_ Git - youtube-dl/blob - youtube_dl/extractor/kaltura.py

   1 # coding: utf-8
   2 from __future__ import unicode_literals
   3
   4 import re
   5 import base64
   6
   7 from .common import InfoExtractor
   8 from ..compat import (
   9     compat_urlparse,
  10     compat_parse_qs,
  11 )
  12 from ..utils import (
  13     clean_html,
  14     ExtractorError,
  15     int_or_none,
  16     unsmuggle_url,
  17     smuggle_url,
  18 )
  19
  20
  21 class KalturaIE(InfoExtractor):
  22     _VALID_URL = r'''(?x)
  23                 (?:
  24                     kaltura:(?P<partner_id>\d+):(?P<id>[0-9a-z_]+)|
  25                     https?://
  26                         (:?(?:www|cdnapi(?:sec)?)\.)?kaltura\.com/
  27                         (?:
  28                             (?:
  29                                 # flash player
  30                                 index\.php/kwidget|
  31                                 # html5 player
  32                                 html5/html5lib/[^/]+/mwEmbedFrame\.php
  33                             )
  34                         )(?:/(?P<path>[^?]+))?(?:\?(?P<query>.*))?
  35                 )
  36                 '''
  37     _SERVICE_URL = 'http://cdnapi.kaltura.com'
  38     _SERVICE_BASE = '/api_v3/index.php'
  39     _TESTS = [
  40         {
  41             'url': 'kaltura:269692:1_1jc2y3e4',
  42             'md5': '3adcbdb3dcc02d647539e53f284ba171',
  43             'info_dict': {
  44                 'id': '1_1jc2y3e4',
  45                 'ext': 'mp4',
  46                 'title': 'Straight from the Heart',
  47                 'upload_date': '20131219',
  48                 'uploader_id': 'mlundberg@wolfgangsvault.com',
  49                 'description': 'The Allman Brothers Band, 12/16/1981',
  50                 'thumbnail': 're:^https?://.*/thumbnail/.*',
  51                 'timestamp': int,
  52             },
  53         },
  54         {
  55             'url': 'http://www.kaltura.com/index.php/kwidget/cache_st/1300318621/wid/_269692/uiconf_id/3873291/entry_id/1_1jc2y3e4',
  56             'only_matching': True,
  57         },
  58         {
  59             'url': 'https://cdnapisec.kaltura.com/index.php/kwidget/wid/_557781/uiconf_id/22845202/entry_id/1_plr1syf3',
  60             'only_matching': True,
  61         },
  62         {
  63             'url': 'https://cdnapisec.kaltura.com/html5/html5lib/v2.30.2/mwEmbedFrame.php/p/1337/uiconf_id/20540612/entry_id/1_sf5ovm7u?wid=_243342',
  64             'only_matching': True,
  65         },
  66         {
  67             # video with subtitles
  68             'url': 'kaltura:111032:1_cw786r8q',
  69             'only_matching': True,
  70         },
  71         {
  72             # video with ttml subtitles (no fileExt)
  73             'url': 'kaltura:1926081:0_l5ye1133',
  74             'info_dict': {
  75                 'id': '0_l5ye1133',
  76                 'ext': 'mp4',
  77                 'title': 'What Can You Do With Python?',
  78                 'upload_date': '20160221',
  79                 'uploader_id': 'stork',
  80                 'thumbnail': 're:^https?://.*/thumbnail/.*',
  81                 'timestamp': int,
  82                 'subtitles': {
  83                     'en': [{
  84                         'ext': 'ttml',
  85                     }],
  86                 },
  87             },
  88             'params': {
  89                 'skip_download': True,
  90             },
  91         }
  92     ]
  93
  94     @staticmethod
  95     def _extract_url(webpage):
  96         mobj = (
  97             re.search(
  98                 r"""(?xs)
  99                     kWidget\.(?:thumb)?[Ee]mbed\(
 100                     \{.*?
 101                         (?P<q1>['\"])wid(?P=q1)\s*:\s*
 102                         (?P<q2>['\"])_?(?P<partner_id>[^'\"]+)(?P=q2),.*?
