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     compat_urllib_parse_urlparse,
  18
  19     determine_ext,
  20     ExtractorError,
  21     unsmuggle_url,
  22     unescapeHTML,
  23 )
  24
  25
  26 class BrightcoveIE(InfoExtractor):
  27     _VALID_URL = r'https?://.*brightcove\.com/(services|viewer).*?\?(?P<query>.*)'
  28     _FEDERATED_URL_TEMPLATE = 'http://c.brightcove.com/services/viewer/htmlFederated?%s'
  29
  30     _TESTS = [
  31         {
  32             # From http://www.8tv.cat/8aldia/videos/xavier-sala-i-martin-aquesta-tarda-a-8-al-dia/
  33             'url': 'http://c.brightcove.com/services/viewer/htmlFederated?playerID=1654948606001&flashID=myExperience&%40videoPlayer=2371591881001',
  34             'md5': '5423e113865d26e40624dce2e4b45d95',
  35             'note': 'Test Brightcove downloads and detection in GenericIE',
  36             'info_dict': {
  37                 'id': '2371591881001',
  38                 'ext': 'mp4',
  39                 'title': 'Xavier Sala i Martín: “Un banc que no presta és un banc zombi que no serveix per a res”',
  40                 'uploader': '8TV',
  41                 'description': 'md5:a950cc4285c43e44d763d036710cd9cd',
  42             }
  43         },
  44         {
  45             # From http://medianetwork.oracle.com/video/player/1785452137001
  46             'url': 'http://c.brightcove.com/services/viewer/htmlFederated?playerID=1217746023001&flashID=myPlayer&%40videoPlayer=1785452137001',
  47             'info_dict': {
  48                 'id': '1785452137001',
  49                 'ext': 'flv',
  50                 'title': 'JVMLS 2012: Arrays 2.0 - Opportunities and Challenges',
  51                 'description': 'John Rose speaks at the JVM Language Summit, August 1, 2012.',
  52                 'uploader': 'Oracle',
  53             },
  54         },
  55         {
  56             # From http://mashable.com/2013/10/26/thermoelectric-bracelet-lets-you-control-your-body-temperature/
  57             'url': 'http://c.brightcove.com/services/viewer/federated_f9?&playerID=1265504713001&publisherID=AQ%7E%7E%2CAAABBzUwv1E%7E%2CxP-xFHVUstiMFlNYfvF4G9yFnNaqCw_9&videoID=2750934548001',
  58             'info_dict': {
  59                 'id': '2750934548001',
  60                 'ext': 'mp4',
  61                 'title': 'This Bracelet Acts as a Personal Thermostat',
  62                 'description': 'md5:547b78c64f4112766ccf4e151c20b6a0',
  63                 'uploader': 'Mashable',
  64             },
  65         },
  66         {
  67             # test that the default referer works
  68             # from http://national.ballet.ca/interact/video/Lost_in_Motion_II/
  69             'url': 'http://link.brightcove.com/services/player/bcpid756015033001?bckey=AQ~~,AAAApYJi_Ck~,GxhXCegT1Dp39ilhXuxMJxasUhVNZiil&bctid=2878862109001',
  70             'info_dict': {
  71                 'id': '2878862109001',
  72                 'ext': 'mp4',
  73                 'title': 'Lost in Motion II',
  74                 'description': 'md5:363109c02998fee92ec02211bd8000df',
  75                 'uploader': 'National Ballet of Canada',
  76             },
  77         },
  78         {
  79             # test flv videos served by akamaihd.net
  80             # From http://www.redbull.com/en/bike/stories/1331655643987/replay-uci-dh-world-cup-2014-from-fort-william
  81             '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',
  82             # The md5 checksum changes on each download
  83             'info_dict': {
  84                 'id': '2996102916001',
  85                 'ext': 'flv',
  86                 'title': 'UCI MTB World Cup 2014: Fort William, UK - Downhill Finals',
  87                 'uploader': 'Red Bull TV',
  88                 'description': 'UCI MTB World Cup 2014: Fort William, UK - Downhill Finals',
  89             },
  90         },
  91         {
  92             # playlist test
  93             # from http://support.brightcove.com/en/video-cloud/docs/playlist-support-single-video-players
  94             'url': 'http://c.brightcove.com/services/viewer/htmlFederated?playerID=3550052898001&playerKey=AQ%7E%7E%2CAAABmA9XpXk%7E%2C-Kp7jNgisre1fG5OdqpAFUTcs0lP_ZoL',
  95             'info_dict': {
  96                 'title': 'Sealife',
  97             },
  98             'playlist_mincount': 7,
  99         },
 100     ]
 101
 102     @classmethod
 103     def _build_brighcove_url(cls, object_str):
 104         """
 105         Build a Brightcove url from a xml string containing
 106         <object class="BrightcoveExperience">{params}</object>
 107         """
 108
 109         # Fix up some stupid HTML, see https://github.com/rg3/youtube-dl/issues/1553
 110         object_str = re.sub(r'(<param name="[^"]+" value="[^"]+")>',
