Raphaël G. Git Repositories - youtubedl/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     # 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             'params': {
  95                 'skip_download': True,
  96             },
  97         }
  98     ]
  99
 100     @staticmethod
 101     def _extract_url(webpage):
 102         mobj = (
 103             re.search(
 104                 r"""(?xs)
 105                     kWidget\.(?:thumb)?[Ee]mbed\(
 106                     \{.*?
 107                         (?P<q1>['\"])wid(?P=q1)\s*:\s*
 108                         (?P<q2>['\"])_?(?P<partner_id>(?:(?!(?P=q2)).)+)(?P=q2),.*?
 109                         (?P<q3>['\"])entry_?[Ii]d(?P=q3)\s*:\s*
 110                         (?P<q4>['\"])(?P<id>(?:(?!(?P=q4)).)+)(?P=q4)(?:,|\s*\})
 111                 """, webpage) or
 112             re.search(
 113                 r'''(?xs)
 114                     (?P<q1>["\'])
 115                         (?:https?:)?//cdnapi(?:sec)?\.kaltura\.com/(?:(?!(?P=q1)).)*(?:p|partner_id)/(?P<partner_id>\d+)(?:(?!(?P=q1)).)*
 116                     (?P=q1).*?
 117                     (?:
 118                         entry_?[Ii]d|
 119                         (?P<q2>["\'])entry_?[Ii]d(?P=q2)
 120                     )\s*:\s*
 121                     (?P<q3>["\'])(?P<id>(?:(?!(?P=q3)).)+)(?P=q3)
 122                 ''', webpage))
 123         if mobj:
 124             embed_info = mobj.groupdict()
 125             url = 'kaltura:%(partner_id)s:%(id)s' % embed_info
 126             escaped_pid = re.escape(embed_info['partner_id'])
 127             service_url = re.search(
 128                 r'<script[^>]+src=["\']((?:https?:)?//.+?)/p/%s/sp/%s00/embedIframeJs' % (escaped_pid, escaped_pid),
 129                 webpage)
 130             if service_url:
 131                 url = smuggle_url(url, {'service_url': service_url.group(1)})
 132             return url
 133
 134     def _kaltura_api_call(self, video_id, actions, service_url=None, *args, **kwargs):
 135         params = actions[0]
 136         if len(actions) > 1:
 137             for i, a in enumerate(actions[1:], start=1):
 138                 for k, v in a.items():
 139                     params['%d:%s' % (i, k)] = v
 140
 141         data = self._download_json(
 142             (service_url or self._SERVICE_URL) + self._SERVICE_BASE,
 143             video_id, query=params, *args, **kwargs)
 144
 145         status = data if len(actions) == 1 else data[0]
 146         if status.get('objectType') == 'KalturaAPIException':
 147             raise ExtractorError(
 148                 '%s said: %s' % (self.IE_NAME, status['message']))
 149
 150         return data
 151
 152     def _get_video_info(self, video_id, partner_id, service_url=None):
 153         actions = [
 154             {
 155                 'action': 'null',
 156                 'apiVersion': '3.1.5',
 157                 'clientTag': 'kdp:v3.8.5',
 158                 'format': 1,  # JSON, 2 = XML, 3 = PHP
 159                 'service': 'multirequest',
 160             },
 161             {
 162                 'expiry': 86400,
 163                 'service': 'session',
 164                 'action': 'startWidgetSession',
 165                 'widgetId': '_%s' % partner_id,
 166             },
 167             {
 168                 'action': 'get',
 169                 'entryId': video_id,
 170                 'service': 'baseentry',
 171                 'ks': '{1:result:ks}',
 172             },
 173             {
 174                 'action': 'getbyentryid',
 175                 'entryId': video_id,
 176                 'service': 'flavorAsset',
 177                 'ks': '{1:result:ks}',
 178             },
 179             {
 180                 'action': 'list',
 181                 'filter:entryIdEqual': video_id,
 182                 'service': 'caption_captionasset',
 183                 'ks': '{1:result:ks}',
 184             },
 185         ]
 186         return self._kaltura_api_call(
 187             video_id, actions, service_url, note='Downloading video info JSON')
 188
 189     def _real_extract(self, url):
 190         url, smuggled_data = unsmuggle_url(url, {})
 191
 192         mobj = re.match(self._VALID_URL, url)
 193         partner_id, entry_id = mobj.group('partner_id', 'id')
 194         ks = None
 195         captions = None
 196         if partner_id and entry_id:
 197             _, info, flavor_assets, captions = self._get_video_info(entry_id, partner_id, smuggled_data.get('service_url'))
 198         else:
 199             path, query = mobj.group('path', 'query')
 200             if not path and not query:
 201                 raise ExtractorError('Invalid URL', expected=True)
 202             params = {}
 203             if query:
 204                 params = compat_parse_qs(query)
 205             if path:
 206                 splitted_path = path.split('/')
 207                 params.update(dict((zip(splitted_path[::2], [[v] for v in splitted_path[1::2]]))))
 208             if 'wid' in params:
 209                 partner_id = params['wid'][0][1:]
 210             elif 'p' in params:
 211                 partner_id = params['p'][0]
 212             else:
 213                 raise ExtractorError('Invalid URL', expected=True)
 214             if 'entry_id' in params:
 215                 entry_id = params['entry_id'][0]
 216                 _, info, flavor_assets, captions = self._get_video_info(entry_id, partner_id)
