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