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 ..utils import (
  10     compat_urllib_parse,
  11     find_xpath_attr,
  12     fix_xml_ampersands,
  13     compat_urlparse,
  14     compat_str,
  15     compat_urllib_request,
  16     compat_parse_qs,
  17
  18     determine_ext,
  19     ExtractorError,
  20     unsmuggle_url,
  21     unescapeHTML,
  22 )
  23
  24
  25 class BrightcoveIE(InfoExtractor):
  26     _VALID_URL = r'https?://.*brightcove\.com/(services|viewer).*\?(?P<query>.*)'
  27     _FEDERATED_URL_TEMPLATE = 'http://c.brightcove.com/services/viewer/htmlFederated?%s'
  28
  29     _TESTS = [
  30         {
  31             # From http://www.8tv.cat/8aldia/videos/xavier-sala-i-martin-aquesta-tarda-a-8-al-dia/
  32             'url': 'http://c.brightcove.com/services/viewer/htmlFederated?playerID=1654948606001&flashID=myExperience&%40videoPlayer=2371591881001',
  33             'md5': '5423e113865d26e40624dce2e4b45d95',
  34             'note': 'Test Brightcove downloads and detection in GenericIE',
  35             'info_dict': {
  36                 'id': '2371591881001',
  37                 'ext': 'mp4',
  38                 'title': 'Xavier Sala i Martín: “Un banc que no presta és un banc zombi que no serveix per a res”',
  39                 'uploader': '8TV',
  40                 'description': 'md5:a950cc4285c43e44d763d036710cd9cd',
  41             }
  42         },
  43         {
  44             # From http://medianetwork.oracle.com/video/player/1785452137001
  45             'url': 'http://c.brightcove.com/services/viewer/htmlFederated?playerID=1217746023001&flashID=myPlayer&%40videoPlayer=1785452137001',
  46             'info_dict': {
  47                 'id': '1785452137001',
  48                 'ext': 'flv',
  49                 'title': 'JVMLS 2012: Arrays 2.0 - Opportunities and Challenges',
  50                 'description': 'John Rose speaks at the JVM Language Summit, August 1, 2012.',
  51                 'uploader': 'Oracle',
  52             },
  53         },
  54         {
  55             # From http://mashable.com/2013/10/26/thermoelectric-bracelet-lets-you-control-your-body-temperature/
  56             'url': 'http://c.brightcove.com/services/viewer/federated_f9?&playerID=1265504713001&publisherID=AQ%7E%7E%2CAAABBzUwv1E%7E%2CxP-xFHVUstiMFlNYfvF4G9yFnNaqCw_9&videoID=2750934548001',
  57             'info_dict': {
  58                 'id': '2750934548001',
  59                 'ext': 'mp4',
  60                 'title': 'This Bracelet Acts as a Personal Thermostat',
  61                 'description': 'md5:547b78c64f4112766ccf4e151c20b6a0',
  62                 'uploader': 'Mashable',
  63             },
  64         },
  65         {
  66             # test that the default referer works
  67             # from http://national.ballet.ca/interact/video/Lost_in_Motion_II/
  68             'url': 'http://link.brightcove.com/services/player/bcpid756015033001?bckey=AQ~~,AAAApYJi_Ck~,GxhXCegT1Dp39ilhXuxMJxasUhVNZiil&bctid=2878862109001',
  69             'info_dict': {
  70                 'id': '2878862109001',
  71                 'ext': 'mp4',
  72                 'title': 'Lost in Motion II',
  73                 'description': 'md5:363109c02998fee92ec02211bd8000df',
  74                 'uploader': 'National Ballet of Canada',
  75             },
  76         },
  77         {
  78             # test flv videos served by akamaihd.net
  79             # From http://www.redbull.com/en/bike/stories/1331655643987/replay-uci-dh-world-cup-2014-from-fort-william
  80             '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',
  81             # The md5 checksum changes on each download
  82             'info_dict': {
  83                 'id': '2996102916001',
  84                 'ext': 'flv',
  85                 'title': 'UCI MTB World Cup 2014: Fort William, UK - Downhill Finals',
  86                 'uploader': 'Red Bull TV',
  87                 'description': 'UCI MTB World Cup 2014: Fort William, UK - Downhill Finals',
  88             },
  89         },
  90     ]
  91
  92     @classmethod
  93     def _build_brighcove_url(cls, object_str):
  94         """
  95         Build a Brightcove url from a xml string containing
  96         <object class="BrightcoveExperience">{params}</object>
  97         """
  98
  99         # Fix up some stupid HTML, see https://github.com/rg3/youtube-dl/issues/1553
