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

   1 # encoding: utf-8
   2 from __future__ import unicode_literals
   3
   4 import re
   5 import json
   6 import xml.etree.ElementTree
   7
   8 from .common import InfoExtractor
   9 from ..compat import (
  10     compat_parse_qs,
  11     compat_str,
  12     compat_urllib_parse,
  13     compat_urllib_parse_urlparse,
  14     compat_urllib_request,
  15     compat_urlparse,
  16     compat_xml_parse_error,
  17 )
  18 from ..utils import (
  19     determine_ext,
  20     ExtractorError,
  21     find_xpath_attr,
  22     fix_xml_ampersands,
  23     unescapeHTML,
  24     unsmuggle_url,
  25     js_to_json,
  26     int_or_none,
  27     parse_iso8601,
  28     extract_attributes,
  29 )
  30
  31
  32 class BrightcoveIE(InfoExtractor):
  33     _VALID_URL = r'(?:https?://.*brightcove\.com/(services|viewer).*?\?|brightcove:)(?P<query>.*)'
  34     _FEDERATED_URL_TEMPLATE = 'http://c.brightcove.com/services/viewer/htmlFederated?%s'
  35
  36     _TESTS = [
  37         {
  38             # From http://www.8tv.cat/8aldia/videos/xavier-sala-i-martin-aquesta-tarda-a-8-al-dia/
  39             'url': 'http://c.brightcove.com/services/viewer/htmlFederated?playerID=1654948606001&flashID=myExperience&%40videoPlayer=2371591881001',
  40             'md5': '5423e113865d26e40624dce2e4b45d95',
  41             'note': 'Test Brightcove downloads and detection in GenericIE',
  42             'info_dict': {
  43                 'id': '2371591881001',
  44                 'ext': 'mp4',
  45                 'title': 'Xavier Sala i Martín: “Un banc que no presta és un banc zombi que no serveix per a res”',
  46                 'uploader': '8TV',
  47                 'description': 'md5:a950cc4285c43e44d763d036710cd9cd',
  48             }
  49         },
  50         {
  51             # From http://medianetwork.oracle.com/video/player/1785452137001
  52             'url': 'http://c.brightcove.com/services/viewer/htmlFederated?playerID=1217746023001&flashID=myPlayer&%40videoPlayer=1785452137001',
  53             'info_dict': {
  54                 'id': '1785452137001',
  55                 'ext': 'flv',
  56                 'title': 'JVMLS 2012: Arrays 2.0 - Opportunities and Challenges',
  57                 'description': 'John Rose speaks at the JVM Language Summit, August 1, 2012.',
  58                 'uploader': 'Oracle',
  59             },
  60         },
  61         {
  62             # From http://mashable.com/2013/10/26/thermoelectric-bracelet-lets-you-control-your-body-temperature/
  63             'url': 'http://c.brightcove.com/services/viewer/federated_f9?&playerID=1265504713001&publisherID=AQ%7E%7E%2CAAABBzUwv1E%7E%2CxP-xFHVUstiMFlNYfvF4G9yFnNaqCw_9&videoID=2750934548001',
  64             'info_dict': {
  65                 'id': '2750934548001',
  66                 'ext': 'mp4',
  67                 'title': 'This Bracelet Acts as a Personal Thermostat',
  68                 'description': 'md5:547b78c64f4112766ccf4e151c20b6a0',
  69                 'uploader': 'Mashable',
  70             },
  71         },
  72         {
  73             # test that the default referer works
  74             # from http://national.ballet.ca/interact/video/Lost_in_Motion_II/
  75             'url': 'http://link.brightcove.com/services/player/bcpid756015033001?bckey=AQ~~,AAAApYJi_Ck~,GxhXCegT1Dp39ilhXuxMJxasUhVNZiil&bctid=2878862109001',
  76             'info_dict': {
  77                 'id': '2878862109001',
  78                 'ext': 'mp4',
  79                 'title': 'Lost in Motion II',
  80                 'description': 'md5:363109c02998fee92ec02211bd8000df',
  81                 'uploader': 'National Ballet of Canada',
  82             },
  83         },
  84         {
  85             # test flv videos served by akamaihd.net
  86             # From http://www.redbull.com/en/bike/stories/1331655643987/replay-uci-dh-world-cup-2014-from-fort-william
  87             'url': 'http://c.brightcove.com/services/viewer/htmlFederated?%40videoPlayer=ref%3ABC2996102916001&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',
  88             # The md5 checksum changes on each download
