Merge 5c1e5a1160a3e65fd2673b6437f3882d348991cb into 2b4fbfce25902d557b86b003cf48f738129efce4

thx seproDev, bashonly: yt-dlp/yt-dlp#12748

.gitignore

@@ -1,3 +1,4 @@
+__pycache__/
 *.pyc
 *.pyo
 *.class
@@ -5,6 +6,7 @@
 *.DS_Store
 wine-py2exe/
 py2exe.log
+.pytest_cache/
 *.kate-swp
 build/
 dist/

test/test_youtube_signature.py

@@ -84,6 +84,21 @@ _SIG_TESTS = [
         '2aq0aqSyOoJXtK73m-uME_jv7-pT15gOFC02RFkGMqWpzEICs69VdbwQ0LDp1v7j8xx92efCJlYFYb1sUkkBSPOlPmXgIARw8JQ0qOAOAA',
         '0QJ8wRAIgXmPlOPSBkkUs1bYFYlJCfe29xxAj7v1pDL0QwbdV96sCIEzpWqMGkFR20CFOg51Tp-7vj_EMu-m37KtXJ2OySqa0q',
     ),
+    (
+        'https://www.youtube.com/s/player/643afba4/tv-player-ias.vflset/tv-player-ias.js',
+        '2aq0aqSyOoJXtK73m-uME_jv7-pT15gOFC02RFkGMqWpzEICs69VdbwQ0LDp1v7j8xx92efCJlYFYb1sUkkBSPOlPmXgIARw8JQ0qOAOAA',
+        'AAOAOq0QJ8wRAIgXmPlOPSBkkUs1bYFYlJCfe29xx8j7vgpDL0QwbdV06sCIEzpWqMGkFR20CFOS21Tp-7vj_EMu-m37KtXJoOy1',
+    ),
+    (
+        'https://www.youtube.com/s/player/363db69b/player_ias.vflset/en_US/base.js',
+        '2aq0aqSyOoJXtK73m-uME_jv7-pT15gOFC02RFkGMqWpzEICs69VdbwQ0LDp1v7j8xx92efCJlYFYb1sUkkBSPOlPmXgIARw8JQ0qOAOAA',
+        '0aqSyOoJXtK73m-uME_jv7-pT15gOFC02RFkGMqWpz2ICs6EVdbwQ0LDp1v7j8xx92efCJlYFYb1sUkkBSPOlPmXgIARw8JQ0qOAOAA',
+    ),
+    (
+        'https://www.youtube.com/s/player/4fcd6e4a/player_ias.vflset/en_US/base.js',
+        '2aq0aqSyOoJXtK73m-uME_jv7-pT15gOFC02RFkGMqWpzEICs69VdbwQ0LDp1v7j8xx92efCJlYFYb1sUkkBSPOlPmXgIARw8JQ0qOAOAA',
+        'wAOAOq0QJ8ARAIgXmPlOPSBkkUs1bYFYlJCfe29xx8q7v1pDL0QwbdV96sCIEzpWqMGkFR20CFOg51Tp-7vj_EMu-m37KtXJoOySqa0',
+    ),
 ]
 
 _NSIG_TESTS = [
@@ -153,7 +168,7 @@ _NSIG_TESTS = [
     ),
     (
         'https://www.youtube.com/s/player/c57c113c/player_ias.vflset/en_US/base.js',
-        '-Txvy6bT5R6LqgnQNx', 'dcklJCnRUHbgSg',
+        'M92UUMHa8PdvPd3wyM', '3hPqLJsiNZx7yA',
     ),
     (
         'https://www.youtube.com/s/player/5a3b6271/player_ias.vflset/en_US/base.js',
@@ -173,7 +188,7 @@ _NSIG_TESTS = [
     ),
     (
         'https://www.youtube.com/s/player/cfa9e7cb/player_ias.vflset/en_US/base.js',
-        'qO0NiMtYQ7TeJnfFG2', 'k9cuJDHNS5O7kQ',
+        'aCi3iElgd2kq0bxVbQ', 'QX1y8jGb2IbZ0w',
     ),
     (
         'https://www.youtube.com/s/player/8c7583ff/player_ias.vflset/en_US/base.js',
@@ -231,10 +246,6 @@ _NSIG_TESTS = [
         'https://www.youtube.com/s/player/f6e09c70/player_ias_tce.vflset/en_US/base.js',
         'W9HJZKktxuYoDTqW', 'jHbbkcaxm54',
     ),
-    (
-        'https://www.youtube.com/s/player/643afba4/player_ias.vflset/en_US/base.js',
-        'W9HJZKktxuYoDTqW', 'larxUlagTRAcSw',
-    ),
     (
         'https://www.youtube.com/s/player/e7567ecf/player_ias_tce.vflset/en_US/base.js',
         'Sy4aDGc0VpYRR9ew_', '5UPOT1VhoZxNLQ',
@@ -259,6 +270,22 @@ _NSIG_TESTS = [
         'https://www.youtube.com/s/player/643afba4/tv-player-ias.vflset/tv-player-ias.js',
         'ir9-V6cdbCiyKxhr', '2PL7ZDYAALMfmA',
     ),
+    (
+        'https://www.youtube.com/s/player/643afba4/player_ias.vflset/en_US/base.js',
+        'W9HJZKktxuYoDTqW', 'larxUlagTRAcSw',
+    ),
+    (
+        'https://www.youtube.com/s/player/363db69b/player_ias.vflset/en_US/base.js',
+        'eWYu5d5YeY_4LyEDc', 'XJQqf-N7Xra3gg',
+    ),
+    (
+        'https://www.youtube.com/s/player/4fcd6e4a/player_ias.vflset/en_US/base.js',
+        'o_L251jm8yhZkWtBW', 'lXoxI3XvToqn6A',
+    ),
+    (
+        'https://www.youtube.com/s/player/4fcd6e4a/tv-player-ias.vflset/tv-player-ias.js',
+        'o_L251jm8yhZkWtBW', 'lXoxI3XvToqn6A',
+    ),
 ]
 
