Merge remote-tracking branch 'source/master' into linkedin-learning-subtitles

youtube_dl/extractor/ard.py

@@ -1,6 +1,7 @@
 # coding: utf-8
 from __future__ import unicode_literals
 
+import json
 import re
 
 from .common import InfoExtractor
@@ -22,7 +23,101 @@ from ..utils import (
 from ..compat import compat_etree_fromstring
 
 
-class ARDMediathekIE(InfoExtractor):
+class ARDMediathekBaseIE(InfoExtractor):
+    _GEO_COUNTRIES = ['DE']
+
+    def _extract_media_info(self, media_info_url, webpage, video_id):
+        media_info = self._download_json(
+            media_info_url, video_id, 'Downloading media JSON')
+        return self._parse_media_info(media_info, video_id, '"fsk"' in webpage)
+
+    def _parse_media_info(self, media_info, video_id, fsk):
+        formats = self._extract_formats(media_info, video_id)
+
+        if not formats:
+            if fsk:
+                raise ExtractorError(
+                    'This video is only available after 20:00', expected=True)
+            elif media_info.get('_geoblocked'):
+                self.raise_geo_restricted(
+                    'This video is not available due to geoblocking',
+                    countries=self._GEO_COUNTRIES)
+
+        self._sort_formats(formats)
+
+        subtitles = {}
+        subtitle_url = media_info.get('_subtitleUrl')
+        if subtitle_url:
+            subtitles['de'] = [{
+                'ext': 'ttml',
+                'url': subtitle_url,
+            }]
+
+        return {
+            'id': video_id,
+            'duration': int_or_none(media_info.get('_duration')),
+            'thumbnail': media_info.get('_previewImage'),
+            'is_live': media_info.get('_isLive') is True,
+            'formats': formats,
+            'subtitles': subtitles,
+        }
+
+    def _extract_formats(self, media_info, video_id):
+        type_ = media_info.get('_type')
+        media_array = media_info.get('_mediaArray', [])
+        formats = []
+        for num, media in enumerate(media_array):
+            for stream in media.get('_mediaStreamArray', []):
+                stream_urls = stream.get('_stream')
+                if not stream_urls:
+                    continue
+                if not isinstance(stream_urls, list):
+                    stream_urls = [stream_urls]
+                quality = stream.get('_quality')
+                server = stream.get('_server')
+                for stream_url in stream_urls:
+                    if not url_or_none(stream_url):
+                        continue
+                    ext = determine_ext(stream_url)
+                    if quality != 'auto' and ext in ('f4m', 'm3u8'):
+                        continue
+                    if ext == 'f4m':
+                        formats.extend(self._extract_f4m_formats(
+                            update_url_query(stream_url, {
+                                'hdcore': '3.1.1',
+                                'plugin': 'aasp-3.1.1.69.124'
+                            }), video_id, f4m_id='hds', fatal=False))
+                    elif ext == 'm3u8':
+                        formats.extend(self._extract_m3u8_formats(
+                            stream_url, video_id, 'mp4', 'm3u8_native',
+                            m3u8_id='hls', fatal=False))
+                    else:
+                        if server and server.startswith('rtmp'):
+                            f = {
+                                'url': server,
+                                'play_path': stream_url,
+                                'format_id': 'a%s-rtmp-%s' % (num, quality),
+                            }
+                        else:
+                            f = {
+                                'url': stream_url,
+                                'format_id': 'a%s-%s-%s' % (num, ext, quality)
+                            }
+                        m = re.search(
+                            r'_(?P<width>\d+)x(?P<height>\d+)\.mp4$',
+                            stream_url)
+                        if m:
+                            f.update({
+                                'width': int(m.group('width')),
+                                'height': int(m.group('height')),
+                            })
+                        if type_ == 'audio':
+                            f['vcodec'] = 'none'
+                        formats.append(f)
+        return formats
+
+
+class ARDMediathekIE(ARDMediathekBaseIE):
     IE_NAME = 'ARD:mediathek'
     _VALID_URL = r'^https?://(?:(?:(?:www|classic)\.)?ardmediathek\.de|mediathek\.(?:daserste|rbb-online)\.de|one\.ard\.de)/(?:.*/)(?P<video_id>[0-9]+|[^0-9][^/\?]+)[^/\?]*(?:\?.*)?'
 