 111                             lambda m: m.group(1) + '/>', object_str)
 112         # Fix up some stupid XML, see https://github.com/rg3/youtube-dl/issues/1608
 113         object_str = object_str.replace('<--', '<!--')
 114         # remove namespace to simplify extraction
 115         object_str = re.sub(r'(<object[^>]*)(xmlns=".*?")', r'\1', object_str)
 116         object_str = fix_xml_ampersands(object_str)
 117
 118         object_doc = xml.etree.ElementTree.fromstring(object_str.encode('utf-8'))
 119
 120         fv_el = find_xpath_attr(object_doc, './param', 'name', 'flashVars')
 121         if fv_el is not None:
 122             flashvars = dict(
 123                 (k, v[0])
 124                 for k, v in compat_parse_qs(fv_el.attrib['value']).items())
 125         else:
 126             flashvars = {}
 127
 128         def find_param(name):
 129             if name in flashvars:
 130                 return flashvars[name]
 131             node = find_xpath_attr(object_doc, './param', 'name', name)
 132             if node is not None:
 133                 return node.attrib['value']
 134             return None
 135
 136         params = {}
 137
 138         playerID = find_param('playerID')
 139         if playerID is None:
 140             raise ExtractorError('Cannot find player ID')
 141         params['playerID'] = playerID
 142
 143         playerKey = find_param('playerKey')
 144         # Not all pages define this value
 145         if playerKey is not None:
 146             params['playerKey'] = playerKey
 147         # The three fields hold the id of the video
 148         videoPlayer = find_param('@videoPlayer') or find_param('videoId') or find_param('videoID')
 149         if videoPlayer is not None:
 150             params['@videoPlayer'] = videoPlayer
 151         linkBase = find_param('linkBaseURL')
 152         if linkBase is not None:
 153             params['linkBaseURL'] = linkBase
 154         data = compat_urllib_parse.urlencode(params)
 155         return cls._FEDERATED_URL_TEMPLATE % data
 156
 157     @classmethod
 158     def _extract_brightcove_url(cls, webpage):
 159         """Try to extract the brightcove url from the webpage, returns None
 160         if it can't be found
 161         """
 162         urls = cls._extract_brightcove_urls(webpage)
 163         return urls[0] if urls else None
 164
 165     @classmethod
 166     def _extract_brightcove_urls(cls, webpage):
 167         """Return a list of all Brightcove URLs from the webpage """
 168
 169         url_m = re.search(
 170             r'<meta\s+property="og:video"\s+content="(https?://(?:secure|c)\.brightcove.com/[^"]+)"',
 171             webpage)
 172         if url_m:
 173             url = unescapeHTML(url_m.group(1))
 174             # Some sites don't add it, we can't download with this url, for example:
 175             # http://www.ktvu.com/videos/news/raw-video-caltrain-releases-video-of-man-almost/vCTZdY/
 176             if 'playerKey' in url or 'videoId' in url:
 177                 return [url]
 178
 179         matches = re.findall(
 180             r'''(?sx)<object
 181             (?:
 182                 [^>]+?class=[\'"][^>]*?BrightcoveExperience.*?[\'"] |
 183                 [^>]*?>\s*<param\s+name="movie"\s+value="https?://[^/]*brightcove\.com/
 184             ).+?</object>''',
 185             webpage)
 186         return [cls._build_brighcove_url(m) for m in matches]
 187
 188     def _real_extract(self, url):
 189         url, smuggled_data = unsmuggle_url(url, {})
 190
 191         # Change the 'videoId' and others field to '@videoPlayer'
 192         url = re.sub(r'(?<=[?&])(videoI(d|D)|bctid)', '%40videoPlayer', url)
 193         # Change bckey (used by bcove.me urls) to playerKey
 194         url = re.sub(r'(?<=[?&])bckey', 'playerKey', url)
 195         mobj = re.match(self._VALID_URL, url)
 196         query_str = mobj.group('query')
 197         query = compat_urlparse.parse_qs(query_str)
 198
 199         videoPlayer = query.get('@videoPlayer')
 200         if videoPlayer:
 201             # We set the original url as the default 'Referer' header
 202             referer = smuggled_data.get('Referer', url)
 203             return self._get_video_info(
 204                 videoPlayer[0], query_str, query, referer=referer)
 205         elif 'playerKey' in query:
 206             player_key = query['playerKey']
 207             return self._get_playlist_info(player_key[0])
 208         else:
 209             raise ExtractorError(
 210                 'Cannot find playerKey= variable. Did you forget quotes in a shell invocation?',
 211                 expected=True)
 212
 213     def _get_video_info(self, video_id, query_str, query, referer=None):
 214         request_url = self._FEDERATED_URL_TEMPLATE % query_str
