yt_dlp/extractor/kaltura.py

   1 import base64
   2 import json
   3 import re
   4
   5 from .common import InfoExtractor
   6 from ..compat import (
   7     compat_urlparse,
   8     compat_parse_qs,
   9 )
  10 from ..utils import (
  11     clean_html,
  12     ExtractorError,
  13     format_field,
  14     int_or_none,
  15     unsmuggle_url,
  16     smuggle_url,
  17     traverse_obj,
  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(?::\d+)?/
  27                         (?:
  28                             (?:
  29                                 # flash player
  30                                 index\.php/(?:kwidget|extwidget/preview)|
  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/service/multirequest'
  39     # See https://github.com/kaltura/server/blob/master/plugins/content/caption/base/lib/model/enums/CaptionType.php
  40     _CAPTION_TYPES = {
  41         1: 'srt',
  42         2: 'ttml',
  43         3: 'vtt',
  44     }
  45     _TESTS = [
  46         {
  47             'url': 'kaltura:269692:1_1jc2y3e4',
  48             'md5': '3adcbdb3dcc02d647539e53f284ba171',
  49             'info_dict': {
  50                 'id': '1_1jc2y3e4',
  51                 'ext': 'mp4',
  52                 'title': 'Straight from the Heart',
  53                 'upload_date': '20131219',
  54                 'uploader_id': 'mlundberg@wolfgangsvault.com',
  55                 'description': 'The Allman Brothers Band, 12/16/1981',
  56                 'thumbnail': 're:^https?://.*/thumbnail/.*',
  57                 'timestamp': int,
  58             },
  59         },
  60         {
  61             'url': 'http://www.kaltura.com/index.php/kwidget/cache_st/1300318621/wid/_269692/uiconf_id/3873291/entry_id/1_1jc2y3e4',
  62             'only_matching': True,
  63         },
  64         {
  65             'url': 'https://cdnapisec.kaltura.com/index.php/kwidget/wid/_557781/uiconf_id/22845202/entry_id/1_plr1syf3',
  66             'only_matching': True,
  67         },
  68         {
  69             'url': 'https://cdnapisec.kaltura.com/html5/html5lib/v2.30.2/mwEmbedFrame.php/p/1337/uiconf_id/20540612/entry_id/1_sf5ovm7u?wid=_243342',
  70             'only_matching': True,
  71         },
  72         {
  73             # video with subtitles
  74             'url': 'kaltura:111032:1_cw786r8q',
  75             'only_matching': True,
  76         },
  77         {
  78             # video with ttml subtitles (no fileExt)
  79             'url': 'kaltura:1926081:0_l5ye1133',
  80             'info_dict': {
  81                 'id': '0_l5ye1133',
  82                 'ext': 'mp4',
  83                 'title': 'What Can You Do With Python?',
  84                 'upload_date': '20160221',
  85                 'uploader_id': 'stork',
  86                 'thumbnail': 're:^https?://.*/thumbnail/.*',
  87                 'timestamp': int,
  88                 'subtitles': {
  89                     'en': [{
  90                         'ext': 'ttml',
  91                     }],
  92                 },
  93             },
  94             'skip': 'Gone. Maybe https://www.safaribooksonline.com/library/tutorials/introduction-to-python-anon/3469/',
  95             'params': {
  96                 'skip_download': True,
  97             },
  98         },
  99         {
 100             'url': 'https://www.kaltura.com/index.php/extwidget/preview/partner_id/1770401/uiconf_id/37307382/entry_id/0_58u8kme7/embed/iframe?&flashvars[streamerType]=auto',
 101             'only_matching': True,
 102         },
 103         {
 104             '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',
 105             'only_matching': True,
 106         },
 107         {
 108             # unavailable source format
 109             'url': 'kaltura:513551:1_66x4rg7o',
 110             'only_matching': True,
 111         }
 112     ]
 113
 114     @staticmethod
 115     def _extract_url(webpage):
 116         urls = KalturaIE._extract_urls(webpage)
 117         return urls[0] if urls else None
 118
 119     @staticmethod
 120     def _extract_urls(webpage):
 121         # Embed codes: https://knowledge.kaltura.com/embedding-kaltura-media-players-your-site
 122         finditer = (
 123             list(re.finditer(
 124                 r"""(?xs)
 125                     kWidget\.(?:thumb)?[Ee]mbed\(
 126                     \{.*?