 103                         (?P<q3>['\"])entry_?[Ii]d(?P=q3)\s*:\s*
 104                         (?P<q4>['\"])(?P<id>[^'\"]+)(?P=q4),
 105                 """, webpage) or
 106             re.search(
 107                 r'''(?xs)
 108                     (?P<q1>["\'])
 109                         (?:https?:)?//cdnapi(?:sec)?\.kaltura\.com/.*?(?:p|partner_id)/(?P<partner_id>\d+).*?
 110                     (?P=q1).*?
 111                     (?:
 112                         entry_?[Ii]d|
 113                         (?P<q2>["\'])entry_?[Ii]d(?P=q2)
 114                     )\s*:\s*
 115                     (?P<q3>["\'])(?P<id>.+?)(?P=q3)
 116                 ''', webpage))
 117         if mobj:
 118             embed_info = mobj.groupdict()
 119             url = 'kaltura:%(partner_id)s:%(id)s' % embed_info
 120             escaped_pid = re.escape(embed_info['partner_id'])
 121             service_url = re.search(
 122                 r'<script[^>]+src=["\']((?:https?:)?//.+?)/p/%s/sp/%s00/embedIframeJs' % (escaped_pid, escaped_pid),
 123                 webpage)
 124             if service_url:
 125                 url = smuggle_url(url, {'service_url': service_url.group(1)})
 126             return url
 127
 128     def _kaltura_api_call(self, video_id, actions, service_url=None, *args, **kwargs):
 129         params = actions[0]
 130         if len(actions) > 1:
 131             for i, a in enumerate(actions[1:], start=1):
 132                 for k, v in a.items():
 133                     params['%d:%s' % (i, k)] = v
 134
 135         data = self._download_json(
 136             (service_url or self._SERVICE_URL) + self._SERVICE_BASE,
 137             video_id, query=params, *args, **kwargs)
 138
 139         status = data if len(actions) == 1 else data[0]
 140         if status.get('objectType') == 'KalturaAPIException':
 141             raise ExtractorError(
 142                 '%s said: %s' % (self.IE_NAME, status['message']))
 143
 144         return data
 145
 146     def _get_video_info(self, video_id, partner_id, service_url=None):
 147         actions = [
 148             {
 149                 'action': 'null',
 150                 'apiVersion': '3.1.5',
 151                 'clientTag': 'kdp:v3.8.5',
 152                 'format': 1,  # JSON, 2 = XML, 3 = PHP
 153                 'service': 'multirequest',
 154             },
 155             {
 156                 'expiry': 86400,
 157                 'service': 'session',
 158                 'action': 'startWidgetSession',
 159                 'widgetId': '_%s' % partner_id,
 160             },
 161             {
 162                 'action': 'get',
 163                 'entryId': video_id,
 164                 'service': 'baseentry',
 165                 'ks': '{1:result:ks}',
 166             },
 167             {
 168                 'action': 'getbyentryid',
 169                 'entryId': video_id,
 170                 'service': 'flavorAsset',
 171                 'ks': '{1:result:ks}',
 172             },
 173             {
 174                 'action': 'list',
 175                 'filter:entryIdEqual': video_id,
 176                 'service': 'caption_captionasset',
 177                 'ks': '{1:result:ks}',
 178             },
 179         ]
 180         return self._kaltura_api_call(
 181             video_id, actions, service_url, note='Downloading video info JSON')
 182
 183     def _real_extract(self, url):
 184         url, smuggled_data = unsmuggle_url(url, {})
 185
 186         mobj = re.match(self._VALID_URL, url)
 187         partner_id, entry_id = mobj.group('partner_id', 'id')
 188         ks = None
 189         captions = None
 190         if partner_id and entry_id:
 191             _, info, flavor_assets, captions = self._get_video_info(entry_id, partner_id, smuggled_data.get('service_url'))
 192         else:
 193             path, query = mobj.group('path', 'query')
 194             if not path and not query:
 195                 raise ExtractorError('Invalid URL', expected=True)
 196             params = {}
 197             if query:
 198                 params = compat_parse_qs(query)
 199             if path:
 200                 splitted_path = path.split('/')
 201                 params.update(dict((zip(splitted_path[::2], [[v] for v in splitted_path[1::2]]))))
 202             if 'wid' in params:
 203                 partner_id = params['wid'][0][1:]
 204             elif 'p' in params:
 205                 partner_id = params['p'][0]
 206             else:
 207                 raise ExtractorError('Invalid URL', expected=True)
 208             if 'entry_id' in params:
 209                 entry_id = params['entry_id'][0]
