_ 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     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             # https://github.com/rg3/youtube-dl/issues/2253
  75             'url': 'http://v.thestar.com/services/player/bcpid2071349530001?bckey=AQ~~,AAAAuO4KaJE~,gatFNwSKdGDmDpIYqNJ-fTHn_c4z_LH_&bctid=3101154703001',
  76             'file': '3101154703001.mp4',
  77             'md5': '0ba9446db037002366bab3b3eb30c88c',
  78             'info_dict': {
  79                 'title': 'Still no power',
  80                 'uploader': 'thestar.com',
  81                 'description': 'Mississauga resident David Farmer is still out of power as a result of the ice storm a month ago. To keep the house warm, Farmer cuts wood from his property for a wood burning stove downstairs.',
  82             }
  83         }
  84     ]
  85
  86     @classmethod
  87     def _build_brighcove_url(cls, object_str):
  88         """
  89         Build a Brightcove url from a xml string containing
  90         <object class="BrightcoveExperience">{params}</object>
  91         """
  92
  93         # Fix up some stupid HTML, see https://github.com/rg3/youtube-dl/issues/1553
  94         object_str = re.sub(r'(<param name="[^"]+" value="[^"]+")>',
  95                             lambda m: m.group(1) + '/>', object_str)
  96         # Fix up some stupid XML, see https://github.com/rg3/youtube-dl/issues/1608
  97         object_str = object_str.replace('<--', '<!--')
  98         object_str = fix_xml_ampersands(object_str)
  99
 100         object_doc = xml.etree.ElementTree.fromstring(object_str)
 101
 102         fv_el = find_xpath_attr(object_doc, './param', 'name', 'flashVars')
 103         if fv_el is not None:
 104             flashvars = dict(
 105                 (k, v[0])
 106                 for k, v in compat_parse_qs(fv_el.attrib['value']).items())
 107         else:
 108             flashvars = {}
 109
 110         def find_param(name):
 111             if name in flashvars:
 112                 return flashvars[name]
 113             node = find_xpath_attr(object_doc, './param', 'name', name)
 114             if node is not None:
 115                 return node.attrib['value']
 116             return None
 117
 118         params = {}
 119
 120         playerID = find_param('playerID')
 121         if playerID is None:
 122             raise ExtractorError('Cannot find player ID')
 123         params['playerID'] = playerID
 124
 125         playerKey = find_param('playerKey')
 126         # Not all pages define this value
 127         if playerKey is not None:
 128             params['playerKey'] = playerKey
 129         # The three fields hold the id of the video
 130         videoPlayer = find_param('@videoPlayer') or find_param('videoId') or find_param('videoID')
 131         if videoPlayer is not None:
 132             params['@videoPlayer'] = videoPlayer
 133         linkBase = find_param('linkBaseURL')
 134         if linkBase is not None:
 135             params['linkBaseURL'] = linkBase
 136         data = compat_urllib_parse.urlencode(params)
 137         return cls._FEDERATED_URL_TEMPLATE % data
 138
 139     @classmethod
 140     def _extract_brightcove_url(cls, webpage):
 141         """Try to extract the brightcove url from the wepbage, returns None
 142         if it can't be found
 143         """
 144
 145         url_m = re.search(r'<meta\s+property="og:video"\s+content="(http://c.brightcove.com/[^"]+)"', webpage)
 146         if url_m:
 147             return url_m.group(1)
 148
 149         m_brightcove = re.search(
 150             r'''(?sx)<object
 151             (?:
 152                 [^>]+?class=([\'"])[^>]*?BrightcoveExperience.*?\1 |
 153                 [^>]*?>\s*<param\s+name="movie"\s+value="https?://[^/]*brightcove\.com/
 154             ).+?</object>''',
 155             webpage)
 156         if m_brightcove is not None:
 157             return cls._build_brighcove_url(m_brightcove.group())
 158         else:
 159             return None
 160
 161     def _real_extract(self, url):
 162         url, smuggled_data = unsmuggle_url(url, {})
 163
 164         # Change the 'videoId' and others field to '@videoPlayer'
 165         url = re.sub(r'(?<=[?&])(videoI(d|D)|bctid)', '%40videoPlayer', url)
