Raphaël G. Git Repositories - youtubedl/blob - youtube_dl/extractor/brightcove.py

   1 # encoding: 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,
  13     compat_urllib_parse_urlparse,
  14     compat_urlparse,
  15     compat_xml_parse_error,
  16 )
  17 from ..utils import (
  18     determine_ext,
  19     ExtractorError,
  20     find_xpath_attr,
  21     fix_xml_ampersands,
  22     float_or_none,
  23     js_to_json,
  24     int_or_none,
  25     parse_iso8601,
  26     sanitized_Request,
  27     unescapeHTML,
  28     unsmuggle_url,
  29 )
  30
  31
  32 class BrightcoveLegacyIE(InfoExtractor):
  33     IE_NAME = 'brightcove:legacy'
  34     _VALID_URL = r'(?:https?://.*brightcove\.com/(services|viewer).*?\?|brightcove:)(?P<query>.*)'
  35     _FEDERATED_URL_TEMPLATE = 'http://c.brightcove.com/services/viewer/htmlFederated?%s'
  36
  37     _TESTS = [
  38         {
  39             # From http://www.8tv.cat/8aldia/videos/xavier-sala-i-martin-aquesta-tarda-a-8-al-dia/
  40             'url': 'http://c.brightcove.com/services/viewer/htmlFederated?playerID=1654948606001&flashID=myExperience&%40videoPlayer=2371591881001',
  41             'md5': '5423e113865d26e40624dce2e4b45d95',
  42             'note': 'Test Brightcove downloads and detection in GenericIE',
  43             'info_dict': {
  44                 'id': '2371591881001',
  45                 'ext': 'mp4',
  46                 'title': 'Xavier Sala i Martín: “Un banc que no presta és un banc zombi que no serveix per a res”',
  47                 'uploader': '8TV',
  48                 'description': 'md5:a950cc4285c43e44d763d036710cd9cd',
  49             }
  50         },
  51         {
  52             # From http://medianetwork.oracle.com/video/player/1785452137001
  53             'url': 'http://c.brightcove.com/services/viewer/htmlFederated?playerID=1217746023001&flashID=myPlayer&%40videoPlayer=1785452137001',
  54             'info_dict': {
  55                 'id': '1785452137001',
  56                 'ext': 'flv',
  57                 'title': 'JVMLS 2012: Arrays 2.0 - Opportunities and Challenges',
  58                 'description': 'John Rose speaks at the JVM Language Summit, August 1, 2012.',
  59                 'uploader': 'Oracle',
  60             },
  61         },
  62         {
  63             # From http://mashable.com/2013/10/26/thermoelectric-bracelet-lets-you-control-your-body-temperature/
  64             'url': 'http://c.brightcove.com/services/viewer/federated_f9?&playerID=1265504713001&publisherID=AQ%7E%7E%2CAAABBzUwv1E%7E%2CxP-xFHVUstiMFlNYfvF4G9yFnNaqCw_9&videoID=2750934548001',
  65             'info_dict': {
  66                 'id': '2750934548001',
  67                 'ext': 'mp4',
  68                 'title': 'This Bracelet Acts as a Personal Thermostat',
  69                 'description': 'md5:547b78c64f4112766ccf4e151c20b6a0',
  70                 'uploader': 'Mashable',
  71             },
  72         },
  73         {
  74             # test that the default referer works
  75             # from http://national.ballet.ca/interact/video/Lost_in_Motion_II/
  76             'url': 'http://link.brightcove.com/services/player/bcpid756015033001?bckey=AQ~~,AAAApYJi_Ck~,GxhXCegT1Dp39ilhXuxMJxasUhVNZiil&bctid=2878862109001',
  77             'info_dict': {
  78                 'id': '2878862109001',
  79                 'ext': 'mp4',
  80                 'title': 'Lost in Motion II',
  81                 'description': 'md5:363109c02998fee92ec02211bd8000df',
  82                 'uploader': 'National Ballet of Canada',
  83             },
  84         },
  85         {
  86             # test flv videos served by akamaihd.net
  87             # From http://www.redbull.com/en/bike/stories/1331655643987/replay-uci-dh-world-cup-2014-from-fort-william
  88             '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',
