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     fix_xml_ampersands,
  13     compat_urlparse,
  14     compat_str,
  15     compat_urllib_request,
  16     compat_parse_qs,
  17
  18     ExtractorError,
  19     unsmuggle_url,
  20     unescapeHTML,
  21 )
  22
  23
  24 class BrightcoveIE(InfoExtractor):
  25     _VALID_URL = r'https?://.*brightcove\.com/(services|viewer).*\?(?P<query>.*)'
  26     _FEDERATED_URL_TEMPLATE = 'http://c.brightcove.com/services/viewer/htmlFederated?%s'
  27
  28     _TESTS = [
  29         {
  30             # From http://www.8tv.cat/8aldia/videos/xavier-sala-i-martin-aquesta-tarda-a-8-al-dia/
  31             'url': 'http://c.brightcove.com/services/viewer/htmlFederated?playerID=1654948606001&flashID=myExperience&%40videoPlayer=2371591881001',
  32             'file': '2371591881001.mp4',
  33             'md5': '5423e113865d26e40624dce2e4b45d95',
  34             'note': 'Test Brightcove downloads and detection in GenericIE',
  35             'info_dict': {
  36                 'title': 'Xavier Sala i Martín: “Un banc que no presta és un banc zombi que no serveix per a res”',
  37                 'uploader': '8TV',
  38                 'description': 'md5:a950cc4285c43e44d763d036710cd9cd',
  39             }
  40         },
  41         {
  42             # From http://medianetwork.oracle.com/video/player/1785452137001
  43             'url': 'http://c.brightcove.com/services/viewer/htmlFederated?playerID=1217746023001&flashID=myPlayer&%40videoPlayer=1785452137001',
  44             'file': '1785452137001.flv',
  45             'info_dict': {
  46                 'title': 'JVMLS 2012: Arrays 2.0 - Opportunities and Challenges',
  47                 'description': 'John Rose speaks at the JVM Language Summit, August 1, 2012.',
  48                 'uploader': 'Oracle',
  49             },
  50         },
  51         {
  52             # From http://mashable.com/2013/10/26/thermoelectric-bracelet-lets-you-control-your-body-temperature/
  53             'url': 'http://c.brightcove.com/services/viewer/federated_f9?&playerID=1265504713001&publisherID=AQ%7E%7E%2CAAABBzUwv1E%7E%2CxP-xFHVUstiMFlNYfvF4G9yFnNaqCw_9&videoID=2750934548001',
  54             'info_dict': {
  55                 'id': '2750934548001',
  56                 'ext': 'mp4',
  57                 'title': 'This Bracelet Acts as a Personal Thermostat',
  58                 'description': 'md5:547b78c64f4112766ccf4e151c20b6a0',
  59                 'uploader': 'Mashable',
  60             },
  61         },
  62         {
  63             # test that the default referer works
  64             # from http://national.ballet.ca/interact/video/Lost_in_Motion_II/
  65             'url': 'http://link.brightcove.com/services/player/bcpid756015033001?bckey=AQ~~,AAAApYJi_Ck~,GxhXCegT1Dp39ilhXuxMJxasUhVNZiil&bctid=2878862109001',
  66             'info_dict': {
  67                 'id': '2878862109001',
  68                 'ext': 'mp4',
  69                 'title': 'Lost in Motion II',
  70                 'description': 'md5:363109c02998fee92ec02211bd8000df',
  71                 'uploader': 'National Ballet of Canada',
  72             },
  73         }
  74     ]
  75
  76     @classmethod
  77     def _build_brighcove_url(cls, object_str):
  78         """
  79         Build a Brightcove url from a xml string containing
  80         <object class="BrightcoveExperience">{params}</object>
  81         """
  82
  83         # Fix up some stupid HTML, see https://github.com/rg3/youtube-dl/issues/1553
  84         object_str = re.sub(r'(<param name="[^"]+" value="[^"]+")>',
  85                             lambda m: m.group(1) + '/>', object_str)
  86         # Fix up some stupid XML, see https://github.com/rg3/youtube-dl/issues/1608
  87         object_str = object_str.replace('<--', '<!--')
  88         object_str = fix_xml_ampersands(object_str)
  89
  90         object_doc = xml.etree.ElementTree.fromstring(object_str.encode('utf-8'))
  91
  92         fv_el = find_xpath_attr(object_doc, './param', 'name', 'flashVars')
  93         if fv_el is not None:
  94             flashvars = dict(
  95                 (k, v[0])
  96                 for k, v in compat_parse_qs(fv_el.attrib['value']).items())
  97         else:
  98             flashvars = {}
  99
 100         def find_param(name):
 101             if name in flashvars:
 102                 return flashvars[name]
 103             node = find_xpath_attr(object_doc, './param', 'name', name)
 104             if node is not None:
 105                 return node.attrib['value']
 106             return None
 107
 108         params = {}
 109
 110         playerID = find_param('playerID')
 111         if playerID is None:
 112             raise ExtractorError('Cannot find player ID')
 113         params['playerID'] = playerID
 114
 115         playerKey = find_param('playerKey')
 116         # Not all pages define this value
 117         if playerKey is not None:
 118             params['playerKey'] = playerKey
 119         # The three fields hold the id of the video
 120         videoPlayer = find_param('@videoPlayer') or find_param('videoId') or find_param('videoID')
 121         if videoPlayer is not None:
 122             params['@videoPlayer'] = videoPlayer
 123         linkBase = find_param('linkBaseURL')
 124         if linkBase is not None:
 125             params['linkBaseURL'] = linkBase
 126         data = compat_urllib_parse.urlencode(params)
 127         return cls._FEDERATED_URL_TEMPLATE % data
 128
 129     @classmethod
 130     def _extract_brightcove_url(cls, webpage):
 131         """Try to extract the brightcove url from the webpage, returns None
 132         if it can't be found