 166         # Change bckey (used by bcove.me urls) to playerKey
 167         url = re.sub(r'(?<=[?&])bckey', 'playerKey', url)
 168         mobj = re.match(self._VALID_URL, url)
 169         query_str = mobj.group('query')
 170         query = compat_urlparse.parse_qs(query_str)
 171
 172         videoPlayer = query.get('@videoPlayer')
 173         if videoPlayer:
 174             # We set the original url as the default 'Referer' header
 175             referer = smuggled_data.get('Referer', url)
 176             return self._get_video_info(
 177                 videoPlayer[0], query_str, query, referer=referer)
 178         else:
 179             player_key = query['playerKey']
 180             return self._get_playlist_info(player_key[0])
 181
 182     def _get_video_info(self, video_id, query_str, query, referer=None):
 183         request_url = self._FEDERATED_URL_TEMPLATE % query_str
 184         req = compat_urllib_request.Request(request_url)
 185         linkBase = query.get('linkBaseURL')
 186         if linkBase is not None:
 187             referer = linkBase[0]
 188         if referer is not None:
 189             req.add_header('Referer', referer)
 190         webpage = self._download_webpage(req, video_id)
 191
 192         self.report_extraction(video_id)
 193         info = self._search_regex(r'var experienceJSON = ({.*?});', webpage, 'json')
 194         info = json.loads(info)['data']
 195         video_info = info['programmedContent']['videoPlayer']['mediaDTO']
 196         video_info['_youtubedl_adServerURL'] = info.get('adServerURL')
 197
 198         return self._extract_video_info(video_info)
 199
 200     def _get_playlist_info(self, player_key):
 201         info_url = 'http://c.brightcove.com/services/json/experience/runtime/?command=get_programming_for_experience&playerKey=%s' % player_key
 202         playlist_info = self._download_webpage(
 203             info_url, player_key, 'Downloading playlist information')
 204
 205         json_data = json.loads(playlist_info)
 206         if 'videoList' not in json_data:
 207             raise ExtractorError('Empty playlist')
 208         playlist_info = json_data['videoList']
 209         videos = [self._extract_video_info(video_info) for video_info in playlist_info['mediaCollectionDTO']['videoDTOs']]
 210
 211         return self.playlist_result(videos, playlist_id=playlist_info['id'],
 212                                     playlist_title=playlist_info['mediaCollectionDTO']['displayName'])
 213
 214     def _extract_video_info(self, video_info):
 215         info = {
 216             'id': compat_str(video_info['id']),
 217             'title': video_info['displayName'].strip(),
 218             'description': video_info.get('shortDescription'),
 219             'thumbnail': video_info.get('videoStillURL') or video_info.get('thumbnailURL'),
 220             'uploader': video_info.get('publisherName'),
 221         }
 222
 223         renditions = video_info.get('renditions')
 224         if renditions:
 225             renditions = sorted(renditions, key=lambda r: r['size'])
 226             info['formats'] = [{
 227                 'url': rend['defaultURL'],
 228                 'height': rend.get('frameHeight'),
 229                 'width': rend.get('frameWidth'),
 230             } for rend in renditions]
 231         elif video_info.get('FLVFullLengthURL') is not None:
 232             info.update({
 233                 'url': video_info['FLVFullLengthURL'],
 234             })
 235
 236         if self._downloader.params.get('include_ads', False):
 237             adServerURL = video_info.get('_youtubedl_adServerURL')
 238             if adServerURL:
 239                 ad_info = {
 240                     '_type': 'url',
 241                     'url': adServerURL,
 242                 }
 243                 if 'url' in info:
 244                     return {
 245                         '_type': 'playlist',
 246                         'title': info['title'],
 247                         'entries': [ad_info, info],
 248                     }
 249                 else:
 250                     return ad_info
 251
 252         if 'url' not in info and not info.get('formats'):
 253             raise ExtractorError('Unable to extract video url for %s' % info['id'])
 254         return info