@@ -63,94 +158,6 @@ class ARDMediathekIE(InfoExtractor):
     def suitable(cls, url):
         return False if ARDBetaMediathekIE.suitable(url) else super(ARDMediathekIE, cls).suitable(url)
 
-    def _extract_media_info(self, media_info_url, webpage, video_id):
-        media_info = self._download_json(
-            media_info_url, video_id, 'Downloading media JSON')
-
-        formats = self._extract_formats(media_info, video_id)
-
-        if not formats:
-            if '"fsk"' in webpage:
-                raise ExtractorError(
-                    'This video is only available after 20:00', expected=True)
-            elif media_info.get('_geoblocked'):
-                raise ExtractorError('This video is not available due to geo restriction', expected=True)
-
-        self._sort_formats(formats)
-
-        duration = int_or_none(media_info.get('_duration'))
-        thumbnail = media_info.get('_previewImage')
-        is_live = media_info.get('_isLive') is True
-
-        subtitles = {}
-        subtitle_url = media_info.get('_subtitleUrl')
-        if subtitle_url:
-            subtitles['de'] = [{
-                'ext': 'ttml',
-                'url': subtitle_url,
-            }]
-
-        return {
-            'id': video_id,
-            'duration': duration,
-            'thumbnail': thumbnail,
-            'is_live': is_live,
-            'formats': formats,
-            'subtitles': subtitles,
-        }
-
-    def _extract_formats(self, media_info, video_id):
-        type_ = media_info.get('_type')
-        media_array = media_info.get('_mediaArray', [])
-        formats = []
-        for num, media in enumerate(media_array):
-            for stream in media.get('_mediaStreamArray', []):
-                stream_urls = stream.get('_stream')
-                if not stream_urls:
-                    continue
-                if not isinstance(stream_urls, list):
-                    stream_urls = [stream_urls]
-                quality = stream.get('_quality')
-                server = stream.get('_server')
-                for stream_url in stream_urls:
-                    if not url_or_none(stream_url):
-                        continue
-                    ext = determine_ext(stream_url)
-                    if quality != 'auto' and ext in ('f4m', 'm3u8'):
-                        continue
-                    if ext == 'f4m':
-                        formats.extend(self._extract_f4m_formats(
-                            update_url_query(stream_url, {
-                                'hdcore': '3.1.1',
-                                'plugin': 'aasp-3.1.1.69.124'
-                            }),
-                            video_id, f4m_id='hds', fatal=False))
-                    elif ext == 'm3u8':
-                        formats.extend(self._extract_m3u8_formats(
-                            stream_url, video_id, 'mp4', m3u8_id='hls', fatal=False))
-                    else:
-                        if server and server.startswith('rtmp'):
-                            f = {
-                                'url': server,
-                                'play_path': stream_url,
-                                'format_id': 'a%s-rtmp-%s' % (num, quality),
-                            }
-                        else:
-                            f = {
-                                'url': stream_url,
-                                'format_id': 'a%s-%s-%s' % (num, ext, quality)
-                            }
-                        m = re.search(r'_(?P<width>\d+)x(?P<height>\d+)\.mp4$', stream_url)
-                        if m:
-                            f.update({
-                                'width': int(m.group('width')),
-                                'height': int(m.group('height')),
-                            })
-                        if type_ == 'audio':
-                            f['vcodec'] = 'none'
-                        formats.append(f)
-        return formats
-
     def _real_extract(self, url):
         # determine video id from url
         m = re.match(self._VALID_URL, url)
@@ -302,19 +309,20 @@ class ARDIE(InfoExtractor):
         }
 
