index.html?videoId=%s' % (
 550                 account_id, player_id, embed, video_id)
 551
 552             # Some brightcove videos may be embedded with video tag only and
 553             # without script tag or any mentioning of brightcove at all. Such
 554             # embeds are considered ambiguous since they are matched based only
 555             # on data-video-id and data-account attributes and in the wild may
 556             # not be brightcove embeds at all. Let's check reconstructed
 557             # brightcove URLs in case of such embeds and only process valid
 558             # ones. By this we ensure there is indeed a brightcove embed.
 559             if not script_tag and not ie._is_valid_url(
 560                     bc_url, video_id, 'possible brightcove video'):
 561                 continue
 562
 563             entries.append(bc_url)
 564
 565         return entries
 566
 567     def _parse_brightcove_metadata(self, json_data, video_id, headers={}):
 568         title = json_data['name'].strip()
 569
 570         formats = []
 571         for source in json_data.get('sources', []):
 572             container = source.get('container')
 573             ext = mimetype2ext(source.get('type'))
 574             src = source.get('src')
 575             if ext == 'ism' or container == 'WVM':
 576                 continue
 577             elif ext == 'm3u8' or container == 'M2TS':
 578                 if not src:
 579                     continue
 580                 formats.extend(self._extract_m3u8_formats(
 581                     src, video_id, 'mp4', 'm3u8_native', m3u8_id='hls', fatal=False))
 582             elif ext == 'mpd':
 583                 if not src:
 584                     continue
 585                 formats.extend(self._extract_mpd_formats(src, video_id, 'dash', fatal=False))
 586             else:
 587                 streaming_src = source.get('streaming_src')
 588                 stream_name, app_name = source.get('stream_name'), source.get('app_name')
 589                 if not src and not streaming_src and (not stream_name or not app_name):
 590                     continue
 591                 tbr = float_or_none(source.get('avg_bitrate'), 1000)
 592                 height = int_or_none(source.get('height'))
 593                 width = int_or_none(source.get('width'))
 594                 f = {
 595                     'tbr': tbr,
 596                     'filesize': int_or_none(source.get('size')),
 597                     'container': container,
 598                     'ext': ext or container.lower(),
 599                 }
 600                 if width == 0 and height == 0:
 601                     f.update({
 602                         'vcodec': 'none',
 603                     })
 604                 else:
 605                     f.update({
 606                         'width': width,
 607                         'height': height,
 608                         'vcodec': source.get('codec'),
 609                     })
 610
 611                 def build_format_id(kind):
 612                     format_id = kind
 613                     if tbr:
 614                         format_id += '-%dk' % int(tbr)
 615                     if height:
 616                         format_id += '-%dp' % height
 617                     return format_id
 618
 619                 if src or streaming_src:
 620                     f.update({
 621                         'url': src or streaming_src,
 622                         'format_id': build_format_id('http' if src else 'http-streaming'),
 623                         'source_preference': 0 if src else -1,
 624                     })
 625                 else:
 626                     f.update({
 627                         'url': app_name,
 628                         'play_path': stream_name,
 629                         'format_id': build_format_id('rtmp'),
 630                     })
 631                 formats.append(f)
 632
 633         errors = json_data.get('errors')
 634         if not formats and errors:
 635             error = errors[0]
 636             raise ExtractorError(
 637                 error.get('message') or error.get('error_subcode') or error['error_code'], expected=True)
 638
 639         self._sort_formats(formats)
 640
 641         for f in formats:
 642             f.setdefault('http_headers', {}).update(headers)
 643
 644         subtitles = {}
 645         for text_track in json_data.get('text_tracks', []):
 646             if text_track.get('src'):
 647                 subtitles.setdefault(text_track.get('srclang'), []).append({
 648                     'url': text_track['src'],
 649                 })
 650
 651         is_live = False
 652         duration = float_or_none(json_data.get('duration'), 1000)
 653         if duration is not None and duration <= 0:
 654             is_live = True
 655
 656         return {
 657             'id': video_id,
 658             'title': self._live_title(title) if is_live else title,
 659             'description': clean_html(json_data.get('description')),
 660             'thumbnail': json_data.get('thumbnail') or json_data.get('poster'),
 661             'duration': duration,
 662             'timestamp': parse_iso8601(json_data.get('published_at')),
 663             'uploader_id': json_data.get('account_id'),
 664             'formats': formats,
 665             'subtitles': subtitles,
 666             'tags': json_data.get('tags', []),
 667             'is_live': is_live,
 668         }
 669
 670     def _real_extract(self, url):
 671         url, smuggled_data = unsmuggle_url(url, {})
 672         self._initialize_geo_bypass(smuggled_data.get('geo_countries'))
 673
 674         account_id, player_id, embed, video_id = re.match(self._VALID_URL, url).groups()
 675
 676         webpage = self._download_webpage(
 677             'http://players.brightcove.net/%s/%s_%s/index.min.js'
 678             % (account_id, player_id, embed), video_id)
 679
 680         policy_key = None
 681
 682         catalog = self._search_regex(
 683             r'catalog\(({.+?})\);', webpage, 'catalog', default=None)
 684         if catalog:
 685             catalog = self._parse_json(
 686                 js_to_json(catalog), video_id, fatal=False)
 687             if catalog:
 688                 policy_key = catalog.get('policyKey')
 689
 690         if not policy_key:
 691             policy_key = self._search_regex(
 692                 r'policyKey\s*:\s*(["\'])(?P<pk>.+?)\1',
 693                 webpage, 'policy key', group='pk')
 694
 695         api_url = 'https://edge.api.brightcove.com/playback/v1/accounts/%s/videos/%s' % (account_id, video_id)
 696         headers = {
 697             'Accept': 'application/json;pk=%s' % policy_key,
 698         }
 699         referrer = smuggled_data.get('referrer')
 700         if referrer:
 701             headers.update({
 702                 'Referer': referrer,
 703                 'Origin': re.search(r'https?://[^/]+', referrer).group(0),
 704             })
 705         try:
 706             json_data = self._download_json(api_url, video_id, headers=headers)
 707         except ExtractorError as e:
 708             if isinstance(e.cause, compat_HTTPError) and e.cause.code == 403:
 709                 json_data = self._parse_json(e.cause.read().decode(), video_id)[0]
 710                 message = json_data.get('message') or json_data['error_code']
 711                 if json_data.get('error_subcode') == 'CLIENT_GEO':
 712                     self.raise_geo_restricted(msg=message)
 713                 raise ExtractorError(message, expected=True)
 714             raise
 715
 716         errors = json_data.get('errors')
 717         if errors and errors[0].get('error_subcode') == 'TVE_AUTH':
 718             custom_fields = json_data['custom_fields']
 719             tve_token = self._extract_mvpd_auth(
 720                 smuggled_data['source_url'], video_id,
 721                 custom_fields['bcadobepassrequestorid'],
 722                 custom_fields['bcadobepassresourceid'])
 723             json_data = self._download_json(
 724                 api_url, video_id, headers={
 725                     'Accept': 'application/json;pk=%s' % policy_key
 726                 }, query={
 727                     'tveToken': tve_token,
 728                 })
 729
 730         return self._parse_brightcove_metadata(
 731             json_data, video_id, headers=headers)