yt_dlp/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     format_field,
  16     int_or_none,
  17     unsmuggle_url,
  18     smuggle_url,
  19 )
  20
  21
  22 class KalturaIE(InfoExtractor):
  23     _VALID_URL = r'''(?x)
  24                 (?:
  25                     kaltura:(?P<partner_id>\d+):(?P<id>[0-9a-z_]+)|
  26                     https?://
  27                         (:?(?:www|cdnapi(?:sec)?)\.)?kaltura\.com(?::\d+)?/
  28                         (?:
  29                             (?:
  30                                 # flash player
  31                                 index\.php/(?:kwidget|extwidget/preview)|
  32                                 # html5 player
  33                                 html5/html5lib/[^/]+/mwEmbedFrame\.php
  34                             )
  35                         )(?:/(?P<path>[^?]+))?(?:\?(?P<query>.*))?
  36                 )
  37                 '''
  38     _SERVICE_URL = 'http://cdnapi.kaltura.com'
  39     _SERVICE_BASE = '/api_v3/index.php'
  40     # See https://github.com/kaltura/server/blob/master/plugins/content/caption/base/lib/model/enums/CaptionType.php
  41     _CAPTION_TYPES = {
  42         1: 'srt',
  43         2: 'ttml',
  44         3: 'vtt',
  45     }
  46     _TESTS = [
  47         {
  48             'url': 'kaltura:269692:1_1jc2y3e4',
  49             'md5': '3adcbdb3dcc02d647539e53f284ba171',
  50             'info_dict': {
  51                 'id': '1_1jc2y3e4',
  52                 'ext': 'mp4',
  53                 'title': 'Straight from the Heart',
  54                 'upload_date': '20131219',
  55                 'uploader_id': 'mlundberg@wolfgangsvault.com',
  56                 'description': 'The Allman Brothers Band, 12/16/1981',
  57                 'thumbnail': 're:^https?://.*/thumbnail/.*',
  58                 'timestamp': int,
  59             },
  60         },
  61         {
  62             'url': 'http://www.kaltura.com/index.php/kwidget/cache_st/1300318621/wid/_269692/uiconf_id/3873291/entry_id/1_1jc2y3e4',
  63             'only_matching': True,
  64         },
  65         {
  66             'url': 'https://cdnapisec.kaltura.com/index.php/kwidget/wid/_557781/uiconf_id/22845202/entry_id/1_plr1syf3',
  67             'only_matching': True,
  68         },
  69         {
  70             'url': 'https://cdnapisec.kaltura.com/html5/html5lib/v2.30.2/mwEmbedFrame.php/p/1337/uiconf_id/20540612/entry_id/1_sf5ovm7u?wid=_243342',
  71             'only_matching': True,
  72         },
  73         {
  74             # video with subtitles
  75             'url': 'kaltura:111032:1_cw786r8q',
  76             'only_matching': True,
  77         },
  78         {
  79             # video with ttml subtitles (no fileExt)
  80             'url': 'kaltura:1926081:0_l5ye1133',
  81             'info_dict': {
  82                 'id': '0_l5ye1133',
  83                 'ext': 'mp4',
  84                 'title': 'What Can You Do With Python?',
  85                 'upload_date': '20160221',
  86                 'uploader_id': 'stork',
  87                 'thumbnail': 're:^https?://.*/thumbnail/.*',
  88                 'timestamp': int,
  89                 'subtitles': {
  90                     'en': [{
  91                         'ext': 'ttml',
  92                     }],
  93                 },
  94             },
  95             'skip': 'Gone. Maybe https://www.safaribooksonline.com/library/tutorials/introduction-to-python-anon/3469/',
  96             'params': {
  97                 'skip_download': True,
  98             },
  99         },
 100         {
 101             'url': 'https://www.kaltura.com/index.php/extwidget/preview/partner_id/1770401/uiconf_id/37307382/entry_id/0_58u8kme7/embed/iframe?&flashvars[streamerType]=auto',
 102             'only_matching': True,
 103         },
 104         {
 105             'url': 'https://www.kaltura.com:443/index.php/extwidget/preview/partner_id/1770401/uiconf_id/37307382/entry_id/0_58u8kme7/embed/iframe?&flashvars[streamerType]=auto',
 106             'only_matching': True,
 107         },
 108         {
 109             # unavailable source format
 110             'url': 'kaltura:513551:1_66x4rg7o',
 111             'only_matching': True,
 112         }
 113     ]
 114
 115     @staticmethod
 116     def _extract_url(webpage):
 117         urls = KalturaIE._extract_urls(webpage)
 118         return urls[0] if urls else None
 119
 120     @staticmethod
 121     def _extract_urls(webpage):
 122         # Embed codes: https://knowledge.kaltura.com/embedding-kaltura-media-players-your-site
 123         finditer = (
 124             list(re.finditer(
 125                 r"""(?xs)
 126                     kWidget\.(?:thumb)?[Ee]mbed\(
 127                     \{.*?