 
-class ARDBetaMediathekIE(InfoExtractor):
+class ARDBetaMediathekIE(ARDMediathekBaseIE):
-    _VALID_URL = r'https://(?:beta|www)\.ardmediathek\.de/[^/]+/(?:player|live)/(?P<video_id>[a-zA-Z0-9]+)(?:/(?P<display_id>[^/?#]+))?'
+    _VALID_URL = r'https://(?:beta|www)\.ardmediathek\.de/(?P<client>[^/]+)/(?:player|live)/(?P<video_id>[a-zA-Z0-9]+)(?:/(?P<display_id>[^/?#]+))?'
     _TESTS = [{
         'url': 'https://beta.ardmediathek.de/ard/player/Y3JpZDovL2Rhc2Vyc3RlLmRlL3RhdG9ydC9mYmM4NGM1NC0xNzU4LTRmZGYtYWFhZS0wYzcyZTIxNGEyMDE/die-robuste-roswita',
-        'md5': '2d02d996156ea3c397cfc5036b5d7f8f',
+        'md5': 'dfdc87d2e7e09d073d5a80770a9ce88f',
         'info_dict': {
             'display_id': 'die-robuste-roswita',
-            'id': 'Y3JpZDovL2Rhc2Vyc3RlLmRlL3RhdG9ydC9mYmM4NGM1NC0xNzU4LTRmZGYtYWFhZS0wYzcyZTIxNGEyMDE',
+            'id': '70153354',
-            'title': 'Tatort: Die robuste Roswita',
+            'title': 'Die robuste Roswita',
             'description': r're:^Der Mord.*trüber ist als die Ilm.',
             'duration': 5316,
-            'thumbnail': 'https://img.ardmediathek.de/standard/00/55/43/59/34/-1774185891/16x9/960?mandant=ard',
+            'thumbnail': 'https://img.ardmediathek.de/standard/00/70/15/33/90/-1852531467/16x9/960?mandant=ard',
-            'upload_date': '20180826',
+            'timestamp': 1577047500,
+            'upload_date': '20191222',
             'ext': 'mp4',
         },
     }, {
@@ -330,71 +338,69 @@ class ARDBetaMediathekIE(InfoExtractor):
         video_id = mobj.group('video_id')
         display_id = mobj.group('display_id') or video_id
 
