[youtube] fix automatic captions extraction(closes #27162)(closes #27388)

youtube_dl/extractor/brightcove.py

@@ -28,6 +28,7 @@ from ..utils import (
     parse_iso8601,
     smuggle_url,
     str_or_none,
+    try_get,
     unescapeHTML,
     unsmuggle_url,
     UnsupportedError,
@@ -600,11 +601,14 @@ class BrightcoveNewIE(AdobePassIE):
         store_pk = lambda x: self._downloader.cache.store('brightcove', policy_key_id, x)
 
         def extract_policy_key():
+            base_url = 'http://players.brightcove.net/%s/%s_%s/' % (account_id, player_id, embed)
+            config = self._download_json(
+                base_url + 'config.json', video_id, fatal=False) or {}
+            policy_key = try_get(
+                config, lambda x: x['video_cloud']['policy_key'])
+            if not policy_key:
                 webpage = self._download_webpage(
-                'http://players.brightcove.net/%s/%s_%s/index.min.js'
-                % (account_id, player_id, embed), video_id)
-
-            policy_key = None
+                    base_url + 'index.min.js', video_id)
 
                 catalog = self._search_regex(
                     r'catalog\(({.+?})\);', webpage, 'catalog', default=None)

youtube_dl/extractor/cbslocal.py

@@ -11,7 +11,47 @@ from ..utils import (
 
 
 class CBSLocalIE(AnvatoIE):
-    _VALID_URL = r'https?://[a-z]+\.cbslocal\.com/(?:\d+/\d+/\d+|video)/(?P<id>[0-9a-z-]+)'
+    _VALID_URL_BASE = r'https?://[a-z]+\.cbslocal\.com/'
+    _VALID_URL = _VALID_URL_BASE + r'video/(?P<id>\d+)'
+
+    _TESTS = [{
+        'url': 'http://newyork.cbslocal.com/video/3580809-a-very-blue-anniversary/',
+        'info_dict': {
+            'id': '3580809',
+            'ext': 'mp4',
+            'title': 'A Very Blue Anniversary',
+            'description': 'CBS2’s Cindy Hsu has more.',
+            'thumbnail': 're:^https?://.*',
+            'timestamp': int,
+            'upload_date': r're:^\d{8}$',
+            'uploader': 'CBS',
+            'subtitles': {
+                'en': 'mincount:5',
+            },
+            'categories': [
+                'Stations\\Spoken Word\\WCBSTV',
+                'Syndication\\AOL',
+                'Syndication\\MSN',
+                'Syndication\\NDN',
+                'Syndication\\Yahoo',
+                'Content\\News',
+                'Content\\News\\Local News',
+            ],
+            'tags': ['CBS 2 News Weekends', 'Cindy Hsu', 'Blue Man Group'],
+        },
+        'params': {
+            'skip_download': True,
+        },
+    }]
+
+    def _real_extract(self, url):
+        mcp_id = self._match_id(url)
+        return self.url_result(
+            'anvato:anvato_cbslocal_app_web_prod_547f3e49241ef0e5d30c79b2efbca5d92c698f67:' + mcp_id, 'Anvato', mcp_id)
+
+
+class CBSLocalArticleIE(AnvatoIE):
+    _VALID_URL = CBSLocalIE._VALID_URL_BASE + r'\d+/\d+/\d+/(?P<id>[0-9a-z-]+)'
 
     _TESTS = [{
         # Anvato backend
@@ -52,31 +92,6 @@ class CBSLocalIE(AnvatoIE):
             # m3u8 download
             'skip_download': True,
         },
-    }, {
-        'url': 'http://newyork.cbslocal.com/video/3580809-a-very-blue-anniversary/',
-        'info_dict': {
-            'id': '3580809',
-            'ext': 'mp4',
-            'title': 'A Very Blue Anniversary',
-            'description': 'CBS2’s Cindy Hsu has more.',
-            'thumbnail': 're:^https?://.*',
-            'timestamp': int,
-            'upload_date': r're:^\d{8}$',
-            'uploader': 'CBS',
-            'subtitles': {
-                'en': 'mincount:5',
-            },
-            'categories': [
-                'Stations\\Spoken Word\\WCBSTV',
-                'Syndication\\AOL',
-                'Syndication\\MSN',
-                'Syndication\\NDN',
-                'Syndication\\Yahoo',
-                'Content\\News',
-                'Content\\News\\Local News',
-            ],
-            'tags': ['CBS 2 News Weekends', 'Cindy Hsu', 'Blue Man Group'],
-        },
     }]
 
