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 ..compat import (
  10     compat_parse_qs,
  11     compat_str,
  12     compat_urllib_parse,
  13     compat_urllib_parse_urlparse,
  14     compat_urllib_request,
  15     compat_urlparse,
  16 )
  17 from ..utils import (
  18     determine_ext,
  19     ExtractorError,
  20     find_xpath_attr,
  21     fix_xml_ampersands,
  22     unescapeHTML,
  23     unsmuggle_url,
  24 )
  25
  26
  27 class BrightcoveIE(InfoExtractor):
  28     _VALID_URL = r'(?:https?://.*brightcove\.com/(services|viewer).*?\?|brightcove:)(?P<query>.*)'
  29     _FEDERATED_URL_TEMPLATE = 'http://c.brightcove.com/services/viewer/htmlFederated?%s'
  30
  31     _TESTS = [
  32         {
  33             # From http://www.8tv.cat/8aldia/videos/xavier-sala-i-martin-aquesta-tarda-a-8-al-dia/
  34             'url': 'http://c.brightcove.com/services/viewer/htmlFederated?playerID=1654948606001&flashID=myExperience&%40videoPlayer=2371591881001',
  35             'md5': '5423e113865d26e40624dce2e4b45d95',
  36             'note': 'Test Brightcove downloads and detection in GenericIE',
  37             'info_dict': {
  38                 'id': '2371591881001',
  39                 'ext': 'mp4',
  40                 'title': 'Xavier Sala i Martín: “Un banc que no presta és un banc zombi que no serveix per a res”',
  41                 'uploader': '8TV',
  42                 'description': 'md5:a950cc4285c43e44d763d036710cd9cd',
  43             }
  44         },
  45         {
  46             # From http://medianetwork.oracle.com/video/player/1785452137001
  47             'url': 'http://c.brightcove.com/services/viewer/htmlFederated?playerID=1217746023001&flashID=myPlayer&%40videoPlayer=1785452137001',
  48             'info_dict': {
  49                 'id': '1785452137001',
  50                 'ext': 'flv',
  51                 'title': 'JVMLS 2012: Arrays 2.0 - Opportunities and Challenges',
  52                 'description': 'John Rose speaks at the JVM Language Summit, August 1, 2012.',
  53                 'uploader': 'Oracle',
  54             },
  55         },
  56         {
  57             # From http://mashable.com/2013/10/26/thermoelectric-bracelet-lets-you-control-your-body-temperature/
  58             'url': 'http://c.brightcove.com/services/viewer/federated_f9?&playerID=1265504713001&publisherID=AQ%7E%7E%2CAAABBzUwv1E%7E%2CxP-xFHVUstiMFlNYfvF4G9yFnNaqCw_9&videoID=2750934548001',
  59             'info_dict': {
  60                 'id': '2750934548001',
  61                 'ext': 'mp4',
  62                 'title': 'This Bracelet Acts as a Personal Thermostat',
  63                 'description': 'md5:547b78c64f4112766ccf4e151c20b6a0',
  64                 'uploader': 'Mashable',
  65             },
  66         },
  67         {
  68             # test that the default referer works
  69             # from http://national.ballet.ca/interact/video/Lost_in_Motion_II/
  70             'url': 'http://link.brightcove.com/services/player/bcpid756015033001?bckey=AQ~~,AAAApYJi_Ck~,GxhXCegT1Dp39ilhXuxMJxasUhVNZiil&bctid=2878862109001',
  71             'info_dict': {
  72                 'id': '2878862109001',
  73                 'ext': 'mp4',
  74                 'title': 'Lost in Motion II',
  75                 'description': 'md5:363109c02998fee92ec02211bd8000df',
  76                 'uploader': 'National Ballet of Canada',
  77             },
  78         },
  79         {
  80             # test flv videos served by akamaihd.net
  81             # From http://www.redbull.com/en/bike/stories/1331655643987/replay-uci-dh-world-cup-2014-from-fort-william
  82             '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',
  83             # The md5 checksum changes on each download
  84             'info_dict': {
  85                 'id': '2996102916001',
  86                 'ext': 'flv',
  87                 'title': 'UCI MTB World Cup 2014: Fort William, UK - Downhill Finals',
  88                 'uploader': 'Red Bull TV',
  89                 'description': 'UCI MTB World Cup 2014: Fort William, UK - Downhill Finals',
  90             },
  91         },
  92         {
  93             # playlist test
  94             # from http://support.brightcove.com/en/video-cloud/docs/playlist-support-single-video-players
  95             'url': 'http://c.brightcove.com/services/viewer/htmlFederated?playerID=3550052898001&playerKey=AQ%7E%7E%2CAAABmA9XpXk%7E%2C-Kp7jNgisre1fG5OdqpAFUTcs0lP_ZoL',
  96             'info_dict': {
  97                 'title': 'Sealife',
  98                 'id': '3550319591001',
  99             },
 100             'playlist_mincount': 7,
 101         },
 102     ]
 103
 104     @classmethod
 105     def _build_brighcove_url(cls, object_str):
 106         """
 107         Build a Brightcove url from a xml string containing
 108         <object class="BrightcoveExperience">{params}</object>
 109         """