 127                         (?P<q1>['"])wid(?P=q1)\s*:\s*
 128                         (?P<q2>['"])_?(?P<partner_id>(?:(?!(?P=q2)).)+)(?P=q2),.*?
 129                         (?P<q3>['"])entry_?[Ii]d(?P=q3)\s*:\s*
 130                         (?P<q4>['"])(?P<id>(?:(?!(?P=q4)).)+)(?P=q4)(?:,|\s*\})
 131                 """, webpage))
 132             or list(re.finditer(
 133                 r'''(?xs)
 134                     (?P<q1>["'])
 135                         (?:https?:)?//cdnapi(?:sec)?\.kaltura\.com(?::\d+)?/(?:(?!(?P=q1)).)*\b(?:p|partner_id)/(?P<partner_id>\d+)(?:(?!(?P=q1)).)*
 136                     (?P=q1).*?
 137                     (?:
 138                         (?:
 139                             entry_?[Ii]d|
 140                             (?P<q2>["'])entry_?[Ii]d(?P=q2)
 141                         )\s*:\s*|
 142                         \[\s*(?P<q2_1>["'])entry_?[Ii]d(?P=q2_1)\s*\]\s*=\s*
 143                     )
 144                     (?P<q3>["'])(?P<id>(?:(?!(?P=q3)).)+)(?P=q3)
 145                 ''', webpage))
 146             or list(re.finditer(
 147                 r'''(?xs)
 148                     <(?:iframe[^>]+src|meta[^>]+\bcontent)=(?P<q1>["'])\s*
 149                       (?:https?:)?//(?:(?:www|cdnapi(?:sec)?)\.)?kaltura\.com/(?:(?!(?P=q1)).)*\b(?:p|partner_id)/(?P<partner_id>\d+)
 150                       (?:(?!(?P=q1)).)*
 151                       [?&;]entry_id=(?P<id>(?:(?!(?P=q1))[^&])+)
 152                       (?:(?!(?P=q1)).)*
 153                     (?P=q1)
 154                 ''', webpage))
 155         )
 156         urls = []
 157         for mobj in finditer:
 158             embed_info = mobj.groupdict()
 159             for k, v in embed_info.items():
 160                 if v:
 161                     embed_info[k] = v.strip()
 162             url = 'kaltura:%(partner_id)s:%(id)s' % embed_info
 163             escaped_pid = re.escape(embed_info['partner_id'])
 164             service_mobj = re.search(
 165                 r'<script[^>]+src=(["\'])(?P<id>(?:https?:)?//(?:(?!\1).)+)/p/%s/sp/%s00/embedIframeJs' % (escaped_pid, escaped_pid),
 166                 webpage)
 167             if service_mobj:
 168                 url = smuggle_url(url, {'service_url': service_mobj.group('id')})
 169             urls.append(url)
 170         return urls
 171
 172     def _kaltura_api_call(self, video_id, actions, service_url=None, *args, **kwargs):
 173         params = actions[0]
 174         params.update({i: a for i, a in enumerate(actions[1:], start=1)})
 175
 176         data = self._download_json(
 177             (service_url or self._SERVICE_URL) + self._SERVICE_BASE,
 178             video_id, data=json.dumps(params).encode('utf-8'),
 179             headers={
 180                 'Content-Type': 'application/json',
 181                 'Accept-Encoding': 'gzip, deflate, br',
 182             }, *args, **kwargs)
 183
 184         for idx, status in enumerate(data):
 185             if not isinstance(status, dict):
 186                 continue
 187             if status.get('objectType') == 'KalturaAPIException':
 188                 raise ExtractorError(
 189                     '%s said: %s (%d)' % (self.IE_NAME, status['message'], idx))
 190
 191         data[1] = traverse_obj(data, (1, 'objects', 0))
 192
 193         return data
 194
 195     def _get_video_info(self, video_id, partner_id, service_url=None):
 196         actions = [
 197             {
 198                 'apiVersion': '3.3.0',
 199                 'clientTag': 'html5:v3.1.0',
 200                 'format': 1,  # JSON, 2 = XML, 3 = PHP
 201                 'ks': '',
 202                 'partnerId': partner_id,
 203             },
 204             {
 205                 'expiry': 86400,
 206                 'service': 'session',
 207                 'action': 'startWidgetSession',
 208                 'widgetId': '_%s' % partner_id,
 209             },
 210             {
 211                 'action': 'list',
 212                 'filter': {'redirectFromEntryId': video_id},
 213                 'service': 'baseentry',
 214                 'ks': '{1:result:ks}',
 215                 'responseProfile': {
 216                     'type': 1,
 217                     'fields': 'createdAt,dataUrl,duration,name,plays,thumbnailUrl,userId',
 218                 },
 219             },
 220             {
 221                 'action': 'getbyentryid',
 222                 'entryId': video_id,