     def _real_extract(self, url):

youtube_dl/extractor/extractors.py

@@ -163,7 +163,10 @@ from .cbc import (
     CBCOlympicsIE,
 )
 from .cbs import CBSIE
-from .cbslocal import CBSLocalIE
+from .cbslocal import (
+    CBSLocalIE,
+    CBSLocalArticleIE,
+)
 from .cbsinteractive import CBSInteractiveIE
 from .cbsnews import (
     CBSNewsEmbedIE,

youtube_dl/extractor/facebook.py

@@ -1,6 +1,7 @@
 # coding: utf-8
 from __future__ import unicode_literals
 
+import json
 import re
 import socket
 
@@ -8,6 +9,7 @@ from .common import InfoExtractor
 from ..compat import (
     compat_etree_fromstring,
     compat_http_client,
+    compat_str,
     compat_urllib_error,
     compat_urllib_parse_unquote,
     compat_urllib_parse_unquote_plus,
@@ -47,7 +49,8 @@ class FacebookIE(InfoExtractor):
                             )\?(?:.*?)(?:v|video_id|story_fbid)=|
                             [^/]+/videos/(?:[^/]+/)?|
                             [^/]+/posts/|
-                            groups/[^/]+/permalink/
+                            groups/[^/]+/permalink/|
+                            watchparty/
                         )|
                     facebook:
                 )
@@ -280,8 +283,18 @@ class FacebookIE(InfoExtractor):
         # data.video.creation_story.attachments[].media
         'url': 'https://www.facebook.com/watch/live/?v=1823658634322275',
         'only_matching': True,
+    }, {
+        'url': 'https://www.facebook.com/watchparty/211641140192478',
+        'info_dict': {
+            'id': '211641140192478',
+        },
+        'playlist_count': 1,
+        'skip': 'Requires logging in',
     }]
     _SUPPORTED_PAGLETS_REGEX = r'(?:pagelet_group_mall|permalink_video_pagelet|hyperfeed_story_id_[0-9a-f]+)'
+    _api_config = {
+        'graphURI': '/api/graphql/'
+    }
 
     @staticmethod
     def _extract_urls(webpage):
@@ -405,6 +418,17 @@ class FacebookIE(InfoExtractor):
 
             self._sort_formats(formats)
 
+        def extract_relay_data(_filter):
+            return self._parse_json(self._search_regex(
+                r'handleWithCustomApplyEach\([^,]+,\s*({.*?%s.*?})\);' % _filter,
+                webpage, 'replay data', default='{}'), video_id, fatal=False) or {}
+
+        def extract_relay_prefetched_data(_filter):
+            replay_data = extract_relay_data(_filter)
+            for require in (replay_data.get('require') or []):
+                if require[0] == 'RelayPrefetchedStreamCache':
+                    return try_get(require, lambda x: x[3][1]['__bbox']['result']['data'], dict) or {}
+
         if not video_data:
             server_js_data = self._parse_json(self._search_regex([
                 r'bigPipe\.onPageletArrive\(({.+?})\)\s*;\s*}\s*\)\s*,\s*["\']onPageletArrive\s+' + self._SUPPORTED_PAGLETS_REGEX,
@@ -413,11 +437,9 @@ class FacebookIE(InfoExtractor):
             video_data = extract_from_jsmods_instances(server_js_data)
 