-        webpage = self._download_webpage(url, display_id)
+        player_page = self._download_json(
-        data_json = self._search_regex(r'window\.__APOLLO_STATE__\s*=\s*(\{.*);\n', webpage, 'json')
+            'https://api.ardmediathek.de/public-gateway',
-        data = self._parse_json(data_json, display_id)
+            display_id, data=json.dumps({
-
+                'query': '''{
-        res = {
+  playerPage(client:"%s", clipId: "%s") {
-            'id': video_id,
+    blockedByFsk
-            'display_id': display_id,
+    broadcastedOn
+    maturityContentRating
+    mediaCollection {
+      _duration
+      _geoblocked
+      _isLive
+      _mediaArray {
+        _mediaStreamArray {
+          _quality
+          _server
+          _stream
         }
-        formats = []
+      }
-        subtitles = {}
+      _previewImage
-        geoblocked = False
+      _subtitleUrl
-        for widget in data.values():
+      _type
-            if widget.get('_geoblocked') is True:
+    }
-                geoblocked = True
+    show {
-            if '_duration' in widget:
+      title
-                res['duration'] = int_or_none(widget['_duration'])
+    }
-            if 'clipTitle' in widget:
+    synopsis
-                res['title'] = widget['clipTitle']
+    title
-            if '_previewImage' in widget:
+    tracking {
-                res['thumbnail'] = widget['_previewImage']
+      atiCustomVars {
-            if 'broadcastedOn' in widget:
+        contentId
-                res['timestamp'] = unified_timestamp(widget['broadcastedOn'])
+      }
-            if 'synopsis' in widget:
+    }
-                res['description'] = widget['synopsis']
+  }
-            subtitle_url = url_or_none(widget.get('_subtitleUrl'))
+}''' % (mobj.group('client'), video_id),
-            if subtitle_url:
+            }).encode(), headers={
-                subtitles.setdefault('de', []).append({
+                'Content-Type': 'application/json'
-                    'ext': 'ttml',
+            })['data']['playerPage']
-                    'url': subtitle_url,
+        title = player_page['title']
+        content_id = str_or_none(try_get(
+            player_page, lambda x: x['tracking']['atiCustomVars']['contentId']))
+        media_collection = player_page.get('mediaCollection') or {}
+        if not media_collection and content_id:
+            media_collection = self._download_json(
+                'https://www.ardmediathek.de/play/media/' + content_id,
+                content_id, fatal=False) or {}
+        info = self._parse_media_info(
+            media_collection, content_id or video_id,
+            player_page.get('blockedByFsk'))
+        age_limit = None
+        description = player_page.get('synopsis')
+        maturity_content_rating = player_page.get('maturityContentRating')
+        if maturity_content_rating:
+            age_limit = int_or_none(maturity_content_rating.lstrip('FSK'))
+        if not age_limit and description:
+            age_limit = int_or_none(self._search_regex(
+                r'\(FSK\s*(\d+)\)\s*$', description, 'age limit', default=None))
+        info.update({
+            'age_limit': age_limit,
+            'display_id': display_id,
+            'title': title,
+            'description': description,
+            'timestamp': unified_timestamp(player_page.get('broadcastedOn')),
+            'series': try_get(player_page, lambda x: x['show']['title']),
         })
-            if '_quality' in widget:
+        return info
-                format_url = url_or_none(try_get(
-                    widget, lambda x: x['_stream']['json'][0]))
-                if not format_url:
-                    continue
-                ext = determine_ext(format_url)
-                if ext == 'f4m':
-                    formats.extend(self._extract_f4m_formats(
-                        format_url + '?hdcore=3.11.0',
-                        video_id, f4m_id='hds', fatal=False))
-                elif ext == 'm3u8':
-                    formats.extend(self._extract_m3u8_formats(
-                        format_url, video_id, 'mp4', m3u8_id='hls',
-                        fatal=False))
-                else:
-                    # HTTP formats are not available when geoblocked is True,
-                    # other formats are fine though
-                    if geoblocked:
-                        continue
-                    quality = str_or_none(widget.get('_quality'))
-                    formats.append({
-                        'format_id': ('http-' + quality) if quality else 'http',
-                        'url': format_url,
-                        'preference': 10,  # Plain HTTP, that's nice
-                    })
-
-        if not formats and geoblocked:
-            self.raise_geo_restricted(
-                msg='This video is not available due to geoblocking',
-                countries=['DE'])
-
-        self._sort_formats(formats)
-        res.update({
-            'subtitles': subtitles,
-            'formats': formats,
-        })
-
-        return res

youtube_dl/extractor/azmedien.py

@@ -47,39 +47,19 @@ class AZMedienIE(InfoExtractor):
         'url': 'https://www.telebaern.tv/telebaern-news/montag-1-oktober-2018-ganze-sendung-133531189#video=0_7xjo9lf1',
         'only_matching': True
     }]
-
+    _API_TEMPL = 'https://www.%s/api/pub/gql/%s/NewsArticleTeaser/cb9f2f81ed22e9b47f4ca64ea3cc5a5d13e88d1d'
     _PARTNER_ID = '1719221'
 
     def _real_extract(self, url):
-        mobj = re.match(self._VALID_URL, url)
+        host, display_id, article_id, entry_id = re.match(self._VALID_URL, url).groups()
-        host = mobj.group('host')
-        video_id = mobj.group('id')
-        entry_id = mobj.group('kaltura_id')
 
         if not entry_id:
-            api_url = 'https://www.%s/api/pub/gql/%s' % (host, host.split('.')[0])
+            entry_id = self._download_json(
-            payload = {
+                self._API_TEMPL % (host, host.split('.')[0]), display_id, query={
-                'query': '''query VideoContext($articleId: ID!) {
+                    'variables': json.dumps({
-                    article: node(id: $articleId) {
+                        'contextId': 'NewsArticle:' + article_id,
-                      ... on Article {
+                    }),
-                        mainAssetRelation {
+                })['data']['context']['mainAsset']['video']['kaltura']['kalturaId']
-                          asset {
-                            ... on VideoAsset {
-                              kalturaId
-                            }
-                          }
-                        }
-                      }
-                    }
-                  }''',
-                'variables': {'articleId': 'Article:%s' % mobj.group('article_id')},
-            }
-            json_data = self._download_json(
-                api_url, video_id, headers={
-                    'Content-Type': 'application/json',
-                },
-                data=json.dumps(payload).encode())
-            entry_id = json_data['data']['article']['mainAssetRelation']['asset']['kalturaId']
 