 100         object_str = re.sub(r'(<param name="[^"]+" value="[^"]+")>',
 101                             lambda m: m.group(1) + '/>', object_str)
 102         # Fix up some stupid XML, see https://github.com/rg3/youtube-dl/issues/1608
 103         object_str = object_str.replace('<--', '<!--')
 104         object_str = fix_xml_ampersands(object_str)
 105
 106         object_doc = xml.etree.ElementTree.fromstring(object_str.encode('utf-8'))
 107
 108         fv_el = find_xpath_attr(object_doc, './param', 'name', 'flashVars')
 109         if fv_el is not None:
 110             flashvars = dict(
 111                 (k, v[0])
 112                 for k, v in compat_parse_qs(fv_el.attrib['value']).items())
 113         else:
 114             flashvars = {}
 115
 116         def find_param(name):
 117             if name in flashvars:
 118                 return flashvars[name]
 119             node = find_xpath_attr(object_doc, './param', 'name', name)
 120             if node is not None:
 121                 return node.attrib['value']
 122             return None
 123
 124         params = {}
 125
 126         playerID = find_param('playerID')
 127         if playerID is None:
 128             raise ExtractorError('Cannot find player ID')
 129         params['playerID'] = playerID
 130
 131         playerKey = find_param('playerKey')
 132         # Not all pages define this value
 133         if playerKey is not None:
 134             params['playerKey'] = playerKey
 135         # The three fields hold the id of the video
 136         videoPlayer = find_param('@videoPlayer') or find_param('videoId') or find_param('videoID')
 137         if videoPlayer is not None:
 138             params['@videoPlayer'] = videoPlayer
 139         linkBase = find_param('linkBaseURL')
 140         if linkBase is not None:
 141             params['linkBaseURL'] = linkBase
 142         data = compat_urllib_parse.urlencode(params)
 143         return cls._FEDERATED_URL_TEMPLATE % data
 144
 145     @classmethod
 146     def _extract_brightcove_url(cls, webpage):
 147         """Try to extract the brightcove url from the webpage, returns None
 148         if it can't be found
 149         """
 150         urls = cls._extract_brightcove_urls(webpage)
 151         return urls[0] if urls else None
 152
 153     @classmethod
 154     def _extract_brightcove_urls(cls, webpage):
 155         """Return a list of all Brightcove URLs from the webpage """
 156
 157         url_m = re.search(r'<meta\s+property="og:video"\s+content="(http://c.brightcove.com/[^"]+)"', webpage)
 158         if url_m:
 159             url = unescapeHTML(url_m.group(1))
 160             # Some sites don't add it, we can't download with this url, for example:
 161             # http://www.ktvu.com/videos/news/raw-video-caltrain-releases-video-of-man-almost/vCTZdY/
 162             if 'playerKey' in url:
 163                 return [url]
 164
 165         matches = re.findall(
 166             r'''(?sx)<object
 167             (?:
 168                 [^>]+?class=[\'"][^>]*?BrightcoveExperience.*?[\'"] |
 169                 [^>]*?>\s*<param\s+name="movie"\s+value="https?://[^/]*brightcove\.com/
 170             ).+?</object>''',
 171             webpage)
 172         return [cls._build_brighcove_url(m) for m in matches]
 173
 174     def _real_extract(self, url):
 175         url, smuggled_data = unsmuggle_url(url, {})
 176
 177         # Change the 'videoId' and others field to '@videoPlayer'
 178         url = re.sub(r'(?<=[?&])(videoI(d|D)|bctid)', '%40videoPlayer', url)
 179         # Change bckey (used by bcove.me urls) to playerKey
 180         url = re.sub(r'(?<=[?&])bckey', 'playerKey', url)
 181         mobj = re.match(self._VALID_URL, url)
 182         query_str = mobj.group('query')
 183         query = compat_urlparse.parse_qs(query_str)
 184
 185         videoPlayer = query.get('@videoPlayer')
 186         if videoPlayer:
 187             # We set the original url as the default 'Referer' header
 188             referer = smuggled_data.get('Referer', url)
 189             return self._get_video_info(
 190                 videoPlayer[0], query_str, query, referer=referer)
 191         elif 'playerKey' in query:
 192             player_key = query['playerKey']
 193             return self._get_playlist_info(player_key[0])
 194         else:
