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 import xml.etree.ElementTree
   7
   8 from .common import InfoExtractor
   9 from ..utils import (
  10     compat_urllib_parse,
  11     find_xpath_attr,
  12     compat_urlparse,
  13     compat_str,
  14     compat_urllib_request,
  15
  16     ExtractorError,
  17     unsmuggle_url,
  18 )
  19
  20
  21 class BrightcoveIE(InfoExtractor):
  22     _VALID_URL = r'https?://.*brightcove\.com/(services|viewer).*\?(?P<query>.*)'
  23     _FEDERATED_URL_TEMPLATE = 'http://c.brightcove.com/services/viewer/htmlFederated?%s'
  24     _PLAYLIST_URL_TEMPLATE = 'http://c.brightcove.com/services/json/experience/runtime/?command=get_programming_for_experience&playerKey=%s'
  25
  26     _TESTS = [
  27         {
  28             # From http://www.8tv.cat/8aldia/videos/xavier-sala-i-martin-aquesta-tarda-a-8-al-dia/
  29             'url': 'http://c.brightcove.com/services/viewer/htmlFederated?playerID=1654948606001&flashID=myExperience&%40videoPlayer=2371591881001',
  30             'file': '2371591881001.mp4',
  31             'md5': '5423e113865d26e40624dce2e4b45d95',
  32             'note': 'Test Brightcove downloads and detection in GenericIE',
  33             'info_dict': {
  34                 'title': 'Xavier Sala i Martín: “Un banc que no presta és un banc zombi que no serveix per a res”',
  35                 'uploader': '8TV',
  36                 'description': 'md5:a950cc4285c43e44d763d036710cd9cd',
  37             }
  38         },
  39         {
  40             # From http://medianetwork.oracle.com/video/player/1785452137001
  41             'url': 'http://c.brightcove.com/services/viewer/htmlFederated?playerID=1217746023001&flashID=myPlayer&%40videoPlayer=1785452137001',
  42             'file': '1785452137001.flv',
  43             'info_dict': {
  44                 'title': 'JVMLS 2012: Arrays 2.0 - Opportunities and Challenges',
  45                 'description': 'John Rose speaks at the JVM Language Summit, August 1, 2012.',
  46                 'uploader': 'Oracle',
  47             },
  48         },
  49         {
  50             # From http://mashable.com/2013/10/26/thermoelectric-bracelet-lets-you-control-your-body-temperature/
  51             'url': 'http://c.brightcove.com/services/viewer/federated_f9?&playerID=1265504713001&publisherID=AQ%7E%7E%2CAAABBzUwv1E%7E%2CxP-xFHVUstiMFlNYfvF4G9yFnNaqCw_9&videoID=2750934548001',
  52             'info_dict': {
  53                 'id': '2750934548001',
  54                 'ext': 'mp4',
  55                 'title': 'This Bracelet Acts as a Personal Thermostat',
  56                 'description': 'md5:547b78c64f4112766ccf4e151c20b6a0',
  57                 'uploader': 'Mashable',
  58             },
  59         },
  60         {
  61             # test that the default referer works
  62             # from http://national.ballet.ca/interact/video/Lost_in_Motion_II/
  63             'url': 'http://link.brightcove.com/services/player/bcpid756015033001?bckey=AQ~~,AAAApYJi_Ck~,GxhXCegT1Dp39ilhXuxMJxasUhVNZiil&bctid=2878862109001',
  64             'info_dict': {
  65                 'id': '2878862109001',
  66                 'ext': 'mp4',
  67                 'title': 'Lost in Motion II',
  68                 'description': 'md5:363109c02998fee92ec02211bd8000df',
  69                 'uploader': 'National Ballet of Canada',
  70             },
  71         },
  72     ]
  73
  74     @classmethod
  75     def _build_brighcove_url(cls, object_str):
  76         """
  77         Build a Brightcove url from a xml string containing
  78         <object class="BrightcoveExperience">{params}</object>
  79         """
  80
  81         # Fix up some stupid HTML, see https://github.com/rg3/youtube-dl/issues/1553
  82         object_str = re.sub(r'(<param name="[^"]+" value="[^"]+")>',
  83                             lambda m: m.group(1) + '/>', object_str)
  84         # Fix up some stupid XML, see https://github.com/rg3/youtube-dl/issues/1608
  85         object_str = object_str.replace('<--', '<!--')
  86
  87         object_doc = xml.etree.ElementTree.fromstring(object_str)
  88         assert 'BrightcoveExperience' in object_doc.attrib['class']
  89         params = {
  90             'playerID': find_xpath_attr(object_doc, './param', 'name', 'playerID').attrib['value'],
  91         }
  92         def find_param(name):
  93             node = find_xpath_attr(object_doc, './param', 'name', name)
  94             if node is not None:
  95                 return node.attrib['value']
  96             return None
  97         playerKey = find_param('playerKey')
  98         # Not all pages define this value
  99         if playerKey is not None:
 100             params['playerKey'] = playerKey