 110
 111         # Fix up some stupid HTML, see https://github.com/rg3/youtube-dl/issues/1553
 112         object_str = re.sub(r'(<param(?:\s+[a-zA-Z0-9_]+="[^"]*")*)>',
 113                             lambda m: m.group(1) + '/>', object_str)
 114         # Fix up some stupid XML, see https://github.com/rg3/youtube-dl/issues/1608
 115         object_str = object_str.replace('<--', '<!--')
 116         # remove namespace to simplify extraction
 117         object_str = re.sub(r'(<object[^>]*)(xmlns=".*?")', r'\1', object_str)
 118         object_str = fix_xml_ampersands(object_str)
 119
 120         object_doc = xml.etree.ElementTree.fromstring(object_str.encode('utf-8'))
 121
 122         fv_el = find_xpath_attr(object_doc, './param', 'name', 'flashVars')
 123         if fv_el is not None:
 124             flashvars = dict(
 125                 (k, v[0])
 126                 for k, v in compat_parse_qs(fv_el.attrib['value']).items())
 127         else:
 128             flashvars = {}
 129
 130         def find_param(name):
 131             if name in flashvars:
 132                 return flashvars[name]
 133             node = find_xpath_attr(object_doc, './param', 'name', name)
 134             if node is not None:
 135                 return node.attrib['value']
 136             return None
 137
 138         params = {}
 139
 140         playerID = find_param('playerID')
 141         if playerID is None:
 142             raise ExtractorError('Cannot find player ID')
 143         params['playerID'] = playerID
 144
 145         playerKey = find_param('playerKey')
 146         # Not all pages define this value
 147         if playerKey is not None:
 148             params['playerKey'] = playerKey
 149         # The three fields hold the id of the video
 150         videoPlayer = find_param('@videoPlayer') or find_param('videoId') or find_param('videoID')
 151         if videoPlayer is not None:
 152             params['@videoPlayer'] = videoPlayer
 153         linkBase = find_param('linkBaseURL')
 154         if linkBase is not None:
 155             params['linkBaseURL'] = linkBase
 156         data = compat_urllib_parse.urlencode(params)
 157         return cls._FEDERATED_URL_TEMPLATE % data
 158
 159     @classmethod
 160     def _extract_brightcove_url(cls, webpage):
 161         """Try to extract the brightcove url from the webpage, returns None
 162         if it can't be found
 163         """
 164         urls = cls._extract_brightcove_urls(webpage)
 165         return urls[0] if urls else None
 166
 167     @classmethod
 168     def _extract_brightcove_urls(cls, webpage):
 169         """Return a list of all Brightcove URLs from the webpage """
 170
 171         url_m = re.search(
 172             r'<meta\s+property="og:video"\s+content="(https?://(?:secure|c)\.brightcove.com/[^"]+)"',
 173             webpage)
 174         if url_m:
 175             url = unescapeHTML(url_m.group(1))
 176             # Some sites don't add it, we can't download with this url, for example:
 177             # http://www.ktvu.com/videos/news/raw-video-caltrain-releases-video-of-man-almost/vCTZdY/
 178             if 'playerKey' in url or 'videoId' in url:
 179                 return [url]
 180
 181         matches = re.findall(
 182             r'''(?sx)<object
 183             (?:
 184                 [^>]+?class=[\'"][^>]*?BrightcoveExperience.*?[\'"] |
 185                 [^>]*?>\s*<param\s+name="movie"\s+value="https?://[^/]*brightcove\.com/
 186             ).+?</object>''',
 187             webpage)
 188         return [cls._build_brighcove_url(m) for m in matches]
 189
 190     def _real_extract(self, url):
 191         url, smuggled_data = unsmuggle_url(url, {})
 192
 193         # Change the 'videoId' and others field to '@videoPlayer'
 194         url = re.sub(r'(?<=[?&])(videoI(d|D)|bctid)', '%40videoPlayer', url)
 195         # Change bckey (used by bcove.me urls) to playerKey
 196         url = re.sub(r'(?<=[?&])bckey', 'playerKey', url)
 197         mobj = re.match(self._VALID_URL, url)
 198         query_str = mobj.group('query')
 199         query = compat_urlparse.parse_qs(query_str)
 200
 201         videoPlayer = query.get('@videoPlayer')
 202         if videoPlayer:
 203             # We set the original url as the default 'Referer' header
 204             referer = smuggled_data.get('Referer', url)
 205             return self._get_video_info(
 206                 videoPlayer[0], query_str, query, referer=referer)
 207         elif 'playerKey' in query:
 208             player_key = query['playerKey']
 209             return self._get_playlist_info(player_key[0])
 210         else:
 211             raise ExtractorError(
 212                 'Cannot find playerKey= variable. Did you forget quotes in a shell invocation?',
 213                 expected=True)
 214
 215     def _get_video_info(self, video_id, query_str, query, referer=None):