         if not video_data:
-            graphql_data = self._parse_json(self._search_regex(
-                r'handleWithCustomApplyEach\([^,]+,\s*({.*?"(?:dash_manifest|playable_url(?:_quality_hd)?)"\s*:\s*"[^"]+".*?})\);',
-                webpage, 'graphql data', default='{}'), video_id, fatal=False) or {}
-            for require in (graphql_data.get('require') or []):
-                if require[0] == 'RelayPrefetchedStreamCache':
+            data = extract_relay_prefetched_data(
+                r'"(?:dash_manifest|playable_url(?:_quality_hd)?)"\s*:\s*"[^"]+"')
+            if data:
                 entries = []
 
                 def parse_graphql_video(video):
@@ -459,8 +481,6 @@ class FacebookIE(InfoExtractor):
                     if media.get('__typename') == 'Video':
                         return parse_graphql_video(media)
 
-                    data = try_get(require, lambda x: x[3][1]['__bbox']['result']['data'], dict) or {}
-
                 nodes = data.get('nodes') or []
                 node = data.get('node') or {}
                 if not nodes and node:
@@ -504,6 +524,43 @@ class FacebookIE(InfoExtractor):
             elif '>You must log in to continue' in webpage:
                 self.raise_login_required()
 
+        if not video_data and '/watchparty/' in url:
+            post_data = {
+                'doc_id': 3731964053542869,
+                'variables': json.dumps({
+                    'livingRoomID': video_id,
+                }),
+            }
+
+            prefetched_data = extract_relay_prefetched_data(r'"login_data"\s*:\s*{')
+            if prefetched_data:
+                lsd = try_get(prefetched_data, lambda x: x['login_data']['lsd'], dict)
+                if lsd:
+                    post_data[lsd['name']] = lsd['value']
+
+            relay_data = extract_relay_data(r'\[\s*"RelayAPIConfigDefaults"\s*,')
+            for define in (relay_data.get('define') or []):
+                if define[0] == 'RelayAPIConfigDefaults':
+                    self._api_config = define[2]
+
+            living_room = self._download_json(
+                urljoin(url, self._api_config['graphURI']), video_id,
+                data=urlencode_postdata(post_data))['data']['living_room']
+
+            entries = []
+            for edge in (try_get(living_room, lambda x: x['recap']['watched_content']['edges']) or []):
+                video = try_get(edge, lambda x: x['node']['video']) or {}
+                v_id = video.get('id')
+                if not v_id:
+                    continue
+                v_id = compat_str(v_id)
+                entries.append(self.url_result(
+                    self._VIDEO_PAGE_TEMPLATE % v_id,
+                    self.ie_key(), v_id, video.get('name')))
+
+            return self.playlist_result(entries, video_id)
+
+        if not video_data:
             # Video info not in first request, do a secondary request using
             # tahoe player specific URL
             tahoe_data = self._download_webpage(

youtube_dl/extractor/sonyliv.py

@@ -1,40 +1,112 @@
 # coding: utf-8
 from __future__ import unicode_literals
 
+import time
+import uuid
+
 from .common import InfoExtractor
-from ..utils import smuggle_url
+from ..compat import compat_HTTPError
+from ..utils import (
+    ExtractorError,
+    int_or_none,
+)
 