  89             'info_dict': {
  90                 'id': '2996102916001',
  91                 'ext': 'flv',
  92                 'title': 'UCI MTB World Cup 2014: Fort William, UK - Downhill Finals',
  93                 'uploader': 'Red Bull TV',
  94                 'description': 'UCI MTB World Cup 2014: Fort William, UK - Downhill Finals',
  95             },
  96         },
  97         {
  98             # playlist test
  99             # from http://support.brightcove.com/en/video-cloud/docs/playlist-support-single-video-players
 100             'url': 'http://c.brightcove.com/services/viewer/htmlFederated?playerID=3550052898001&playerKey=AQ%7E%7E%2CAAABmA9XpXk%7E%2C-Kp7jNgisre1fG5OdqpAFUTcs0lP_ZoL',
 101             'info_dict': {
 102                 'title': 'Sealife',
 103                 'id': '3550319591001',
 104             },
 105             'playlist_mincount': 7,
 106         },
 107     ]
 108
 109     @classmethod
 110     def _build_brighcove_url(cls, object_str):
 111         """
 112         Build a Brightcove url from a xml string containing
 113         <object class="BrightcoveExperience">{params}</object>
 114         """
 115
 116         # Fix up some stupid HTML, see https://github.com/rg3/youtube-dl/issues/1553
 117         object_str = re.sub(r'(<param(?:\s+[a-zA-Z0-9_]+="[^"]*")*)>',
 118                             lambda m: m.group(1) + '/>', object_str)
 119         # Fix up some stupid XML, see https://github.com/rg3/youtube-dl/issues/1608
 120         object_str = object_str.replace('<--', '<!--')
 121         # remove namespace to simplify extraction
 122         object_str = re.sub(r'(<object[^>]*)(xmlns=".*?")', r'\1', object_str)
 123         object_str = fix_xml_ampersands(object_str)
 124
 125         try:
 126             object_doc = xml.etree.ElementTree.fromstring(object_str.encode('utf-8'))
 127         except compat_xml_parse_error:
 128             return
 129
 130         fv_el = find_xpath_attr(object_doc, './param', 'name', 'flashVars')
 131         if fv_el is not None:
 132             flashvars = dict(
 133                 (k, v[0])
 134                 for k, v in compat_parse_qs(fv_el.attrib['value']).items())
 135         else:
 136             flashvars = {}
 137
 138         def find_param(name):
 139             if name in flashvars:
 140                 return flashvars[name]
 141             node = find_xpath_attr(object_doc, './param', 'name', name)
 142             if node is not None:
 143                 return node.attrib['value']
 144             return None
 145
 146         params = {}
 147
 148         playerID = find_param('playerID')
 149         if playerID is None:
 150             raise ExtractorError('Cannot find player ID')
 151         params['playerID'] = playerID
 152
 153         playerKey = find_param('playerKey')
 154         # Not all pages define this value
 155         if playerKey is not None:
 156             params['playerKey'] = playerKey
 157         # The three fields hold the id of the video
 158         videoPlayer = find_param('@videoPlayer') or find_param('videoId') or find_param('videoID')
 159         if videoPlayer is not None:
 160             params['@videoPlayer'] = videoPlayer
 161         linkBase = find_param('linkBaseURL')
 162         if linkBase is not None:
 163             params['linkBaseURL'] = linkBase
 164         return cls._make_brightcove_url(params)
 165
 166     @classmethod
 167     def _build_brighcove_url_from_js(cls, object_js):
 168         # The layout of JS is as follows:
 169         # customBC.createVideo = function (width, height, playerID, playerKey, videoPlayer, VideoRandomID) {
 170         #   // build Brightcove <object /> XML
 171         # }
 172         m = re.search(
 173             r'''(?x)customBC.\createVideo\(
 174                 .*?                                                  # skipping width and height
 175                 ["\'](?P<playerID>\d+)["\']\s*,\s*                   # playerID
 176                 ["\'](?P<playerKey>AQ[^"\']{48})[^"\']*["\']\s*,\s*  # playerKey begins with AQ and is 50 characters