 128                         (?P<q1>['"])wid(?P=q1)\s*:\s*
 129                         (?P<q2>['"])_?(?P<partner_id>(?:(?!(?P=q2)).)+)(?P=q2),.*?
 130                         (?P<q3>['"])entry_?[Ii]d(?P=q3)\s*:\s*
 131                         (?P<q4>['"])(?P<id>(?:(?!(?P=q4)).)+)(?P=q4)(?:,|\s*\})
 132                 """, webpage))
 133             or list(re.finditer(
 134                 r'''(?xs)
 135                     (?P<q1>["'])
 136                         (?:https?:)?//cdnapi(?:sec)?\.kaltura\.com(?::\d+)?/(?:(?!(?P=q1)).)*\b(?:p|partner_id)/(?P<partner_id>\d+)(?:(?!(?P=q1)).)*
 137                     (?P=q1).*?
 138                     (?:
 139                         (?:
 140                             entry_?[Ii]d|
 141                             (?P<q2>["'])entry_?[Ii]d(?P=q2)
 142                         )\s*:\s*|
 143                         \[\s*(?P<q2_1>["'])entry_?[Ii]d(?P=q2_1)\s*\]\s*=\s*
 144                     )
 145                     (?P<q3>["'])(?P<id>(?:(?!(?P=q3)).)+)(?P=q3)
 146                 ''', webpage))
 147             or list(re.finditer(
 148                 r'''(?xs)
 149                     <(?:iframe[^>]+src|meta[^>]+\bcontent)=(?P<q1>["'])\s*
 150                       (?:https?:)?//(?:(?:www|cdnapi(?:sec)?)\.)?kaltura\.com/(?:(?!(?P=q1)).)*\b(?:p|partner_id)/(?P<partner_id>\d+)
 151                       (?:(?!(?P=q1)).)*
 152                       [?&;]entry_id=(?P<id>(?:(?!(?P=q1))[^&])+)
 153                       (?:(?!(?P=q1)).)*
 154                     (?P=q1)
 155                 ''', webpage))
 156         )
 157         urls = []
 158         for mobj in finditer:
 159             embed_info = mobj.groupdict()
 160             for k, v in embed_info.items():
 161                 if v:
 162                     embed_info[k] = v.strip()
 163             url = 'kaltura:%(partner_id)s:%(id)s' % embed_info
 164             escaped_pid = re.escape(embed_info['partner_id'])
 165             service_mobj = re.search(
 166                 r'<script[^>]+src=(["\'])(?P<id>(?:https?:)?//(?:(?!\1).)+)/p/%s/sp/%s00/embedIframeJs' % (escaped_pid, escaped_pid),
 167                 webpage)
 168             if service_mobj:
 169                 url = smuggle_url(url, {'service_url': service_mobj.group('id')})
 170             urls.append(url)
 171         return urls
 172
 173     def _kaltura_api_call(self, video_id, actions, service_url=None, *args, **kwargs):
 174         params = actions[0]
 175         if len(actions) > 1:
 176             for i, a in enumerate(actions[1:], start=1):
 177                 for k, v in a.items():
 178                     params['%d:%s' % (i, k)] = v
 179
 180         data = self._download_json(
 181             (service_url or self._SERVICE_URL) + self._SERVICE_BASE,
 182             video_id, query=params, *args, **kwargs)
 183
 184         status = data if len(actions) == 1 else data[0]
 185         if status.get('objectType') == 'KalturaAPIException':
 186             raise ExtractorError(
 187                 '%s said: %s' % (self.IE_NAME, status['message']))
 188
 189         return data
 190
 191     def _get_video_info(self, video_id, partner_id, service_url=None):
 192         actions = [
 193             {
 194                 'action': 'null',
 195                 'apiVersion': '3.1.5',
 196                 'clientTag': 'kdp:v3.8.5',
 197                 'format': 1,  # JSON, 2 = XML, 3 = PHP
 198                 'service': 'multirequest',
 199             },
 200             {
 201                 'expiry': 86400,
 202                 'service': 'session',
 203                 'action': 'startWidgetSession',
 204                 'widgetId': '_%s' % partner_id,
 205             },
 206             {
 207                 'action': 'get',
 208                 'entryId': video_id,
 209                 'service': 'baseentry',
 210                 'ks': '{1:result:ks}',
 211                 'responseProfile:fields': 'createdAt,dataUrl,duration,name,plays,thumbnailUrl,userId',
 212                 'responseProfile:type': 1,
 213             },
 214             {
 215                 'action': 'getbyentryid',
 216                 'entryId': video_id,
 217                 'service': 'flavorAsset',
 218                 'ks': '{1:result:ks}',
 219             },
 220             {