 
 class SonyLIVIE(InfoExtractor):
-    _VALID_URL = r'https?://(?:www\.)?sonyliv\.com/details/[^/]+/(?P<id>\d+)'
+    _VALID_URL = r'https?://(?:www\.)?sonyliv\.com/(?:s(?:how|port)s/[^/]+|movies|clip|trailer|music-videos)/[^/?#&]+-(?P<id>\d+)'
     _TESTS = [{
-        'url': "http://www.sonyliv.com/details/episodes/5024612095001/Ep.-1---Achaari-Cheese-Toast---Bachelor's-Delight",
+        'url': 'https://www.sonyliv.com/shows/bachelors-delight-1700000113/achaari-cheese-toast-1000022678?watch=true',
         'info_dict': {
-            'title': "Ep. 1 - Achaari Cheese Toast - Bachelor's Delight",
-            'id': 'ref:5024612095001',
+            'title': 'Bachelors Delight - Achaari Cheese Toast',
+            'id': '1000022678',
             'ext': 'mp4',
-            'upload_date': '20170923',
-            'description': 'md5:7f28509a148d5be9d0782b4d5106410d',
-            'uploader_id': '5182475815001',
-            'timestamp': 1506200547,
+            'upload_date': '20200411',
+            'description': 'md5:3957fa31d9309bf336ceb3f37ad5b7cb',
+            'timestamp': 1586632091,
+            'duration': 185,
+            'season_number': 1,
+            'episode': 'Achaari Cheese Toast',
+            'episode_number': 1,
+            'release_year': 2016,
         },
         'params': {
             'skip_download': True,
         },
-        'add_ie': ['BrightcoveNew'],
     }, {
-        'url': 'http://www.sonyliv.com/details/full%20movie/4951168986001/Sei-Raat-(Bangla)',
+        'url': 'https://www.sonyliv.com/movies/tahalka-1000050121?watch=true',
+        'only_matching': True,
+    }, {
+        'url': 'https://www.sonyliv.com/clip/jigarbaaz-1000098925',
+        'only_matching': True,
+    }, {
+        'url': 'https://www.sonyliv.com/trailer/sandwiched-forever-1000100286?watch=true',
+        'only_matching': True,
+    }, {
+        'url': 'https://www.sonyliv.com/sports/india-tour-of-australia-2020-21-1700000286/cricket-hls-day-3-1st-test-aus-vs-ind-19-dec-2020-1000100959?watch=true',
+        'only_matching': True,
+    }, {
+        'url': 'https://www.sonyliv.com/music-videos/yeh-un-dinon-ki-baat-hai-1000018779',
         'only_matching': True,
     }]
+    _GEO_COUNTRIES = ['IN']
+    _TOKEN = None
 
-    # BRIGHTCOVE_URL_TEMPLATE = 'http://players.brightcove.net/4338955589001/default_default/index.html?videoId=%s'
-    BRIGHTCOVE_URL_TEMPLATE = 'http://players.brightcove.net/5182475815001/default_default/index.html?videoId=ref:%s'
+    def _call_api(self, version, path, video_id):
+        headers = {}
+        if self._TOKEN:
+            headers['security_token'] = self._TOKEN
+        try:
+            return self._download_json(
+                'https://apiv2.sonyliv.com/AGL/%s/A/ENG/WEB/%s' % (version, path),
+                video_id, headers=headers)['resultObj']
+        except ExtractorError as e:
+            if isinstance(e.cause, compat_HTTPError) and e.cause.code == 403:
+                message = self._parse_json(
+                    e.cause.read().decode(), video_id)['message']
+                if message == 'Geoblocked Country':
+                    self.raise_geo_restricted(countries=self._GEO_COUNTRIES)
+                raise ExtractorError(message)
+            raise
+
+    def _real_initialize(self):
+        self._TOKEN = self._call_api('1.4', 'ALL/GETTOKEN', None)
 