  89             # The md5 checksum changes on each download
  90             'info_dict': {
  91                 'id': '2996102916001',
  92                 'ext': 'flv',
  93                 'title': 'UCI MTB World Cup 2014: Fort William, UK - Downhill Finals',
  94                 'uploader': 'Red Bull TV',
  95                 'description': 'UCI MTB World Cup 2014: Fort William, UK - Downhill Finals',
  96             },
  97         },
  98         {
  99             # playlist test
 100             # from http://support.brightcove.com/en/video-cloud/docs/playlist-support-single-video-players
 101             'url': 'http://c.brightcove.com/services/viewer/htmlFederated?playerID=3550052898001&playerKey=AQ%7E%7E%2CAAABmA9XpXk%7E%2C-Kp7jNgisre1fG5OdqpAFUTcs0lP_ZoL',
 102             'info_dict': {
 103                 'title': 'Sealife',
 104                 'id': '3550319591001',
 105             },
 106             'playlist_mincount': 7,
 107         },
 108     ]
 109
 110     @classmethod
 111     def _build_brighcove_url(cls, object_str):
 112         """
 113         Build a Brightcove url from a xml string containing
 114         <object class="BrightcoveExperience">{params}</object>
 115         """
 116
 117         # Fix up some stupid HTML, see https://github.com/rg3/youtube-dl/issues/1553
 118         object_str = re.sub(r'(<param(?:\s+[a-zA-Z0-9_]+="[^"]*")*)>',
 119                             lambda m: m.group(1) + '/>', object_str)
 120         # Fix up some stupid XML, see https://github.com/rg3/youtube-dl/issues/1608
 121         object_str = object_str.replace('<--', '<!--')
 122         # remove namespace to simplify extraction
 123         object_str = re.sub(r'(<object[^>]*)(xmlns=".*?")', r'\1', object_str)
 124         object_str = fix_xml_ampersands(object_str)
 125
 126         try:
 127             object_doc = compat_etree_fromstring(object_str.encode('utf-8'))
 128         except compat_xml_parse_error:
 129             return
 130
 131         fv_el = find_xpath_attr(object_doc, './param', 'name', 'flashVars')
 132         if fv_el is not None:
 133             flashvars = dict(
 134                 (k, v[0])
 135                 for k, v in compat_parse_qs(fv_el.attrib['value']).items())
 136         else:
 137             flashvars = {}
 138
 139         def find_param(name):
 140             if name in flashvars:
 141                 return flashvars[name]
 142             node = find_xpath_attr(object_doc, './param', 'name', name)
 143             if node is not None:
 144                 return node.attrib['value']
 145             return None
 146
 147         params = {}
 148
 149         playerID = find_param('playerID')
 150         if playerID is None:
 151             raise ExtractorError('Cannot find player ID')
 152         params['playerID'] = playerID
 153
 154         playerKey = find_param('playerKey')
 155         # Not all pages define this value
 156         if playerKey is not None:
 157             params['playerKey'] = playerKey
 158         # The three fields hold the id of the video
 159         videoPlayer = find_param('@videoPlayer') or find_param('videoId') or find_param('videoID')
 160         if videoPlayer is not None:
 161             params['@videoPlayer'] = videoPlayer
 162         linkBase = find_param('linkBaseURL')
 163         if linkBase is not None:
 164             params['linkBaseURL'] = linkBase
 165         return cls._make_brightcove_url(params)
 166
 167     @classmethod
 168     def _build_brighcove_url_from_js(cls, object_js):
 169         # The layout of JS is as follows:
 170         # customBC.createVideo = function (width, height, playerID, playerKey, videoPlayer, VideoRandomID) {
 171         #   // build Brightcove <object /> XML
 172         # }
 173         m = re.search(
 174             r'''(?x)customBC.\createVideo\(