 101         # The three fields hold the id of the video
 102         videoPlayer = find_param('@videoPlayer') or find_param('videoId') or find_param('videoID')
 103         if videoPlayer is not None:
 104             params['@videoPlayer'] = videoPlayer
 105         linkBase = find_param('linkBaseURL')
 106         if linkBase is not None:
 107             params['linkBaseURL'] = linkBase
 108         data = compat_urllib_parse.urlencode(params)
 109         return cls._FEDERATED_URL_TEMPLATE % data
 110
 111     @classmethod
 112     def _extract_brightcove_url(cls, webpage):
 113         """Try to extract the brightcove url from the wepbage, returns None
 114         if it can't be found
 115         """
 116         m_brightcove = re.search(
 117             r'<object[^>]+?class=([\'"])[^>]*?BrightcoveExperience.*?\1.+?</object>',
 118             webpage, re.DOTALL)
 119         if m_brightcove is not None:
 120             return cls._build_brighcove_url(m_brightcove.group())
 121         else:
 122             return None
 123
 124     def _real_extract(self, url):
 125         url, smuggled_data = unsmuggle_url(url, {})
 126
 127         # Change the 'videoId' and others field to '@videoPlayer'
 128         url = re.sub(r'(?<=[?&])(videoI(d|D)|bctid)', '%40videoPlayer', url)
 129         # Change bckey (used by bcove.me urls) to playerKey
 130         url = re.sub(r'(?<=[?&])bckey', 'playerKey', url)
 131         mobj = re.match(self._VALID_URL, url)
 132         query_str = mobj.group('query')
 133         query = compat_urlparse.parse_qs(query_str)
 134
 135         videoPlayer = query.get('@videoPlayer')
 136         if videoPlayer:
 137             # We set the original url as the default 'Referer' header
 138             referer = smuggled_data.get('Referer', url)
 139             return self._get_video_info(
 140                 videoPlayer[0], query_str, query, referer=referer)
 141         else:
 142             player_key = query['playerKey']
 143             return self._get_playlist_info(player_key[0])
 144
 145     def _get_video_info(self, video_id, query_str, query, referer=None):
 146         request_url = self._FEDERATED_URL_TEMPLATE % query_str
 147         req = compat_urllib_request.Request(request_url)
 148         linkBase = query.get('linkBaseURL')
 149         if linkBase is not None:
 150             referer = linkBase[0]
 151         if referer is not None:
 152             req.add_header('Referer', referer)
 153         webpage = self._download_webpage(req, video_id)
 154
 155         self.report_extraction(video_id)
 156         info = self._search_regex(r'var experienceJSON = ({.*?});', webpage, 'json')
 157         info = json.loads(info)['data']
 158         video_info = info['programmedContent']['videoPlayer']['mediaDTO']
 159
 160         return self._extract_video_info(video_info)
 161
 162     def _get_playlist_info(self, player_key):
 163         playlist_info = self._download_webpage(self._PLAYLIST_URL_TEMPLATE % player_key,
 164                                                player_key, 'Downloading playlist information')
 165
 166         json_data = json.loads(playlist_info)
 167         if 'videoList' not in json_data:
 168             raise ExtractorError('Empty playlist')
 169         playlist_info = json_data['videoList']
 170         videos = [self._extract_video_info(video_info) for video_info in playlist_info['mediaCollectionDTO']['videoDTOs']]
 171
 172         return self.playlist_result(videos, playlist_id=playlist_info['id'],
 173                                     playlist_title=playlist_info['mediaCollectionDTO']['displayName'])
 174
 175     def _extract_video_info(self, video_info):
 176         info = {
 177             'id': compat_str(video_info['id']),
 178             'title': video_info['displayName'],
 179             'description': video_info.get('shortDescription'),
 180             'thumbnail': video_info.get('videoStillURL') or video_info.get('thumbnailURL'),
 181             'uploader': video_info.get('publisherName'),
 182         }
 183
 184         renditions = video_info.get('renditions')
 185         if renditions:
 186             renditions = sorted(renditions, key=lambda r: r['size'])
 187             info['formats'] = [{
 188                 'url': rend['defaultURL'],
 189                 'height': rend.get('frameHeight'),
 190                 'width': rend.get('frameWidth'),
 191             } for rend in renditions]
 192         elif video_info.get('FLVFullLengthURL') is not None:
 193             info.update({
 194                 'url': video_info['FLVFullLengthURL'],
 195             })
 196         else:
 197             raise ExtractorError('Unable to extract video url for %s' % info['id'])
 198         return info