youtube_dl/extractor/brightcove.py

   1 # encoding: utf-8
   2
   3 import re
   4 import json
   5 import xml.etree.ElementTree
   6
   7 from .common import InfoExtractor
   8 from ..utils import (
   9     compat_urllib_parse,
  10     find_xpath_attr,
  11     compat_urlparse,
  12     compat_str,
  13
  14     ExtractorError,
  15 )
  16
  17 class BrightcoveIE(InfoExtractor):
  18     _VALID_URL = r'https?://.*brightcove\.com/(services|viewer).*\?(?P<query>.*)'
  19     _FEDERATED_URL_TEMPLATE = 'http://c.brightcove.com/services/viewer/htmlFederated?%s'
  20     _PLAYLIST_URL_TEMPLATE = 'http://c.brightcove.com/services/json/experience/runtime/?command=get_programming_for_experience&playerKey=%s'
  21
  22     _TESTS = [
  23         {
  24             # From http://www.8tv.cat/8aldia/videos/xavier-sala-i-martin-aquesta-tarda-a-8-al-dia/
  25             u'url': u'http://c.brightcove.com/services/viewer/htmlFederated?playerID=1654948606001&flashID=myExperience&%40videoPlayer=2371591881001',
  26             u'file': u'2371591881001.mp4',
  27             u'md5': u'8eccab865181d29ec2958f32a6a754f5',
  28             u'note': u'Test Brightcove downloads and detection in GenericIE',
  29             u'info_dict': {
  30                 u'title': u'Xavier Sala i Martín: “Un banc que no presta és un banc zombi que no serveix per a res”',
  31                 u'uploader': u'8TV',
  32                 u'description': u'md5:a950cc4285c43e44d763d036710cd9cd',
  33             }
  34         },
  35         {
  36             # From http://medianetwork.oracle.com/video/player/1785452137001
  37             u'url': u'http://c.brightcove.com/services/viewer/htmlFederated?playerID=1217746023001&flashID=myPlayer&%40videoPlayer=1785452137001',
  38             u'file': u'1785452137001.flv',
  39             u'info_dict': {
  40                 u'title': u'JVMLS 2012: Arrays 2.0 - Opportunities and Challenges',
  41                 u'description': u'John Rose speaks at the JVM Language Summit, August 1, 2012.',
  42                 u'uploader': u'Oracle',
  43             },
  44         },
  45         {
  46             # From http://mashable.com/2013/10/26/thermoelectric-bracelet-lets-you-control-your-body-temperature/
  47             u'url': u'http://c.brightcove.com/services/viewer/federated_f9?&playerID=1265504713001&publisherID=AQ%7E%7E%2CAAABBzUwv1E%7E%2CxP-xFHVUstiMFlNYfvF4G9yFnNaqCw_9&videoID=2750934548001',
  48             u'info_dict': {
  49                 u'id': u'2750934548001',
  50                 u'ext': u'mp4',
  51                 u'title': u'This Bracelet Acts as a Personal Thermostat',
  52                 u'description': u'md5:547b78c64f4112766ccf4e151c20b6a0',
  53                 u'uploader': u'Mashable',
  54             },
  55         },
  56     ]
  57
  58     @classmethod
  59     def _build_brighcove_url(cls, object_str):
  60         """
  61         Build a Brightcove url from a xml string containing
  62         <object class="BrightcoveExperience">{params}</object>
  63         """
  64
  65         # Fix up some stupid HTML, see https://github.com/rg3/youtube-dl/issues/1553
  66         object_str = re.sub(r'(<param name="[^"]+" value="[^"]+")>',
  67                             lambda m: m.group(1) + '/>', object_str)
  68         # Fix up some stupid XML, see https://github.com/rg3/youtube-dl/issues/1608
  69         object_str = object_str.replace(u'<--', u'<!--')
  70
  71         object_doc = xml.etree.ElementTree.fromstring(object_str)
  72         assert u'BrightcoveExperience' in object_doc.attrib['class']
  73         params = {'flashID': object_doc.attrib['id'],
  74                   'playerID': find_xpath_attr(object_doc, './param', 'name', 'playerID').attrib['value'],
  75                   }
  76         playerKey = find_xpath_attr(object_doc, './param', 'name', 'playerKey')
  77         # Not all pages define this value