 216         request_url = self._FEDERATED_URL_TEMPLATE % query_str
 217         req = compat_urllib_request.Request(request_url)
 218         linkBase = query.get('linkBaseURL')
 219         if linkBase is not None:
 220             referer = linkBase[0]
 221         if referer is not None:
 222             req.add_header('Referer', referer)
 223         webpage = self._download_webpage(req, video_id)
 224
 225         error_msg = self._html_search_regex(
 226             r"<h1>We're sorry.</h1>([\s\n]*<p>.*?</p>)+", webpage,
 227             'error message', default=None)
 228         if error_msg is not None:
 229             raise ExtractorError(
 230                 'brightcove said: %s' % error_msg, expected=True)
 231
 232         self.report_extraction(video_id)
 233         info = self._search_regex(r'var experienceJSON = ({.*});', webpage, 'json')
 234         info = json.loads(info)['data']
 235         video_info = info['programmedContent']['videoPlayer']['mediaDTO']
 236         video_info['_youtubedl_adServerURL'] = info.get('adServerURL')
 237
 238         return self._extract_video_info(video_info)
 239
 240     def _get_playlist_info(self, player_key):
 241         info_url = 'http://c.brightcove.com/services/json/experience/runtime/?command=get_programming_for_experience&playerKey=%s' % player_key
 242         playlist_info = self._download_webpage(
 243             info_url, player_key, 'Downloading playlist information')
 244
 245         json_data = json.loads(playlist_info)
 246         if 'videoList' not in json_data:
 247             raise ExtractorError('Empty playlist')
 248         playlist_info = json_data['videoList']
 249         videos = [self._extract_video_info(video_info) for video_info in playlist_info['mediaCollectionDTO']['videoDTOs']]
 250
 251         return self.playlist_result(videos, playlist_id='%s' % playlist_info['id'],
 252                                     playlist_title=playlist_info['mediaCollectionDTO']['displayName'])
 253
 254     def _extract_video_info(self, video_info):
 255         info = {
 256             'id': compat_str(video_info['id']),
 257             'title': video_info['displayName'].strip(),
 258             'description': video_info.get('shortDescription'),
 259             'thumbnail': video_info.get('videoStillURL') or video_info.get('thumbnailURL'),
 260             'uploader': video_info.get('publisherName'),
 261         }
 262
 263         renditions = video_info.get('renditions')
 264         if renditions:
 265             formats = []
 266             for rend in renditions:
 267                 url = rend['defaultURL']
 268                 if not url:
 269                     continue
 270                 ext = None
 271                 if rend['remote']:
 272                     url_comp = compat_urllib_parse_urlparse(url)
 273                     if url_comp.path.endswith('.m3u8'):
 274                         formats.extend(
 275                             self._extract_m3u8_formats(url, info['id'], 'mp4'))
 276                         continue
 277                     elif 'akamaihd.net' in url_comp.netloc:
 278                         # This type of renditions are served through
 279                         # akamaihd.net, but they don't use f4m manifests
 280                         url = url.replace('control/', '') + '?&v=3.3.0&fp=13&r=FEEFJ&g=RTSJIMBMPFPB'
 281                         ext = 'flv'
 282                 if ext is None:
 283                     ext = determine_ext(url)
 284                 size = rend.get('size')
 285                 formats.append({
 286                     'url': url,
 287                     'ext': ext,
 288                     'height': rend.get('frameHeight'),
 289                     'width': rend.get('frameWidth'),
 290                     'filesize': size if size != 0 else None,
 291                 })
 292             self._sort_formats(formats)
 293             info['formats'] = formats
 294         elif video_info.get('FLVFullLengthURL') is not None:
 295             info.update({
 296                 'url': video_info['FLVFullLengthURL'],
 297             })
 298
 299         if self._downloader.params.get('include_ads', False):
 300             adServerURL = video_info.get('_youtubedl_adServerURL')
 301             if adServerURL:
 302                 ad_info = {
 303                     '_type': 'url',
 304                     'url': adServerURL,
 305                 }
 306                 if 'url' in info:
 307                     return {
 308                         '_type': 'playlist',
 309                         'title': info['title'],
 310                         'entries': [ad_info, info],
 311                     }
 312                 else:
 313                     return ad_info
 314
 315         if 'url' not in info and not info.get('formats'):
 316             raise ExtractorError('Unable to extract video url for %s' % info['id'])
 317         return info