     def _real_extract(self, url):
-        brightcove_id = self._match_id(url)
-        return self.url_result(
-            smuggle_url(self.BRIGHTCOVE_URL_TEMPLATE % brightcove_id, {
-                'geo_countries': ['IN'],
-                'referrer': url,
-            }),
-            'BrightcoveNew', brightcove_id)
+        video_id = self._match_id(url)
+        content = self._call_api(
+            '1.5', 'IN/CONTENT/VIDEOURL/VOD/' + video_id, video_id)
+        if content.get('isEncrypted'):
+            raise ExtractorError('This video is DRM protected.', expected=True)
+        dash_url = content['videoURL']
+        headers = {
+            'x-playback-session-id': '%s-%d' % (uuid.uuid4().hex, time.time() * 1000)
+        }
+        formats = self._extract_mpd_formats(
+            dash_url, video_id, mpd_id='dash', headers=headers, fatal=False)
+        formats.extend(self._extract_m3u8_formats(
+            dash_url.replace('.mpd', '.m3u8').replace('/DASH/', '/HLS/'),
+            video_id, 'mp4', m3u8_id='hls', headers=headers, fatal=False))
+        for f in formats:
+            f.setdefault('http_headers', {}).update(headers)
+        self._sort_formats(formats)
+
+        metadata = self._call_api(
+            '1.6', 'IN/DETAIL/' + video_id, video_id)['containers'][0]['metadata']
+        title = metadata['title']
+        episode = metadata.get('episodeTitle')
+        if episode and title != episode:
+            title += ' - ' + episode
+
+        return {
+            'id': video_id,
+            'title': title,
+            'formats': formats,
+            'thumbnail': content.get('posterURL'),
+            'description': metadata.get('longDescription') or metadata.get('shortDescription'),
+            'timestamp': int_or_none(metadata.get('creationDate'), 1000),
+            'duration': int_or_none(metadata.get('duration')),
+            'season_number': int_or_none(metadata.get('season')),
+            'episode': episode,
+            'episode_number': int_or_none(metadata.get('episodeNumber')),
+            'release_year': int_or_none(metadata.get('year')),
+        }

youtube_dl/extractor/streetvoice.py

@@ -2,25 +2,40 @@
 from __future__ import unicode_literals
 
 from .common import InfoExtractor
-from ..compat import compat_str
-from ..utils import unified_strdate
+from ..utils import (
+    int_or_none,
+    parse_iso8601,
+    str_or_none,
+    strip_or_none,
+    try_get,
+    urljoin,
+)
 
 
 class StreetVoiceIE(InfoExtractor):
     _VALID_URL = r'https?://(?:.+?\.)?streetvoice\.com/[^/]+/songs/(?P<id>[0-9]+)'
     _TESTS = [{
-        'url': 'http://streetvoice.com/skippylu/songs/94440/',
-        'md5': '15974627fc01a29e492c98593c2fd472',
+        'url': 'https://streetvoice.com/skippylu/songs/123688/',
+        'md5': '0eb535970629a5195685355f3ed60bfd',
         'info_dict': {
-            'id': '94440',
+            'id': '123688',
             'ext': 'mp3',
-            'title': '輸',
-            'description': 'Crispy脆樂團 - 輸',
-            'thumbnail': r're:^https?://.*\.jpg$',
-            'duration': 260,
-            'upload_date': '20091018',
+            'title': '流浪',
+            'description': 'md5:8eb0bfcc9dcd8aa82bd6efca66e3fea6',
+            'thumbnail': r're:^https?://.*\.jpg',
+            'duration': 270,
+            'upload_date': '20100923',
             'uploader': 'Crispy脆樂團',
             'uploader_id': '627810',
+            'uploader_url': 're:^https?://streetvoice.com/skippylu/',
+            'timestamp': 1285261661,
+            'view_count': int,
+            'like_count': int,
+            'comment_count': int,
+            'repost_count': int,
+            'track': '流浪',
+            'track_id': '123688',
+            'album': '2010',
         }
     }, {
         'url': 'http://tw.streetvoice.com/skippylu/songs/94440/',
@@ -29,21 +44,57 @@ class StreetVoiceIE(InfoExtractor):
 