 215         req = compat_urllib_request.Request(request_url)
 216         linkBase = query.get('linkBaseURL')
 217         if linkBase is not None:
 218             referer = linkBase[0]
 219         if referer is not None:
 220             req.add_header('Referer', referer)
 221         webpage = self._download_webpage(req, video_id)
 222
 223         error_msg = self._html_search_regex(
 224             r"<h1>We're sorry.</h1>([\s\n]*<p>.*?</p>)+", webpage,
 225             'error message', default=None)
 226         if error_msg is not None:
 227             raise ExtractorError(
 228                 'brightcove said: %s' % error_msg, expected=True)
 229
 230         self.report_extraction(video_id)
 231         info = self._search_regex(r'var experienceJSON = ({.*});', webpage, 'json')
 232         info = json.loads(info)['data']
 233         video_info = info['programmedContent']['videoPlayer']['mediaDTO']
 234         video_info['_youtubedl_adServerURL'] = info.get('adServerURL')
 235
 236         return self._extract_video_info(video_info)
 237
 238     def _get_playlist_info(self, player_key):
 239         info_url = 'http://c.brightcove.com/services/json/experience/runtime/?command=get_programming_for_experience&playerKey=%s' % player_key
 240         playlist_info = self._download_webpage(
 241             info_url, player_key, 'Downloading playlist information')
 242
 243         json_data = json.loads(playlist_info)
 244         if 'videoList' not in json_data:
 245             raise ExtractorError('Empty playlist')
 246         playlist_info = json_data['videoList']
 247         videos = [self._extract_video_info(video_info) for video_info in playlist_info['mediaCollectionDTO']['videoDTOs']]
 248
 249         return self.playlist_result(videos, playlist_id=playlist_info['id'],
 250                                     playlist_title=playlist_info['mediaCollectionDTO']['displayName'])
 251
 252     def _extract_video_info(self, video_info):
 253         info = {
 254             'id': compat_str(video_info['id']),
 255             'title': video_info['displayName'].strip(),
 256             'description': video_info.get('shortDescription'),
 257             'thumbnail': video_info.get('videoStillURL') or video_info.get('thumbnailURL'),
 258             'uploader': video_info.get('publisherName'),
 259         }
 260
 261         renditions = video_info.get('renditions')
 262         if renditions:
 263             formats = []
 264             for rend in renditions:
 265                 url = rend['defaultURL']
 266                 if not url:
 267                     continue
 268                 if rend['remote']:
 269                     url_comp = compat_urllib_parse_urlparse(url)
 270                     if url_comp.path.endswith('.m3u8'):
 271                         formats.extend(
 272                             self._extract_m3u8_formats(url, info['id'], 'mp4'))
 273                         continue
 274                     elif 'akamaihd.net' in url_comp.netloc:
 275                         # This type of renditions are served through
 276                         # akamaihd.net, but they don't use f4m manifests
 277                         url = url.replace('control/', '') + '?&v=3.3.0&fp=13&r=FEEFJ&g=RTSJIMBMPFPB'
 278                         ext = 'flv'
 279                 else:
 280                     ext = determine_ext(url)
 281                 size = rend.get('size')
 282                 formats.append({
 283                     'url': url,
 284                     'ext': ext,
 285                     'height': rend.get('frameHeight'),
 286                     'width': rend.get('frameWidth'),
 287                     'filesize': size if size != 0 else None,
 288                 })
 289             self._sort_formats(formats)
 290             info['formats'] = formats
 291         elif video_info.get('FLVFullLengthURL') is not None:
 292             info.update({
 293                 'url': video_info['FLVFullLengthURL'],
 294             })
 295
 296         if self._downloader.params.get('include_ads', False):
 297             adServerURL = video_info.get('_youtubedl_adServerURL')
 298             if adServerURL:
 299                 ad_info = {
 300                     '_type': 'url',
 301                     'url': adServerURL,
 302                 }
 303                 if 'url' in info:
 304                     return {
 305                         '_type': 'playlist',
 306                         'title': info['title'],
 307                         'entries': [ad_info, info],
 308                     }
 309                 else:
 310                     return ad_info
 311
 312         if 'url' not in info and not info.get('formats'):
 313             raise ExtractorError('Unable to extract video url for %s' % info['id'])
 314         return info