         return self.url_result(
             'kaltura:%s:%s' % (self._PARTNER_ID, entry_id),

youtube_dl/extractor/businessinsider.py

@@ -9,21 +9,26 @@ class BusinessInsiderIE(InfoExtractor):
     _VALID_URL = r'https?://(?:[^/]+\.)?businessinsider\.(?:com|nl)/(?:[^/]+/)*(?P<id>[^/?#&]+)'
     _TESTS = [{
         'url': 'http://uk.businessinsider.com/how-much-radiation-youre-exposed-to-in-everyday-life-2016-6',
-        'md5': 'ca237a53a8eb20b6dc5bd60564d4ab3e',
+        'md5': 'ffed3e1e12a6f950aa2f7d83851b497a',
         'info_dict': {
-            'id': 'hZRllCfw',
+            'id': 'cjGDb0X9',
             'ext': 'mp4',
-            'title': "Here's how much radiation you're exposed to in everyday life",
+            'title': "Bananas give you more radiation exposure than living next to a nuclear power plant",
-            'description': 'md5:9a0d6e2c279948aadaa5e84d6d9b99bd',
+            'description': 'md5:0175a3baf200dd8fa658f94cade841b3',
-            'upload_date': '20170709',
+            'upload_date': '20160611',
-            'timestamp': 1499606400,
+            'timestamp': 1465675620,
-        },
-        'params': {
-            'skip_download': True,
         },
     }, {
         'url': 'https://www.businessinsider.nl/5-scientifically-proven-things-make-you-less-attractive-2017-7/',
-        'only_matching': True,
+        'md5': '43f438dbc6da0b89f5ac42f68529d84a',
+        'info_dict': {
+            'id': '5zJwd4FK',
+            'ext': 'mp4',
+            'title': 'Deze dingen zorgen ervoor dat je minder snel een date scoort',
+            'description': 'md5:2af8975825d38a4fed24717bbe51db49',
+            'upload_date': '20170705',
+            'timestamp': 1499270528,
+        },
     }, {
         'url': 'http://www.businessinsider.com/excel-index-match-vlookup-video-how-to-2015-2?IR=T',
         'only_matching': True,
@@ -35,7 +40,8 @@ class BusinessInsiderIE(InfoExtractor):
         jwplatform_id = self._search_regex(
             (r'data-media-id=["\']([a-zA-Z0-9]{8})',
              r'id=["\']jwplayer_([a-zA-Z0-9]{8})',
-             r'id["\']?\s*:\s*["\']?([a-zA-Z0-9]{8})'),
+             r'id["\']?\s*:\s*["\']?([a-zA-Z0-9]{8})',
+             r'(?:jwplatform\.com/players/|jwplayer_)([a-zA-Z0-9]{8})'),
             webpage, 'jwplatform id')
         return self.url_result(
             'jwplatform:%s' % jwplatform_id, ie=JWPlatformIE.ie_key(),

youtube_dl/extractor/ivi.py

@@ -239,7 +239,7 @@ class IviCompilationIE(InfoExtractor):
             self.url_result(
                 'http://www.ivi.ru/watch/%s/%s' % (compilation_id, serie), IviIE.ie_key())
             for serie in re.findall(
-                r'<a href="/watch/%s/(\d+)"[^>]+data-id="\1"' % compilation_id, html)]
+                r'<a\b[^>]+\bhref=["\']/watch/%s/(\d+)["\']' % compilation_id, html)]
 
     def _real_extract(self, url):
         mobj = re.match(self._VALID_URL, url)

youtube_dl/extractor/srmediathek.py

@@ -1,14 +1,14 @@
 # coding: utf-8
 from __future__ import unicode_literals
 
-from .ard import ARDMediathekIE
+from .ard import ARDMediathekBaseIE
 from ..utils import (
     ExtractorError,
     get_element_by_attribute,
 )
 