 223                 'service': 'flavorAsset',
 224                 'ks': '{1:result:ks}',
 225             },
 226             {
 227                 'action': 'list',
 228                 'filter:entryIdEqual': video_id,
 229                 'service': 'caption_captionasset',
 230                 'ks': '{1:result:ks}',
 231             },
 232         ]
 233         return self._kaltura_api_call(
 234             video_id, actions, service_url, note='Downloading video info JSON')
 235
 236     def _real_extract(self, url):
 237         url, smuggled_data = unsmuggle_url(url, {})
 238
 239         mobj = self._match_valid_url(url)
 240         partner_id, entry_id = mobj.group('partner_id', 'id')
 241         ks = None
 242         captions = None
 243         if partner_id and entry_id:
 244             _, info, flavor_assets, captions = self._get_video_info(entry_id, partner_id, smuggled_data.get('service_url'))
 245         else:
 246             path, query = mobj.group('path', 'query')
 247             if not path and not query:
 248                 raise ExtractorError('Invalid URL', expected=True)
 249             params = {}
 250             if query:
 251                 params = compat_parse_qs(query)
 252             if path:
 253                 splitted_path = path.split('/')
 254                 params.update(dict((zip(splitted_path[::2], [[v] for v in splitted_path[1::2]]))))
 255             if 'wid' in params:
 256                 partner_id = params['wid'][0][1:]
 257             elif 'p' in params:
 258                 partner_id = params['p'][0]
 259             elif 'partner_id' in params:
 260                 partner_id = params['partner_id'][0]
 261             else:
 262                 raise ExtractorError('Invalid URL', expected=True)
 263             if 'entry_id' in params:
 264                 entry_id = params['entry_id'][0]
 265                 _, info, flavor_assets, captions = self._get_video_info(entry_id, partner_id)
 266             elif 'uiconf_id' in params and 'flashvars[referenceId]' in params:
 267                 reference_id = params['flashvars[referenceId]'][0]
 268                 webpage = self._download_webpage(url, reference_id)
 269                 entry_data = self._parse_json(self._search_regex(
 270                     r'window\.kalturaIframePackageData\s*=\s*({.*});',
 271                     webpage, 'kalturaIframePackageData'),
 272                     reference_id)['entryResult']
 273                 info, flavor_assets = entry_data['meta'], entry_data['contextData']['flavorAssets']
 274                 entry_id = info['id']
 275                 # Unfortunately, data returned in kalturaIframePackageData lacks
 276                 # captions so we will try requesting the complete data using
 277                 # regular approach since we now know the entry_id
 278                 try:
 279                     _, info, flavor_assets, captions = self._get_video_info(
 280                         entry_id, partner_id)
 281                 except ExtractorError:
 282                     # Regular scenario failed but we already have everything
 283                     # extracted apart from captions and can process at least
 284                     # with this
 285                     pass
 286             else:
 287                 raise ExtractorError('Invalid URL', expected=True)
 288             ks = params.get('flashvars[ks]', [None])[0]
 289
 290         source_url = smuggled_data.get('source_url')
 291         if source_url:
 292             referrer = base64.b64encode(
 293                 '://'.join(compat_urlparse.urlparse(source_url)[:2])
 294                 .encode('utf-8')).decode('utf-8')
 295         else:
 296             referrer = None
 297
 298         def sign_url(unsigned_url):
 299             if ks:
 300                 unsigned_url += '/ks/%s' % ks
 301             if referrer:
 302                 unsigned_url += '?referrer=%s' % referrer
 303             return unsigned_url
 304
 305         data_url = info['dataUrl']
 306         if '/flvclipper/' in data_url:
 307             data_url = re.sub(r'/flvclipper/.*', '/serveFlavor', data_url)
 308
 309         formats = []
 310         subtitles = {}
 311         for f in flavor_assets:
 312             # Continue if asset is not ready
 313             if f.get('status') != 2:
 314                 continue
 315             # Original format that's not available (e.g. kaltura:1926081:0_c03e1b5g)
 316             # skip for now.