 221                 'action': 'list',
 222                 'filter:entryIdEqual': video_id,
 223                 'service': 'caption_captionasset',
 224                 'ks': '{1:result:ks}',
 225             },
 226         ]
 227         return self._kaltura_api_call(
 228             video_id, actions, service_url, note='Downloading video info JSON')
 229
 230     def _real_extract(self, url):
 231         url, smuggled_data = unsmuggle_url(url, {})
 232
 233         mobj = self._match_valid_url(url)
 234         partner_id, entry_id = mobj.group('partner_id', 'id')
 235         ks = None
 236         captions = None
 237         if partner_id and entry_id:
 238             _, info, flavor_assets, captions = self._get_video_info(entry_id, partner_id, smuggled_data.get('service_url'))
 239         else:
 240             path, query = mobj.group('path', 'query')
 241             if not path and not query:
 242                 raise ExtractorError('Invalid URL', expected=True)
 243             params = {}
 244             if query:
 245                 params = compat_parse_qs(query)
 246             if path:
 247                 splitted_path = path.split('/')
 248                 params.update(dict((zip(splitted_path[::2], [[v] for v in splitted_path[1::2]]))))
 249             if 'wid' in params:
 250                 partner_id = params['wid'][0][1:]
 251             elif 'p' in params:
 252                 partner_id = params['p'][0]
 253             elif 'partner_id' in params:
 254                 partner_id = params['partner_id'][0]
 255             else:
 256                 raise ExtractorError('Invalid URL', expected=True)
 257             if 'entry_id' in params:
 258                 entry_id = params['entry_id'][0]
 259                 _, info, flavor_assets, captions = self._get_video_info(entry_id, partner_id)
 260             elif 'uiconf_id' in params and 'flashvars[referenceId]' in params:
 261                 reference_id = params['flashvars[referenceId]'][0]
 262                 webpage = self._download_webpage(url, reference_id)
 263                 entry_data = self._parse_json(self._search_regex(
 264                     r'window\.kalturaIframePackageData\s*=\s*({.*});',
 265                     webpage, 'kalturaIframePackageData'),
 266                     reference_id)['entryResult']
 267                 info, flavor_assets = entry_data['meta'], entry_data['contextData']['flavorAssets']
 268                 entry_id = info['id']
 269                 # Unfortunately, data returned in kalturaIframePackageData lacks
 270                 # captions so we will try requesting the complete data using
 271                 # regular approach since we now know the entry_id
 272                 try:
 273                     _, info, flavor_assets, captions = self._get_video_info(
 274                         entry_id, partner_id)
 275                 except ExtractorError:
 276                     # Regular scenario failed but we already have everything
 277                     # extracted apart from captions and can process at least
 278                     # with this
 279                     pass
 280             else:
 281                 raise ExtractorError('Invalid URL', expected=True)
 282             ks = params.get('flashvars[ks]', [None])[0]
 283
 284         source_url = smuggled_data.get('source_url')
 285         if source_url:
 286             referrer = base64.b64encode(
 287                 '://'.join(compat_urlparse.urlparse(source_url)[:2])
 288                 .encode('utf-8')).decode('utf-8')
 289         else:
 290             referrer = None
 291
 292         def sign_url(unsigned_url):
 293             if ks:
 294                 unsigned_url += '/ks/%s' % ks
 295             if referrer:
 296                 unsigned_url += '?referrer=%s' % referrer
 297             return unsigned_url
 298
 299         data_url = info['dataUrl']
 300         if '/flvclipper/' in data_url:
 301             data_url = re.sub(r'/flvclipper/.*', '/serveFlavor', data_url)
 302
 303         formats = []
 304         for f in flavor_assets:
 305             # Continue if asset is not ready
 306             if f.get('status') != 2:
 307                 continue
 308             # Original format that's not available (e.g. kaltura:1926081:0_c03e1b5g)
 309             # skip for now.