 
-class SRMediathekIE(ARDMediathekIE):
+class SRMediathekIE(ARDMediathekBaseIE):
     IE_NAME = 'sr:mediathek'
     IE_DESC = 'Saarländischer Rundfunk'
     _VALID_URL = r'https?://sr-mediathek(?:\.sr-online)?\.de/index\.php\?.*?&id=(?P<id>[0-9]+)'

youtube_dl/extractor/stretchinternet.py

@@ -5,44 +5,28 @@ from ..utils import int_or_none
 
 
 class StretchInternetIE(InfoExtractor):
-    _VALID_URL = r'https?://portal\.stretchinternet\.com/[^/]+/portal\.htm\?.*?\beventId=(?P<id>\d+)'
+    _VALID_URL = r'https?://portal\.stretchinternet\.com/[^/]+/(?:portal|full)\.htm\?.*?\beventId=(?P<id>\d+)'
     _TEST = {
-        'url': 'https://portal.stretchinternet.com/umary/portal.htm?eventId=313900&streamType=video',
+        'url': 'https://portal.stretchinternet.com/umary/portal.htm?eventId=573272&streamType=video',
         'info_dict': {
-            'id': '313900',
+            'id': '573272',
             'ext': 'mp4',
-            'title': 'Augustana (S.D.) Baseball vs University of Mary',
+            'title': 'University of Mary Wrestling vs. Upper Iowa',
-            'description': 'md5:7578478614aae3bdd4a90f578f787438',
+            'timestamp': 1575668361,
-            'timestamp': 1490468400,
+            'upload_date': '20191206',
-            'upload_date': '20170325',
         }
     }
 
     def _real_extract(self, url):
         video_id = self._match_id(url)
 
-        stream = self._download_json(
-            'https://neo-client.stretchinternet.com/streamservice/v1/media/stream/v%s'
-            % video_id, video_id)
-
-        video_url = 'https://%s' % stream['source']
-
         event = self._download_json(
-            'https://neo-client.stretchinternet.com/portal-ws/getEvent.json',
+            'https://api.stretchinternet.com/trinity/event/tcg/' + video_id,
-            video_id, query={
+            video_id)[0]
-                'clientID': 99997,
-                'eventID': video_id,
-                'token': 'asdf',
-            })['event']
-
-        title = event.get('title') or event['mobileTitle']
-        description = event.get('customText')
-        timestamp = int_or_none(event.get('longtime'))
 
         return {
             'id': video_id,
-            'title': title,
+            'title': event['title'],
-            'description': description,
+            'timestamp': int_or_none(event.get('dateCreated'), 1000),
-            'timestamp': timestamp,
+            'url': 'https://' + event['media'][0]['url'],
-            'url': video_url,
         }

youtube_dl/extractor/twentyfourvideo.py

@@ -18,7 +18,7 @@ class TwentyFourVideoIE(InfoExtractor):
                     https?://
                         (?P<host>
                             (?:(?:www|porno)\.)?24video\.
-                            (?:net|me|xxx|sexy?|tube|adult|site)
+                            (?:net|me|xxx|sexy?|tube|adult|site|vip)
                         )/
                         (?:
                             video/(?:(?:view|xml)/)?|
@@ -59,6 +59,9 @@ class TwentyFourVideoIE(InfoExtractor):
     }, {
         'url': 'https://porno.24video.net/video/2640421-vsya-takaya-gibkaya-i-v-masle',
         'only_matching': True,
+    }, {
+        'url': 'https://www.24video.vip/video/view/1044982',
+        'only_matching': True,
     }]
 
     def _real_extract(self, url):

youtube_dl/extractor/voicerepublic.py

@@ -1,17 +1,12 @@
 from __future__ import unicode_literals
 
-import re
-
 from .common import InfoExtractor
-from ..compat import (
+from ..compat import compat_str
-    compat_str,
-    compat_urlparse,
-)
 from ..utils import (
     ExtractorError,
     determine_ext,
     int_or_none,
-    sanitized_Request,
+    urljoin,
 )
 