 195             raise ExtractorError(
 196                 'Cannot find playerKey= variable. Did you forget quotes in a shell invocation?',
 197                 expected=True)
 198
 199     def _get_video_info(self, video_id, query_str, query, referer=None):
 200         request_url = self._FEDERATED_URL_TEMPLATE % query_str
 201         req = compat_urllib_request.Request(request_url)
 202         linkBase = query.get('linkBaseURL')
 203         if linkBase is not None:
 204             referer = linkBase[0]
 205         if referer is not None:
 206             req.add_header('Referer', referer)
 207         webpage = self._download_webpage(req, video_id)
 208
 209         self.report_extraction(video_id)
 210         info = self._search_regex(r'var experienceJSON = ({.*});', webpage, 'json')
 211         info = json.loads(info)['data']
 212         video_info = info['programmedContent']['videoPlayer']['mediaDTO']
 213         video_info['_youtubedl_adServerURL'] = info.get('adServerURL')
 214
 215         return self._extract_video_info(video_info)
 216
 217     def _get_playlist_info(self, player_key):
 218         info_url = 'http://c.brightcove.com/services/json/experience/runtime/?command=get_programming_for_experience&playerKey=%s' % player_key
 219         playlist_info = self._download_webpage(
 220             info_url, player_key, 'Downloading playlist information')
 221
 222         json_data = json.loads(playlist_info)
 223         if 'videoList' not in json_data:
 224             raise ExtractorError('Empty playlist')
 225         playlist_info = json_data['videoList']
 226         videos = [self._extract_video_info(video_info) for video_info in playlist_info['mediaCollectionDTO']['videoDTOs']]
 227
 228         return self.playlist_result(videos, playlist_id=playlist_info['id'],
 229                                     playlist_title=playlist_info['mediaCollectionDTO']['displayName'])
 230
 231     def _extract_video_info(self, video_info):
 232         info = {
 233             'id': compat_str(video_info['id']),
 234             'title': video_info['displayName'].strip(),
 235             'description': video_info.get('shortDescription'),
 236             'thumbnail': video_info.get('videoStillURL') or video_info.get('thumbnailURL'),
 237             'uploader': video_info.get('publisherName'),
 238         }
 239
 240         renditions = video_info.get('renditions')
 241         if renditions:
 242             formats = []
 243             for rend in renditions:
 244                 url = rend['defaultURL']
 245                 if rend['remote']:
 246                     # This type of renditions are served through akamaihd.net,
 247                     # but they don't use f4m manifests
 248                     url = url.replace('control/', '') + '?&v=3.3.0&fp=13&r=FEEFJ&g=RTSJIMBMPFPB'
 249                     ext = 'flv'
 250                 else:
 251                     ext = determine_ext(url)
 252                 size = rend.get('size')
 253                 formats.append({
 254                     'url': url,
 255                     'ext': ext,
 256                     'height': rend.get('frameHeight'),
 257                     'width': rend.get('frameWidth'),
 258                     'filesize': size if size != 0 else None,
 259                 })
 260             self._sort_formats(formats)
 261             info['formats'] = formats
 262         elif video_info.get('FLVFullLengthURL') is not None:
 263             info.update({
 264                 'url': video_info['FLVFullLengthURL'],
 265             })
 266
 267         if self._downloader.params.get('include_ads', False):
 268             adServerURL = video_info.get('_youtubedl_adServerURL')
 269             if adServerURL:
 270                 ad_info = {
 271                     '_type': 'url',
 272                     'url': adServerURL,
 273                 }
 274                 if 'url' in info:
 275                     return {
 276                         '_type': 'playlist',
 277                         'title': info['title'],
 278                         'entries': [ad_info, info],
 279                     }
 280                 else:
 281                     return ad_info
 282
 283         if 'url' not in info and not info.get('formats'):
 284             raise ExtractorError('Unable to extract video url for %s' % info['id'])
 285         return info