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