 
@@ -26,8 +21,7 @@ class VoiceRepublicIE(InfoExtractor):
             'ext': 'm4a',
             'title': 'Watching the Watchers: Building a Sousveillance State',
             'description': 'Secret surveillance programs have metadata too. The people and companies that operate secret surveillance programs can be surveilled.',
-            'thumbnail': r're:^https?://.*\.(?:png|jpg)$',
+            'duration': 1556,
-            'duration': 1800,
             'view_count': int,
         }
     }, {
@@ -38,63 +32,31 @@ class VoiceRepublicIE(InfoExtractor):
     def _real_extract(self, url):
         display_id = self._match_id(url)
 
-        req = sanitized_Request(
+        webpage = self._download_webpage(url, display_id)
-            compat_urlparse.urljoin(url, '/talks/%s' % display_id))
-        # Older versions of Firefox get redirected to an "upgrade browser" page
-        req.add_header('User-Agent', 'youtube-dl')
-        webpage = self._download_webpage(req, display_id)
 
         if '>Queued for processing, please stand by...<' in webpage:
             raise ExtractorError(
                 'Audio is still queued for processing', expected=True)
 
-        config = self._search_regex(
+        talk = self._parse_json(self._search_regex(
-            r'(?s)return ({.+?});\s*\n', webpage,
+            r'initialSnapshot\s*=\s*({.+?});',
-            'data', default=None)
+            webpage, 'talk'), display_id)['talk']
-        data = self._parse_json(config, display_id, fatal=False) if config else None
+        title = talk['title']
-        if data:
-            title = data['title']
-            description = data.get('teaser')
-            talk_id = compat_str(data.get('talk_id') or display_id)
-            talk = data['talk']
-            duration = int_or_none(talk.get('duration'))
         formats = [{
-                'url': compat_urlparse.urljoin(url, talk_url),
+            'url': urljoin(url, talk_url),
             'format_id': format_id,
             'ext': determine_ext(talk_url) or format_id,
             'vcodec': 'none',
-            } for format_id, talk_url in talk['links'].items()]
+        } for format_id, talk_url in talk['media_links'].items()]
-        else:
-            title = self._og_search_title(webpage)
-            description = self._html_search_regex(
-                r"(?s)<div class='talk-teaser'[^>]*>(.+?)</div>",
-                webpage, 'description', fatal=False)
-            talk_id = self._search_regex(
-                [r"id='jc-(\d+)'", r"data-shareable-id='(\d+)'"],
-                webpage, 'talk id', default=None) or display_id
-            duration = None
-            player = self._search_regex(
-                r"class='vr-player jp-jplayer'([^>]+)>", webpage, 'player')
-            formats = [{
-                'url': compat_urlparse.urljoin(url, talk_url),
-                'format_id': format_id,
-                'ext': determine_ext(talk_url) or format_id,
-                'vcodec': 'none',
-            } for format_id, talk_url in re.findall(r"data-([^=]+)='([^']+)'", player)]
         self._sort_formats(formats)
 
-        thumbnail = self._og_search_thumbnail(webpage)
-        view_count = int_or_none(self._search_regex(
-            r"class='play-count[^']*'>\s*(\d+) plays",
-            webpage, 'play count', fatal=False))
-
         return {
-            'id': talk_id,
+            'id': compat_str(talk.get('id') or display_id),
             'display_id': display_id,
             'title': title,
-            'description': description,
+            'description': talk.get('teaser'),
-            'thumbnail': thumbnail,
+            'thumbnail': talk.get('image_url'),
-            'duration': duration,
+            'duration': int_or_none(talk.get('archived_duration')),
-            'view_count': view_count,
+            'view_count': int_or_none(talk.get('play_count')),
             'formats': formats,
         }