 210                 _, info, flavor_assets, captions = self._get_video_info(entry_id, partner_id)
 211             elif 'uiconf_id' in params and 'flashvars[referenceId]' in params:
 212                 reference_id = params['flashvars[referenceId]'][0]
 213                 webpage = self._download_webpage(url, reference_id)
 214                 entry_data = self._parse_json(self._search_regex(
 215                     r'window\.kalturaIframePackageData\s*=\s*({.*});',
 216                     webpage, 'kalturaIframePackageData'),
 217                     reference_id)['entryResult']
 218                 info, flavor_assets = entry_data['meta'], entry_data['contextData']['flavorAssets']
 219                 entry_id = info['id']
 220                 # Unfortunately, data returned in kalturaIframePackageData lacks
 221                 # captions so we will try requesting the complete data using
 222                 # regular approach since we now know the entry_id
 223                 try:
 224                     _, info, flavor_assets, captions = self._get_video_info(
 225                         entry_id, partner_id)
 226                 except ExtractorError:
 227                     # Regular scenario failed but we already have everything
 228                     # extracted apart from captions and can process at least
 229                     # with this
 230                     pass
 231             else:
 232                 raise ExtractorError('Invalid URL', expected=True)
 233             ks = params.get('flashvars[ks]', [None])[0]
 234
 235         source_url = smuggled_data.get('source_url')
 236         if source_url:
 237             referrer = base64.b64encode(
 238                 '://'.join(compat_urlparse.urlparse(source_url)[:2])
 239                 .encode('utf-8')).decode('utf-8')
 240         else:
 241             referrer = None
 242
 243         def sign_url(unsigned_url):
 244             if ks:
 245                 unsigned_url += '/ks/%s' % ks
 246             if referrer:
 247                 unsigned_url += '?referrer=%s' % referrer
 248             return unsigned_url
 249
 250         data_url = info['dataUrl']
 251         if '/flvclipper/' in data_url:
 252             data_url = re.sub(r'/flvclipper/.*', '/serveFlavor', data_url)
 253
 254         formats = []
 255         for f in flavor_assets:
 256             # Continue if asset is not ready
 257             if f.get('status') != 2:
 258                 continue
 259             video_url = sign_url(
 260                 '%s/flavorId/%s' % (data_url, f['id']))
 261             formats.append({
 262                 'format_id': '%(fileExt)s-%(bitrate)s' % f,
 263                 'ext': f.get('fileExt'),
 264                 'tbr': int_or_none(f['bitrate']),
 265                 'fps': int_or_none(f.get('frameRate')),
 266                 'filesize_approx': int_or_none(f.get('size'), invscale=1024),
 267                 'container': f.get('containerFormat'),
 268                 'vcodec': f.get('videoCodecId'),
 269                 'height': int_or_none(f.get('height')),
 270                 'width': int_or_none(f.get('width')),
 271                 'url': video_url,
 272             })
 273         if '/playManifest/' in data_url:
 274             m3u8_url = sign_url(data_url.replace(
 275                 'format/url', 'format/applehttp'))
 276             formats.extend(self._extract_m3u8_formats(
 277                 m3u8_url, entry_id, 'mp4', 'm3u8_native',
 278                 m3u8_id='hls', fatal=False))
 279
 280         self._sort_formats(formats)
 281
 282         subtitles = {}
 283         if captions:
 284             for caption in captions.get('objects', []):
 285                 # Continue if caption is not ready
 286                 if f.get('status') != 2:
 287                     continue
 288                 subtitles.setdefault(caption.get('languageCode') or caption.get('language'), []).append({
 289                     'url': '%s/api_v3/service/caption_captionasset/action/serve/captionAssetId/%s' % (self._SERVICE_URL, caption['id']),
 290                     'ext': caption.get('fileExt', 'ttml'),
 291                 })
 292
 293         return {
 294             'id': entry_id,
 295             'title': info['name'],
 296             'formats': formats,
 297             'subtitles': subtitles,
 298             'description': clean_html(info.get('description')),
 299             'thumbnail': info.get('thumbnailUrl'),
 300             'duration': info.get('duration'),
 301             'timestamp': info.get('createdAt'),
 302             'uploader_id': info.get('userId'),
 303             'view_count': info.get('plays'),
 304         }