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