 175                 .*?                                                  # skipping width and height
 176                 ["\'](?P<playerID>\d+)["\']\s*,\s*                   # playerID
 177                 ["\'](?P<playerKey>AQ[^"\']{48})[^"\']*["\']\s*,\s*  # playerKey begins with AQ and is 50 characters
 178                                                                      # in length, however it's appended to itself
 179                                                                      # in places, so truncate
 180                 ["\'](?P<videoID>\d+)["\']                           # @videoPlayer
 181             ''', object_js)
 182         if m:
 183             return cls._make_brightcove_url(m.groupdict())
 184
 185     @classmethod
 186     def _make_brightcove_url(cls, params):
 187         data = compat_urllib_parse.urlencode(params)
 188         return cls._FEDERATED_URL_TEMPLATE % data
 189
 190     @classmethod
 191     def _extract_brightcove_url(cls, webpage):
 192         """Try to extract the brightcove url from the webpage, returns None
 193         if it can't be found
 194         """
 195         urls = cls._extract_brightcove_urls(webpage)
 196         return urls[0] if urls else None
 197
 198     @classmethod
 199     def _extract_brightcove_urls(cls, webpage):
 200         """Return a list of all Brightcove URLs from the webpage """
 201
 202         url_m = re.search(
 203             r'<meta\s+property=[\'"]og:video[\'"]\s+content=[\'"](https?://(?:secure|c)\.brightcove.com/[^\'"]+)[\'"]',
 204             webpage)
 205         if url_m:
 206             url = unescapeHTML(url_m.group(1))
 207             # Some sites don't add it, we can't download with this url, for example:
 208             # http://www.ktvu.com/videos/news/raw-video-caltrain-releases-video-of-man-almost/vCTZdY/
 209             if 'playerKey' in url or 'videoId' in url:
 210                 return [url]
 211
 212         matches = re.findall(
 213             r'''(?sx)<object
 214             (?:
 215                 [^>]+?class=[\'"][^>]*?BrightcoveExperience.*?[\'"] |
 216                 [^>]*?>\s*<param\s+name="movie"\s+value="https?://[^/]*brightcove\.com/
 217             ).+?>\s*</object>''',
 218             webpage)
 219         if matches:
 220             return list(filter(None, [cls._build_brighcove_url(m) for m in matches]))
 221
 222         return list(filter(None, [
 223             cls._build_brighcove_url_from_js(custom_bc)
 224             for custom_bc in re.findall(r'(customBC\.createVideo\(.+?\);)', webpage)]))
 225
 226     def _real_extract(self, url):
 227         url, smuggled_data = unsmuggle_url(url, {})
 228
 229         # Change the 'videoId' and others field to '@videoPlayer'
 230         url = re.sub(r'(?<=[?&])(videoI(d|D)|bctid)', '%40videoPlayer', url)
 231         # Change bckey (used by bcove.me urls) to playerKey
 232         url = re.sub(r'(?<=[?&])bckey', 'playerKey', url)
 233         mobj = re.match(self._VALID_URL, url)
 234         query_str = mobj.group('query')
 235         query = compat_urlparse.parse_qs(query_str)
 236
 237         videoPlayer = query.get('@videoPlayer')
 238         if videoPlayer:
 239             # We set the original url as the default 'Referer' header
 240             referer = smuggled_data.get('Referer', url)
 241             return self._get_video_info(
 242                 videoPlayer[0], query_str, query, referer=referer)
 243         elif 'playerKey' in query:
 244             player_key = query['playerKey']
 245             return self._get_playlist_info(player_key[0])
 246         else:
 247             raise ExtractorError(
 248                 'Cannot find playerKey= variable. Did you forget quotes in a shell invocation?',
 249                 expected=True)
 250
 251     def _get_video_info(self, video_id, query_str, query, referer=None):