 317             if f.get('fileExt') == 'chun':
 318                 continue
 319             # DRM-protected video, cannot be decrypted
 320             if not self.get_param('allow_unplayable_formats') and f.get('fileExt') == 'wvm':
 321                 continue
 322             if not f.get('fileExt'):
 323                 # QT indicates QuickTime; some videos have broken fileExt
 324                 if f.get('containerFormat') == 'qt':
 325                     f['fileExt'] = 'mov'
 326                 else:
 327                     f['fileExt'] = 'mp4'
 328             video_url = sign_url(
 329                 '%s/flavorId/%s' % (data_url, f['id']))
 330             format_id = '%(fileExt)s-%(bitrate)s' % f
 331             # Source format may not be available (e.g. kaltura:513551:1_66x4rg7o)
 332             if f.get('isOriginal') is True and not self._is_valid_url(
 333                     video_url, entry_id, format_id):
 334                 continue
 335             # audio-only has no videoCodecId (e.g. kaltura:1926081:0_c03e1b5g
 336             # -f mp4-56)
 337             vcodec = 'none' if 'videoCodecId' not in f and f.get(
 338                 'frameRate') == 0 else f.get('videoCodecId')
 339             formats.append({
 340                 'format_id': format_id,
 341                 'ext': f.get('fileExt'),
 342                 'tbr': int_or_none(f['bitrate']),
 343                 'fps': int_or_none(f.get('frameRate')),
 344                 'filesize_approx': int_or_none(f.get('size'), invscale=1024),
 345                 'container': f.get('containerFormat'),
 346                 'vcodec': vcodec,
 347                 'height': int_or_none(f.get('height')),
 348                 'width': int_or_none(f.get('width')),
 349                 'url': video_url,
 350             })
 351         if '/playManifest/' in data_url:
 352             m3u8_url = sign_url(data_url.replace(
 353                 'format/url', 'format/applehttp'))
 354             fmts, subs = self._extract_m3u8_formats_and_subtitles(
 355                 m3u8_url, entry_id, 'mp4', 'm3u8_native',
 356                 m3u8_id='hls', fatal=False)
 357             formats.extend(fmts)
 358             self._merge_subtitles(subs, target=subtitles)
 359
 360         self._sort_formats(formats)
 361
 362         if captions:
 363             for caption in captions.get('objects', []):
 364                 # Continue if caption is not ready
 365                 if caption.get('status') != 2:
 366                     continue
 367                 if not caption.get('id'):
 368                     continue
 369                 caption_format = int_or_none(caption.get('format'))
 370                 subtitles.setdefault(caption.get('languageCode') or caption.get('language'), []).append({
 371                     'url': '%s/api_v3/service/caption_captionasset/action/serve/captionAssetId/%s' % (self._SERVICE_URL, caption['id']),
 372                     'ext': caption.get('fileExt') or self._CAPTION_TYPES.get(caption_format) or 'ttml',
 373                 })
 374
 375         return {
 376             'id': entry_id,
 377             'title': info['name'],
 378             'formats': formats,
 379             'subtitles': subtitles,
 380             'description': clean_html(info.get('description')),
 381             'thumbnail': info.get('thumbnailUrl'),
 382             'duration': info.get('duration'),
 383             'timestamp': info.get('createdAt'),
 384             'uploader_id': format_field(info, 'userId', ignore=('None', None)),
 385             'view_count': int_or_none(info.get('plays')),
 386         }