 133         """
 134         urls = cls._extract_brightcove_urls(webpage)
 135         return urls[0] if urls else None
 136
 137     @classmethod
 138     def _extract_brightcove_urls(cls, webpage):
 139         """Return a list of all Brightcove URLs from the webpage """
 140
 141         url_m = re.search(r'<meta\s+property="og:video"\s+content="(http://c.brightcove.com/[^"]+)"', webpage)
 142         if url_m:
 143             url = unescapeHTML(url_m.group(1))
 144             # Some sites don't add it, we can't download with this url, for example:
 145             # http://www.ktvu.com/videos/news/raw-video-caltrain-releases-video-of-man-almost/vCTZdY/
 146             if 'playerKey' in url:
 147                 return [url]
 148
 149         matches = re.findall(
 150             r'''(?sx)<object
 151             (?:
 152                 [^>]+?class=[\'"][^>]*?BrightcoveExperience.*?[\'"] |
 153                 [^>]*?>\s*<param\s+name="movie"\s+value="https?://[^/]*brightcove\.com/
 154             ).+?</object>''',
 155             webpage)
 156         return [cls._build_brighcove_url(m) for m in matches]
 157
 158     def _real_extract(self, url):
 159         url, smuggled_data = unsmuggle_url(url, {})
 160
 161         # Change the 'videoId' and others field to '@videoPlayer'
 162         url = re.sub(r'(?<=[?&])(videoI(d|D)|bctid)', '%40videoPlayer', url)
 163         # Change bckey (used by bcove.me urls) to playerKey
 164         url = re.sub(r'(?<=[?&])bckey', 'playerKey', url)
 165         mobj = re.match(self._VALID_URL, url)
 166         query_str = mobj.group('query')
 167         query = compat_urlparse.parse_qs(query_str)
 168
 169         videoPlayer = query.get('@videoPlayer')
 170         if videoPlayer:
 171             # We set the original url as the default 'Referer' header
 172             referer = smuggled_data.get('Referer', url)
 173             return self._get_video_info(
 174                 videoPlayer[0], query_str, query, referer=referer)
 175         else:
 176             player_key = query['playerKey']
 177             return self._get_playlist_info(player_key[0])
 178
 179     def _get_video_info(self, video_id, query_str, query, referer=None):
 180         request_url = self._FEDERATED_URL_TEMPLATE % query_str
 181         req = compat_urllib_request.Request(request_url)
 182         linkBase = query.get('linkBaseURL')
 183         if linkBase is not None:
 184             referer = linkBase[0]
 185         if referer is not None:
 186             req.add_header('Referer', referer)
 187         webpage = self._download_webpage(req, video_id)
 188
 189         self.report_extraction(video_id)
 190         info = self._search_regex(r'var experienceJSON = ({.*?});', webpage, 'json')
 191         info = json.loads(info)['data']
 192         video_info = info['programmedContent']['videoPlayer']['mediaDTO']
 193         video_info['_youtubedl_adServerURL'] = info.get('adServerURL')
 194
 195         return self._extract_video_info(video_info)
 196
 197     def _get_playlist_info(self, player_key):
 198         info_url = 'http://c.brightcove.com/services/json/experience/runtime/?command=get_programming_for_experience&playerKey=%s' % player_key
 199         playlist_info = self._download_webpage(
 200             info_url, player_key, 'Downloading playlist information')
 201
 202         json_data = json.loads(playlist_info)
 203         if 'videoList' not in json_data:
 204             raise ExtractorError('Empty playlist')
 205         playlist_info = json_data['videoList']
 206         videos = [self._extract_video_info(video_info) for video_info in playlist_info['mediaCollectionDTO']['videoDTOs']]
 207
 208         return self.playlist_result(videos, playlist_id=playlist_info['id'],
 209                                     playlist_title=playlist_info['mediaCollectionDTO']['displayName'])
 210
 211     def _extract_video_info(self, video_info):
 212         info = {
 213             'id': compat_str(video_info['id']),
 214             'title': video_info['displayName'].strip(),
 215             'description': video_info.get('shortDescription'),
 216             'thumbnail': video_info.get('videoStillURL') or video_info.get('thumbnailURL'),
 217             'uploader': video_info.get('publisherName'),
 218         }
 219
 220         renditions = video_info.get('renditions')
 221         if renditions:
 222             renditions = sorted(renditions, key=lambda r: r['size'])
 223             info['formats'] = [{
 224                 'url': rend['defaultURL'],
 225                 'height': rend.get('frameHeight'),
 226                 'width': rend.get('frameWidth'),
 227             } for rend in renditions]
 228         elif video_info.get('FLVFullLengthURL') is not None:
 229             info.update({
 230                 'url': video_info['FLVFullLengthURL'],
 231             })
 232
 233         if self._downloader.params.get('include_ads', False):
 234             adServerURL = video_info.get('_youtubedl_adServerURL')
 235             if adServerURL:
 236                 ad_info = {
 237                     '_type': 'url',
 238                     'url': adServerURL,
 239                 }
 240                 if 'url' in info:
 241                     return {
 242                         '_type': 'playlist',
 243                         'title': info['title'],
 244                         'entries': [ad_info, info],
 245                     }
 246                 else:
 247                     return ad_info
 248
 249         if 'url' not in info and not info.get('formats'):
 250             raise ExtractorError('Unable to extract video url for %s' % info['id'])
 251         return info