 252         request_url = self._FEDERATED_URL_TEMPLATE % query_str
 253         req = sanitized_Request(request_url)
 254         linkBase = query.get('linkBaseURL')
 255         if linkBase is not None:
 256             referer = linkBase[0]
 257         if referer is not None:
 258             req.add_header('Referer', referer)
 259         webpage = self._download_webpage(req, video_id)
 260
 261         error_msg = self._html_search_regex(
 262             r"<h1>We're sorry.</h1>([\s\n]*<p>.*?</p>)+", webpage,
 263             'error message', default=None)
 264         if error_msg is not None:
 265             raise ExtractorError(
 266                 'brightcove said: %s' % error_msg, expected=True)
 267
 268         self.report_extraction(video_id)
 269         info = self._search_regex(r'var experienceJSON = ({.*});', webpage, 'json')
 270         info = json.loads(info)['data']
 271         video_info = info['programmedContent']['videoPlayer']['mediaDTO']
 272         video_info['_youtubedl_adServerURL'] = info.get('adServerURL')
 273
 274         return self._extract_video_info(video_info)
 275
 276     def _get_playlist_info(self, player_key):
 277         info_url = 'http://c.brightcove.com/services/json/experience/runtime/?command=get_programming_for_experience&playerKey=%s' % player_key
 278         playlist_info = self._download_webpage(
 279             info_url, player_key, 'Downloading playlist information')
 280
 281         json_data = json.loads(playlist_info)
 282         if 'videoList' not in json_data:
 283             raise ExtractorError('Empty playlist')
 284         playlist_info = json_data['videoList']
 285         videos = [self._extract_video_info(video_info) for video_info in playlist_info['mediaCollectionDTO']['videoDTOs']]
 286
 287         return self.playlist_result(videos, playlist_id='%s' % playlist_info['id'],
 288                                     playlist_title=playlist_info['mediaCollectionDTO']['displayName'])
 289
 290     def _extract_video_info(self, video_info):
 291         info = {
 292             'id': compat_str(video_info['id']),
 293             'title': video_info['displayName'].strip(),
 294             'description': video_info.get('shortDescription'),
 295             'thumbnail': video_info.get('videoStillURL') or video_info.get('thumbnailURL'),
 296             'uploader': video_info.get('publisherName'),
 297         }
 298
 299         renditions = video_info.get('renditions')
 300         if renditions:
 301             formats = []
 302             for rend in renditions:
 303                 url = rend['defaultURL']
 304                 if not url:
 305                     continue
 306                 ext = None
 307                 if rend['remote']:
 308                     url_comp = compat_urllib_parse_urlparse(url)
 309                     if url_comp.path.endswith('.m3u8'):
 310                         formats.extend(
 311                             self._extract_m3u8_formats(url, info['id'], 'mp4'))
 312                         continue
 313                     elif 'akamaihd.net' in url_comp.netloc:
 314                         # This type of renditions are served through
 315                         # akamaihd.net, but they don't use f4m manifests
 316                         url = url.replace('control/', '') + '?&v=3.3.0&fp=13&r=FEEFJ&g=RTSJIMBMPFPB'
 317                         ext = 'flv'
 318                 if ext is None:
 319                     ext = determine_ext(url)
 320                 size = rend.get('size')
 321                 formats.append({
 322                     'url': url,
 323                     'ext': ext,
 324                     'height': rend.get('frameHeight'),
 325                     'width': rend.get('frameWidth'),
 326                     'filesize': size if size != 0 else None,
 327                 })
 328             self._sort_formats(formats)
