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