 217             elif 'uiconf_id' in params and 'flashvars[referenceId]' in params:
 218                 reference_id = params['flashvars[referenceId]'][0]
 219                 webpage = self._download_webpage(url, reference_id)
 220                 entry_data = self._parse_json(self._search_regex(
 221                     r'window\.kalturaIframePackageData\s*=\s*({.*});',
 222                     webpage, 'kalturaIframePackageData'),
 223                     reference_id)['entryResult']
 224                 info, flavor_assets = entry_data['meta'], entry_data['contextData']['flavorAssets']
 225                 entry_id = info['id']
 226                 # Unfortunately, data returned in kalturaIframePackageData lacks
 227                 # captions so we will try requesting the complete data using
 228                 # regular approach since we now know the entry_id
 229                 try:
 230                     _, info, flavor_assets, captions = self._get_video_info(
 231                         entry_id, partner_id)
 232                 except ExtractorError:
 233                     # Regular scenario failed but we already have everything
 234                     # extracted apart from captions and can process at least
 235                     # with this
 236                     pass
 237             else:
 238                 raise ExtractorError('Invalid URL', expected=True)
 239             ks = params.get('flashvars[ks]', [None])[0]
 240
 241         source_url = smuggled_data.get('source_url')
 242         if source_url:
 243             referrer = base64.b64encode(
 244                 '://'.join(compat_urlparse.urlparse(source_url)[:2])
 245                 .encode('utf-8')).decode('utf-8')
 246         else:
 247             referrer = None
 248
 249         def sign_url(unsigned_url):
 250             if ks:
 251                 unsigned_url += '/ks/%s' % ks
 252             if referrer:
 253                 unsigned_url += '?referrer=%s' % referrer
 254             return unsigned_url
 255
 256         data_url = info['dataUrl']
 257         if '/flvclipper/' in data_url:
 258             data_url = re.sub(r'/flvclipper/.*', '/serveFlavor', data_url)
 259
 260         formats = []
 261         for f in flavor_assets:
 262             # Continue if asset is not ready
 263             if f.get('status') != 2:
 264                 continue
 265             # Original format that's not available (e.g. kaltura:1926081:0_c03e1b5g)
 266             # skip for now.
 267             if f.get('fileExt') == 'chun':
 268                 continue
 269             if not f.get('fileExt'):
 270                 # QT indicates QuickTime; some videos have broken fileExt
 271                 if f.get('containerFormat') == 'qt':
 272                     f['fileExt'] = 'mov'
 273                 else:
 274                     f['fileExt'] = 'mp4'
 275             video_url = sign_url(
 276                 '%s/flavorId/%s' % (data_url, f['id']))
 277             # audio-only has no videoCodecId (e.g. kaltura:1926081:0_c03e1b5g
 278             # -f mp4-56)
 279             vcodec = 'none' if 'videoCodecId' not in f and f.get(
 280                 'frameRate') == 0 else f.get('videoCodecId')
 281             formats.append({
 282                 'format_id': '%(fileExt)s-%(bitrate)s' % f,
 283                 'ext': f.get('fileExt'),
 284                 'tbr': int_or_none(f['bitrate']),
 285                 'fps': int_or_none(f.get('frameRate')),
 286                 'filesize_approx': int_or_none(f.get('size'), invscale=1024),
 287                 'container': f.get('containerFormat'),
 288                 'vcodec': vcodec,
 289                 'height': int_or_none(f.get('height')),
 290                 'width': int_or_none(f.get('width')),
 291                 'url': video_url,
 292             })
 293         if '/playManifest/' in data_url:
 294             m3u8_url = sign_url(data_url.replace(
 295                 'format/url', 'format/applehttp'))
 296             formats.extend(self._extract_m3u8_formats(
 297                 m3u8_url, entry_id, 'mp4', 'm3u8_native',
 298                 m3u8_id='hls', fatal=False))
 299
 300         self._sort_formats(formats)
 301
 302         subtitles = {}
 303         if captions:
 304             for caption in captions.get('objects', []):
 305                 # Continue if caption is not ready
 306                 if f.get('status') != 2:
 307                     continue
 308                 if not caption.get('id'):
 309                     continue
 310                 caption_format = int_or_none(caption.get('format'))
 311                 subtitles.setdefault(caption.get('languageCode') or caption.get('language'), []).append({
 312                     'url': '%s/api_v3/service/caption_captionasset/action/serve/captionAssetId/%s' % (self._SERVICE_URL, caption['id']),
 313                     'ext': caption.get('fileExt') or self._CAPTION_TYPES.get(caption_format) or 'ttml',
 314                 })
 315
 316         return {
 317             'id': entry_id,
 318             'title': info['name'],
 319             'formats': formats,
 320             'subtitles': subtitles,
 321             'description': clean_html(info.get('description')),
 322             'thumbnail': info.get('thumbnailUrl'),
 323             'duration': info.get('duration'),
 324             'timestamp': info.get('createdAt'),
 325             'uploader_id': info.get('userId') if info.get('userId') != 'None' else None,
 326             'view_count': info.get('plays'),
 327         }