     def _real_extract(self, url):
         song_id = self._match_id(url)
-
-        song = self._download_json(
-            'https://streetvoice.com/api/v1/public/song/%s/' % song_id, song_id, data=b'')
-
+        base_url = 'https://streetvoice.com/api/v4/song/%s/' % song_id
+        song = self._download_json(base_url, song_id, query={
+            'fields': 'album,comments_count,created_at,id,image,length,likes_count,name,nickname,plays_count,profile,share_count,synopsis,user,username',
+        })
         title = song['name']
-        author = song['user']['nickname']
+
+        formats = []
+        for suffix, format_id in [('hls/file', 'hls'), ('file', 'http'), ('file/original', 'original')]:
+            f_url = (self._download_json(
+                base_url + suffix + '/', song_id,
+                'Downloading %s format URL' % format_id,
+                data=b'', fatal=False) or {}).get('file')
+            if not f_url:
+                continue
+            f = {
+                'ext': 'mp3',
+                'format_id': format_id,
+                'url': f_url,
+                'vcodec': 'none',
+            }
+            if format_id == 'hls':
+                f['protocol'] = 'm3u8_native'
+            abr = self._search_regex(r'\.mp3\.(\d+)k', f_url, 'bitrate', default=None)
+            if abr:
+                abr = int(abr)
+                f.update({
+                    'abr': abr,
+                    'tbr': abr,
+                })
+            formats.append(f)
+
+        user = song.get('user') or {}
+        username = user.get('username')
+        get_count = lambda x: int_or_none(song.get(x + '_count'))
 
         return {
             'id': song_id,
-            'url': song['file'],
+            'formats': formats,
             'title': title,
-            'description': '%s - %s' % (author, title),
-            'thumbnail': self._proto_relative_url(song.get('image'), 'http:'),
-            'duration': song.get('length'),
-            'upload_date': unified_strdate(song.get('created_at')),
-            'uploader': author,
-            'uploader_id': compat_str(song['user']['id']),
+            'description': strip_or_none(song.get('synopsis')),
+            'thumbnail': song.get('image'),
+            'duration': int_or_none(song.get('length')),
+            'timestamp': parse_iso8601(song.get('created_at')),
+            'uploader': try_get(user, lambda x: x['profile']['nickname']),
+            'uploader_id': str_or_none(user.get('id')),
+            'uploader_url': urljoin(url, '/%s/' % username) if username else None,
+            'view_count': get_count('plays'),
+            'like_count': get_count('likes'),
+            'comment_count': get_count('comments'),
+            'repost_count': get_count('share'),
+            'track': title,
+            'track_id': song_id,
+            'album': try_get(song, lambda x: x['album']['name']),
         }

youtube_dl/extractor/youtube.py

@@ -1322,17 +1322,16 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
             return self._parse_json(
                 uppercase_escape(config), video_id, fatal=False)
 
-    def _get_automatic_captions(self, video_id, webpage):
+    def _get_automatic_captions(self, video_id, player_response, player_config):
         """We need the webpage for getting the captions url, pass it as an
            argument to speed up the process."""
         self.to_screen('%s: Looking for automatic captions' % video_id)
-        player_config = self._get_ytplayer_config(video_id, webpage)
         err_msg = 'Couldn\'t find automatic captions for %s' % video_id
-        if not player_config:
+        if not (player_response or player_config):
             self._downloader.report_warning(err_msg)
             return {}
         try:
-            args = player_config['args']
+            args = player_config.get('args') if player_config else {}
             caption_url = args.get('ttsurl')
             if caption_url:
                 timestamp = args['timestamp']
@@ -1391,10 +1390,6 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
                 return captions
 
             # New captions format as of 22.06.2017
-            player_response = args.get('player_response')
-            if player_response and isinstance(player_response, compat_str):
-                player_response = self._parse_json(
-                    player_response, video_id, fatal=False)
             if player_response:
                 renderer = player_response['captions']['playerCaptionsTracklistRenderer']
                 base_url = renderer['captionTracks'][0]['baseUrl']
@@ -1652,6 +1647,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
         # Get video info
         video_info = {}
         embed_webpage = None
+        ytplayer_config = None
 
         if re.search(r'["\']status["\']\s*:\s*["\']LOGIN_REQUIRED', video_webpage) is not None:
             age_gate = True
@@ -2276,7 +2272,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
 
         # subtitles
         video_subtitles = self.extract_subtitles(video_id, video_webpage)
-        automatic_captions = self.extract_automatic_captions(video_id, video_webpage)
+        automatic_captions = self.extract_automatic_captions(video_id, player_response, ytplayer_config)
 
         video_duration = try_get(
             video_info, lambda x: int_or_none(x['length_seconds'][0]))