 310             if f.get('fileExt') == 'chun':
 311                 continue
 312             # DRM-protected video, cannot be decrypted
 313             if not self.get_param('allow_unplayable_formats') and f.get('fileExt') == 'wvm':
 314                 continue
 315             if not f.get('fileExt'):
 316                 # QT indicates QuickTime; some videos have broken fileExt
 317                 if f.get('containerFormat') == 'qt':
 318                     f['fileExt'] = 'mov'
 319                 else:
 320                     f['fileExt'] = 'mp4'
 321             video_url = sign_url(
 322                 '%s/flavorId/%s' % (data_url, f['id']))
 323             format_id = '%(fileExt)s-%(bitrate)s' % f
 324             # Source format may not be available (e.g. kaltura:513551:1_66x4rg7o)
 325             if f.get('isOriginal') is True and not self._is_valid_url(
 326                     video_url, entry_id, format_id):
 327                 continue
 328             # audio-only has no videoCodecId (e.g. kaltura:1926081:0_c03e1b5g
 329             # -f mp4-56)
 330             vcodec = 'none' if 'videoCodecId' not in f and f.get(
 331                 'frameRate') == 0 else f.get('videoCodecId')
 332             formats.append({
 333                 'format_id': format_id,
 334                 'ext': f.get('fileExt'),
 335                 'tbr': int_or_none(f['bitrate']),
 336                 'fps': int_or_none(f.get('frameRate')),
 337                 'filesize_approx': int_or_none(f.get('size'), invscale=1024),
 338                 'container': f.get('containerFormat'),
 339                 'vcodec': vcodec,
 340                 'height': int_or_none(f.get('height')),
 341                 'width': int_or_none(f.get('width')),
 342                 'url': video_url,
 343             })
 344         if '/playManifest/' in data_url:
 345             m3u8_url = sign_url(data_url.replace(
 346                 'format/url', 'format/applehttp'))
 347             formats.extend(self._extract_m3u8_formats(
 348                 m3u8_url, entry_id, 'mp4', 'm3u8_native',
 349                 m3u8_id='hls', fatal=False))
 350
 351         self._sort_formats(formats)
 352
 353         subtitles = {}
 354         if captions:
 355             for caption in captions.get('objects', []):
 356                 # Continue if caption is not ready
 357                 if caption.get('status') != 2:
 358                     continue
 359                 if not caption.get('id'):
 360                     continue
 361                 caption_format = int_or_none(caption.get('format'))
 362                 subtitles.setdefault(caption.get('languageCode') or caption.get('language'), []).append({
 363                     'url': '%s/api_v3/service/caption_captionasset/action/serve/captionAssetId/%s' % (self._SERVICE_URL, caption['id']),
 364                     'ext': caption.get('fileExt') or self._CAPTION_TYPES.get(caption_format) or 'ttml',
 365                 })
 366
 367         return {
 368             'id': entry_id,
 369             'title': info['name'],
 370             'formats': formats,
 371             'subtitles': subtitles,
 372             'description': clean_html(info.get('description')),
 373             'thumbnail': info.get('thumbnailUrl'),
 374             'duration': info.get('duration'),
 375             'timestamp': info.get('createdAt'),
 376             'uploader_id': format_field(info, 'userId', ignore=('None', None)),
 377             'view_count': info.get('plays'),
 378         }