 177                                                                      # in length, however it's appended to itself
 178                                                                      # in places, so truncate
 179                 ["\'](?P<videoID>\d+)["\']                           # @videoPlayer
 180             ''', object_js)
 181         if m:
 182             return cls._make_brightcove_url(m.groupdict())
 183
 184     @classmethod
 185     def _make_brightcove_url(cls, params):
 186         data = compat_urllib_parse.urlencode(params)
 187         return cls._FEDERATED_URL_TEMPLATE % data
 188
 189     @classmethod
 190     def _extract_brightcove_url(cls, webpage):
 191         """Try to extract the brightcove url from the webpage, returns None
 192         if it can't be found
 193         """
 194         urls = cls._extract_brightcove_urls(webpage)
 195         return urls[0] if urls else None
 196
 197     @classmethod
 198     def _extract_brightcove_urls(cls, webpage):
 199         """Return a list of all Brightcove URLs from the webpage """
 200
 201         url_m = re.search(
 202             r'<meta\s+property=[\'"]og:video[\'"]\s+content=[\'"](https?://(?:secure|c)\.brightcove.com/[^\'"]+)[\'"]',
 203             webpage)
 204         if url_m:
 205             url = unescapeHTML(url_m.group(1))
 206             # Some sites don't add it, we can't download with this url, for example:
 207             # http://www.ktvu.com/videos/news/raw-video-caltrain-releases-video-of-man-almost/vCTZdY/
 208             if 'playerKey' in url or 'videoId' in url:
 209                 return [url]
 210
 211         matches = re.findall(
 212             r'''(?sx)<object
 213             (?:
 214                 [^>]+?class=[\'"][^>]*?BrightcoveExperience.*?[\'"] |
 215                 [^>]*?>\s*<param\s+name="movie"\s+value="https?://[^/]*brightcove\.com/
 216             ).+?>\s*</object>''',
 217             webpage)
 218         if matches:
 219             return list(filter(None, [cls._build_brighcove_url(m) for m in matches]))
 220
 221         return list(filter(None, [
 222             cls._build_brighcove_url_from_js(custom_bc)
 223             for custom_bc in re.findall(r'(customBC\.createVideo\(.+?\);)', webpage)]))
 224
 225     def _real_extract(self, url):
 226         url, smuggled_data = unsmuggle_url(url, {})
 227
 228         # Change the 'videoId' and others field to '@videoPlayer'
 229         url = re.sub(r'(?<=[?&])(videoI(d|D)|bctid)', '%40videoPlayer', url)
 230         # Change bckey (used by bcove.me urls) to playerKey
 231         url = re.sub(r'(?<=[?&])bckey', 'playerKey', url)
 232         mobj = re.match(self._VALID_URL, url)
 233         query_str = mobj.group('query')
 234         query = compat_urlparse.parse_qs(query_str)
 235
 236         videoPlayer = query.get('@videoPlayer')
 237         if videoPlayer:
 238             # We set the original url as the default 'Referer' header
 239             referer = smuggled_data.get('Referer', url)
 240             return self._get_video_info(
 241                 videoPlayer[0], query_str, query, referer=referer)
 242         elif 'playerKey' in query:
 243             player_key = query['playerKey']
 244             return self._get_playlist_info(player_key[0])
 245         else:
 246             raise ExtractorError(
 247                 'Cannot find playerKey= variable. Did you forget quotes in a shell invocation?',
 248                 expected=True)
 249
 250     def _get_video_info(self, video_id, query_str, query, referer=None):
 251         request_url = self._FEDERATED_URL_TEMPLATE % query_str
 252         req = compat_urllib_request.Request(request_url)
 253         linkBase = query.get('linkBaseURL')
 254         if linkBase is not None:
 255             referer = linkBase[0]
 256         if referer is not None:
 257             req.add_header('Referer', referer)
 258         webpage = self._download_webpage(req, video_id)
 259
 260         error_msg = self._html_search_regex(
 261             r"<h1>We're sorry.</h1>([\s\n]*<p>.*?</p>)+", webpage,
 262             'error message', default=None)
 263         if error_msg is not None:
 264             raise ExtractorError(