  78         if playerKey is not None:
  79             params['playerKey'] = playerKey.attrib['value']
  80         videoPlayer = find_xpath_attr(object_doc, './param', 'name', '@videoPlayer')
  81         if videoPlayer is not None:
  82             params['@videoPlayer'] = videoPlayer.attrib['value']
  83         data = compat_urllib_parse.urlencode(params)
  84         return cls._FEDERATED_URL_TEMPLATE % data
  85
  86     @classmethod
  87     def _extract_brightcove_url(cls, webpage):
  88         """Try to extract the brightcove url from the wepbage, returns None
  89         if it can't be found
  90         """
  91         m_brightcove = re.search(
  92             r'<object[^>]+?class=([\'"])[^>]*?BrightcoveExperience.*?\1.+?</object>',
  93             webpage, re.DOTALL)
  94         if m_brightcove is not None:
  95             return cls._build_brighcove_url(m_brightcove.group())
  96         else:
  97             return None
  98
  99     def _real_extract(self, url):
 100         # Change the 'videoId' or 'videoID' field to '@videoPlayer'
 101         url = re.sub(r'(?<=[?&])videoI(d|D)', '%40videoPlayer', url)
 102         mobj = re.match(self._VALID_URL, url)
 103         query_str = mobj.group('query')
 104         query = compat_urlparse.parse_qs(query_str)
 105
 106         videoPlayer = query.get('@videoPlayer')
 107         if videoPlayer:
 108             return self._get_video_info(videoPlayer[0], query_str)
 109         else:
 110             player_key = query['playerKey']
 111             return self._get_playlist_info(player_key[0])
 112
 113     def _get_video_info(self, video_id, query):
 114         request_url = self._FEDERATED_URL_TEMPLATE % query
 115         webpage = self._download_webpage(request_url, video_id)
 116
 117         self.report_extraction(video_id)
 118         info = self._search_regex(r'var experienceJSON = ({.*?});', webpage, 'json')
 119         info = json.loads(info)['data']
 120         video_info = info['programmedContent']['videoPlayer']['mediaDTO']
 121
 122         return self._extract_video_info(video_info)
 123
 124     def _get_playlist_info(self, player_key):
 125         playlist_info = self._download_webpage(self._PLAYLIST_URL_TEMPLATE % player_key,
 126                                                player_key, u'Downloading playlist information')
 127
 128         json_data = json.loads(playlist_info)
 129         if 'videoList' not in json_data:
 130             raise ExtractorError(u'Empty playlist')
 131         playlist_info = json_data['videoList']
 132         videos = [self._extract_video_info(video_info) for video_info in playlist_info['mediaCollectionDTO']['videoDTOs']]
 133
 134         return self.playlist_result(videos, playlist_id=playlist_info['id'],
 135                                     playlist_title=playlist_info['mediaCollectionDTO']['displayName'])
 136
 137     def _extract_video_info(self, video_info):
 138         info = {
 139             'id': compat_str(video_info['id']),
 140             'title': video_info['displayName'],
 141             'description': video_info.get('shortDescription'),
 142             'thumbnail': video_info.get('videoStillURL') or video_info.get('thumbnailURL'),
 143             'uploader': video_info.get('publisherName'),
 144         }
 145
 146         renditions = video_info.get('renditions')
 147         if renditions:
 148             renditions = sorted(renditions, key=lambda r: r['size'])
 149             best_format = renditions[-1]
 150             info.update({
 151                 'url': best_format['defaultURL'],
 152             })
 153         elif video_info.get('FLVFullLengthURL') is not None:
 154             info.update({
 155                 'url': video_info['FLVFullLengthURL'],
 156             })
 157         else:
 158             raise ExtractorError(u'Unable to extract video url for %s' % info['id'])
 159         return info