 329             info['formats'] = formats
 330         elif video_info.get('FLVFullLengthURL') is not None:
 331             info.update({
 332                 'url': video_info['FLVFullLengthURL'],
 333             })
 334
 335         if self._downloader.params.get('include_ads', False):
 336             adServerURL = video_info.get('_youtubedl_adServerURL')
 337             if adServerURL:
 338                 ad_info = {
 339                     '_type': 'url',
 340                     'url': adServerURL,
 341                 }
 342                 if 'url' in info:
 343                     return {
 344                         '_type': 'playlist',
 345                         'title': info['title'],
 346                         'entries': [ad_info, info],
 347                     }
 348                 else:
 349                     return ad_info
 350
 351         if 'url' not in info and not info.get('formats'):
 352             raise ExtractorError('Unable to extract video url for %s' % info['id'])
 353         return info
 354
 355
 356 class BrightcoveNewIE(InfoExtractor):
 357     IE_NAME = 'brightcove:new'
 358     _VALID_URL = r'https?://players\.brightcove\.net/(?P<account_id>\d+)/(?P<player_id>[^/]+)_(?P<embed>[^/]+)/index\.html\?.*videoId=(?P<video_id>\d+)'
 359     _TESTS = [{
 360         'url': 'http://players.brightcove.net/929656772001/e41d32dc-ec74-459e-a845-6c69f7b724ea_default/index.html?videoId=4463358922001',
 361         'md5': 'c8100925723840d4b0d243f7025703be',
 362         'info_dict': {
 363             'id': '4463358922001',
 364             'ext': 'mp4',
 365             'title': 'Meet the man behind Popcorn Time',
 366             'description': 'md5:eac376a4fe366edc70279bfb681aea16',
 367             'duration': 165.768,
 368             'timestamp': 1441391203,
 369             'upload_date': '20150904',
 370             'uploader_id': '929656772001',
 371             'formats': 'mincount:22',
 372         },
 373     }, {
 374         # with rtmp streams
 375         'url': 'http://players.brightcove.net/4036320279001/5d112ed9-283f-485f-a7f9-33f42e8bc042_default/index.html?videoId=4279049078001',
 376         'info_dict': {
 377             'id': '4279049078001',
 378             'ext': 'mp4',
 379             'title': 'Titansgrave: Chapter 0',
 380             'description': 'Titansgrave: Chapter 0',
 381             'duration': 1242.058,
 382             'timestamp': 1433556729,
 383             'upload_date': '20150606',
 384             'uploader_id': '4036320279001',
 385             'formats': 'mincount:41',
 386         },
 387         'params': {
 388             'skip_download': True,
 389         }
 390     }]
 391
 392     @staticmethod
 393     def _extract_urls(webpage):
 394         # Reference:
 395         # 1. http://docs.brightcove.com/en/video-cloud/brightcove-player/guides/publish-video.html#setvideoiniframe
 396         # 2. http://docs.brightcove.com/en/video-cloud/brightcove-player/guides/publish-video.html#setvideousingjavascript)
 397         # 3. http://docs.brightcove.com/en/video-cloud/brightcove-player/guides/embed-in-page.html
 398
 399         entries = []
 400
 401         # Look for iframe embeds [1]
 402         for _, url in re.findall(
 403                 r'<iframe[^>]+src=(["\'])((?:https?:)//players\.brightcove\.net/\d+/[^/]+/index\.html.+?)\1', webpage):
 404             entries.append(url)
 405
 406         # Look for embed_in_page embeds [2]
 407         for video_id, account_id, player_id, embed in re.findall(
 408                 # According to examples from [3] it's unclear whether video id
 409                 # may be optional and what to do when it is
 410                 r'''(?sx)
 411                     <video[^>]+
 412                         data-video-id=["\'](\d+)["\'][^>]*>.*?
 413                     </video>.*?