 265                 'brightcove said: %s' % error_msg, expected=True)
 266
 267         self.report_extraction(video_id)
 268         info = self._search_regex(r'var experienceJSON = ({.*});', webpage, 'json')
 269         info = json.loads(info)['data']
 270         video_info = info['programmedContent']['videoPlayer']['mediaDTO']
 271         video_info['_youtubedl_adServerURL'] = info.get('adServerURL')
 272
 273         return self._extract_video_info(video_info)
 274
 275     def _get_playlist_info(self, player_key):
 276         info_url = 'http://c.brightcove.com/services/json/experience/runtime/?command=get_programming_for_experience&playerKey=%s' % player_key
 277         playlist_info = self._download_webpage(
 278             info_url, player_key, 'Downloading playlist information')
 279
 280         json_data = json.loads(playlist_info)
 281         if 'videoList' not in json_data:
 282             raise ExtractorError('Empty playlist')
 283         playlist_info = json_data['videoList']
 284         videos = [self._extract_video_info(video_info) for video_info in playlist_info['mediaCollectionDTO']['videoDTOs']]
 285
 286         return self.playlist_result(videos, playlist_id='%s' % playlist_info['id'],
 287                                     playlist_title=playlist_info['mediaCollectionDTO']['displayName'])
 288
 289     def _extract_video_info(self, video_info):
 290         info = {
 291             'id': compat_str(video_info['id']),
 292             'title': video_info['displayName'].strip(),
 293             'description': video_info.get('shortDescription'),
 294             'thumbnail': video_info.get('videoStillURL') or video_info.get('thumbnailURL'),
 295             'uploader': video_info.get('publisherName'),
 296         }
 297
 298         renditions = video_info.get('renditions')
 299         if renditions:
 300             formats = []
 301             for rend in renditions:
 302                 url = rend['defaultURL']
 303                 if not url:
 304                     continue
 305                 ext = None
 306                 if rend['remote']:
 307                     url_comp = compat_urllib_parse_urlparse(url)
 308                     if url_comp.path.endswith('.m3u8'):
 309                         formats.extend(
 310                             self._extract_m3u8_formats(url, info['id'], 'mp4'))
 311                         continue
 312                     elif 'akamaihd.net' in url_comp.netloc:
 313                         # This type of renditions are served through
 314                         # akamaihd.net, but they don't use f4m manifests
 315                         url = url.replace('control/', '') + '?&v=3.3.0&fp=13&r=FEEFJ&g=RTSJIMBMPFPB'
 316                         ext = 'flv'
 317                 if ext is None:
 318                     ext = determine_ext(url)
 319                 size = rend.get('size')
 320                 formats.append({
 321                     'url': url,
 322                     'ext': ext,
 323                     'height': rend.get('frameHeight'),
 324                     'width': rend.get('frameWidth'),
 325                     'filesize': size if size != 0 else None,
 326                 })
 327             self._sort_formats(formats)
 328             info['formats'] = formats
 329         elif video_info.get('FLVFullLengthURL') is not None:
 330             info.update({
 331                 'url': video_info['FLVFullLengthURL'],
 332             })
 333
 334         if self._downloader.params.get('include_ads', False):
 335             adServerURL = video_info.get('_youtubedl_adServerURL')
 336             if adServerURL:
 337                 ad_info = {
 338                     '_type': 'url',
 339                     'url': adServerURL,
 340                 }
 341                 if 'url' in info:
 342                     return {
 343                         '_type': 'playlist',
 344                         'title': info['title'],
 345                         'entries': [ad_info, info],
 346                     }
 347                 else:
 348                     return ad_info
 349