 414                     <script[^>]+
 415                         src=["\'](?:https?:)?//players\.brightcove\.net/
 416                         (\d+)/([\da-f-]+)_([^/]+)/index\.min\.js
 417                 ''', webpage):
 418             entries.append(
 419                 'http://players.brightcove.net/%s/%s_%s/index.html?videoId=%s'
 420                 % (account_id, player_id, embed, video_id))
 421
 422         return entries
 423
 424     def _real_extract(self, url):
 425         account_id, player_id, embed, video_id = re.match(self._VALID_URL, url).groups()
 426
 427         webpage = self._download_webpage(
 428             'http://players.brightcove.net/%s/%s_%s/index.min.js'
 429             % (account_id, player_id, embed), video_id)
 430
 431         policy_key = None
 432
 433         catalog = self._search_regex(
 434             r'catalog\(({.+?})\);', webpage, 'catalog', default=None)
 435         if catalog:
 436             catalog = self._parse_json(
 437                 js_to_json(catalog), video_id, fatal=False)
 438             if catalog:
 439                 policy_key = catalog.get('policyKey')
 440
 441         if not policy_key:
 442             policy_key = self._search_regex(
 443                 r'policyKey\s*:\s*(["\'])(?P<pk>.+?)\1',
 444                 webpage, 'policy key', group='pk')
 445
 446         req = sanitized_Request(
 447             'https://edge.api.brightcove.com/playback/v1/accounts/%s/videos/%s'
 448             % (account_id, video_id),
 449             headers={'Accept': 'application/json;pk=%s' % policy_key})
 450         json_data = self._download_json(req, video_id)
 451
 452         title = json_data['name']
 453
 454         formats = []
 455         for source in json_data.get('sources', []):
 456             source_type = source.get('type')
 457             src = source.get('src')
 458             if source_type == 'application/x-mpegURL':
 459                 if not src:
 460                     continue
 461                 m3u8_formats = self._extract_m3u8_formats(
 462                     src, video_id, 'mp4', entry_protocol='m3u8_native',
 463                     m3u8_id='hls', fatal=False)
 464                 if m3u8_formats:
 465                     formats.extend(m3u8_formats)
 466             else:
 467                 streaming_src = source.get('streaming_src')
 468                 stream_name, app_name = source.get('stream_name'), source.get('app_name')
 469                 if not src and not streaming_src and (not stream_name or not app_name):
 470                     continue
 471                 tbr = float_or_none(source.get('avg_bitrate'), 1000)
 472                 height = int_or_none(source.get('height'))
 473                 f = {
 474                     'tbr': tbr,
 475                     'width': int_or_none(source.get('width')),
 476                     'height': height,
 477                     'filesize': int_or_none(source.get('size')),
 478                     'container': source.get('container'),
 479                     'vcodec': source.get('codec'),
 480                     'ext': source.get('container').lower(),
 481                 }
 482
 483                 def build_format_id(kind):
 484                     format_id = kind
 485                     if tbr:
 486                         format_id += '-%dk' % int(tbr)
 487                     if height:
 488                         format_id += '-%dp' % height
 489                     return format_id
 490
 491                 if src or streaming_src:
 492                     f.update({
 493                         'url': src or streaming_src,
 494                         'format_id': build_format_id('http' if src else 'http-streaming'),
 495                         'preference': 2 if src else 1,
 496                     })
 497                 else:
 498                     f.update({
 499                         'url': app_name,
 500                         'play_path': stream_name,
 501                         'format_id': build_format_id('rtmp'),
 502                     })
 503                 formats.append(f)
 504         self._sort_formats(formats)
 505
 506         description = json_data.get('description')
 507         thumbnail = json_data.get('thumbnail')
 508         timestamp = parse_iso8601(json_data.get('published_at'))
 509         duration = float_or_none(json_data.get('duration'), 1000)
 510         tags = json_data.get('tags', [])
 511
 512         return {
 513             'id': video_id,
 514             'title': title,
 515             'description': description,
 516             'thumbnail': thumbnail,
 517             'duration': duration,
 518             'timestamp': timestamp,
 519             'uploader_id': account_id,
 520             'formats': formats,
 521             'tags': tags,
 522         }