 350         if 'url' not in info and not info.get('formats'):
 351             raise ExtractorError('Unable to extract video url for %s' % info['id'])
 352         return info
 353
 354
 355 class BrightcoveInPageEmbedIE(InfoExtractor):
 356     _VALID_URL = r'https?://players\.brightcove\.net/(?P<account_id>\d+)/([a-z0-9-]+)_([a-z]+)/index.html?.*videoId=(?P<video_id>\d+)'
 357     TEST = {
 358         'url': 'http://players.brightcove.net/929656772001/e41d32dc-ec74-459e-a845-6c69f7b724ea_default/index.html?videoId=4463358922001',
 359         'info_dict': {
 360             'id': '4463358922001',
 361             'ext': 'flv',
 362             'title': 'Meet the man behind Popcorn Time',
 363             'description': 'md5:a950cc4285c43e44d763d036710cd9cd',
 364             'duration': 165768,
 365         }
 366     }
 367
 368     @staticmethod
 369     def _extract_url(webpage):
 370         video_attributes = re.search(r'(?s)<video([^>]*)>.*?</(?:video|audio)>', webpage)
 371         if video_attributes:
 372             video_attributes = extract_attributes(video_attributes.group(), r'(?s)\s*data-(account|video-id|playlist-id|policy-key|player|embed)\s*=\s*["\']([^"\']+)["\']')
 373             account_id = video_attributes.get('account')
 374             player_id = video_attributes.get('player')
 375             embed = video_attributes.get('embed')
 376             video_id = video_attributes.get('video-id')
 377             if account_id and player_id and embed and video_id:
 378                 return 'http://players.brightcove.net/%s/%s_%s/index.html?videoId=%s' % (account_id, player_id, embed, video_id)
 379         return None
 380
 381     def _real_extract(self, url):
 382         mobj = re.match(self._VALID_URL, url)
 383         account_id, player_id, embed, video_id = mobj.groups()
 384
 385         webpage = self._download_webpage('http://players.brightcove.net/%s/%s_%s/index.min.js' % (account_id, player_id, embed), video_id)
 386
 387         catalog = self._parse_json(
 388             js_to_json(
 389                 self._search_regex(
 390                     r'catalog\(({[^}]+})\);',
 391                     webpage,
 392                     'catalog'
 393                 )
 394             ),
 395             video_id
 396         )
 397         policy_key = catalog['policyKey']
 398
 399         req = compat_urllib_request.Request(
 400             'https://edge.api.brightcove.com/playback/v1/accounts/%s/videos/%s' % (account_id, video_id),
 401             headers={'Accept': 'application/json;pk=%s' % policy_key})
 402         json_data = self._download_json(req, video_id)
 403
 404         title = json_data['name']
 405         description = json_data.get('description')
 406         thumbnail = json_data.get('name')
 407         timestamp = parse_iso8601(json_data.get('published_at'))
 408         duration = int_or_none(json_data.get('duration'))
 409
 410         formats = []
 411         for source in json_data.get('sources'):
 412             source_type = source.get('type')
 413             if source_type == 'application/x-mpegURL':
 414                 formats.extend(self._extract_m3u8_formats(source.get('src'), video_id))
 415             else:
 416                 src = source.get('src') or source.get('streaming_src')
 417                 if src:
 418                     formats.append({
 419                         'url': src,
 420                         'abr': source.get('avg_bitrate'),
 421                         'width': int_or_none(source.get('width')),
 422                         'height': int_or_none(source.get('height')),
 423                         'filesize': source.get('size'),
 424                         'container': source.get('container'),
 425                         'vcodec': source.get('container'),
 426                     })
 427
 428         self._sort_formats(formats)
 429
 430         return {
 431             'id': video_id,
 432             'title': title,
 433             'description': description,
 434             'thumbnail': thumbnail,
 435             'timestamp': timestamp,
 436             'duration': duration,
 437             'formats': formats,
 438         }