 
@@ -271,6 +298,8 @@ class TestPlayerInfo(unittest.TestCase):
             ('https://www.youtube.com/s/player/64dddad9/player-plasma-ias-phone-en_US.vflset/base.js', '64dddad9'),
             ('https://www.youtube.com/s/player/64dddad9/player-plasma-ias-phone-de_DE.vflset/base.js', '64dddad9'),
             ('https://www.youtube.com/s/player/64dddad9/player-plasma-ias-tablet-en_US.vflset/base.js', '64dddad9'),
+            ('https://www.youtube.com/s/player/e7567ecf/player_ias_tce.vflset/en_US/base.js', 'e7567ecf'),
+            ('https://www.youtube.com/s/player/643afba4/tv-player-ias.vflset/tv-player-ias.js', '643afba4'),
             # obsolete
             ('https://www.youtube.com/yts/jsbin/player_ias-vfle4-e03/en_US/base.js', 'vfle4-e03'),
             ('https://www.youtube.com/yts/jsbin/player_ias-vfl49f_g4/en_US/base.js', 'vfl49f_g4'),
@@ -280,8 +309,9 @@ class TestPlayerInfo(unittest.TestCase):
             ('https://s.ytimg.com/yts/jsbin/html5player-en_US-vflXGBaUN.js', 'vflXGBaUN'),
             ('https://s.ytimg.com/yts/jsbin/html5player-en_US-vflKjOTVq/html5player.js', 'vflKjOTVq'),
         )
+        ie = YoutubeIE(FakeYDL({'cachedir': False}))
         for player_url, expected_player_id in PLAYER_URLS:
-            player_id = YoutubeIE._extract_player_info(player_url)
+            player_id = ie._extract_player_info(player_url)
             self.assertEqual(player_id, expected_player_id)
 
 
@@ -301,8 +331,8 @@ class TestSignature(unittest.TestCase):
 def t_factory(name, sig_func, url_pattern):
     def make_tfunc(url, sig_input, expected_sig):
         m = url_pattern.match(url)
-        assert m, '%r should follow URL format' % url
-        test_id = m.group('id')
+        assert m, '{0!r} should follow URL format'.format(url)
+        test_id = re.sub(r'[/.-]', '_', m.group('id') or m.group('compat_id'))
 
         def test_func(self):
             basename = 'player-{0}-{1}.js'.format(name, test_id)
@@ -335,12 +365,16 @@ def n_sig(jscode, sig_input):
 
 
 make_sig_test = t_factory(
-    'signature', signature, re.compile(r'.*(?:-|/player/)(?P<id>[a-zA-Z0-9_-]+)(?:/.+\.js|(?:/watch_as3|/html5player)?\.[a-z]+)$'))
+    'signature', signature,
+    re.compile(r'''(?x)
+        .+/(?P<h5>html5)?player(?(h5)(?:-en_US)?-|/)(?P<id>[a-zA-Z0-9/._-]+)
+        (?(h5)/(?:watch_as3|html5player))?\.js$
+    '''))
 for test_spec in _SIG_TESTS:
     make_sig_test(*test_spec)
 
 make_nsig_test = t_factory(
-    'nsig', n_sig, re.compile(r'.+/player/(?P<id>[a-zA-Z0-9_-]+)/.+.js$'))
+    'nsig', n_sig, re.compile(r'.+/player/(?P<id>[a-zA-Z0-9_/.-]+)\.js$'))
 for test_spec in _NSIG_TESTS:
     make_nsig_test(*test_spec)
 

youtube_dl/extractor/extractors.py

@@ -1328,7 +1328,10 @@ from .trovo import (
 from .trunews import TruNewsIE
 from .trutv import TruTVIE
 from .tube8 import Tube8IE
-from .tubitv import TubiTvIE
+from .tubitv import (
+    TubiTvIE,
+    TubiTvShowIE,
+)
 from .tumblr import TumblrIE
 from .tunein import (
     TuneInClipIE,

youtube_dl/extractor/tubitv.py

@@ -1,23 +1,62 @@
 # coding: utf-8
 from __future__ import unicode_literals
 
-import re
-
 from .common import InfoExtractor
 from ..utils import (
     ExtractorError,
+    get_element_by_id,
     int_or_none,
+    join_nonempty,
+    js_to_json,
+    merge_dicts,
+    parse_age_limit,
     sanitized_Request,
+    strip_or_none,
+    T,
+    traverse_obj,
+    url_or_none,
     urlencode_postdata,
 )
 
 
 class TubiTvIE(InfoExtractor):
-    _VALID_URL = r'https?://(?:www\.)?tubitv\.com/(?:video|movies|tv-shows)/(?P<id>[0-9]+)'
+    IE_NAME = 'tubitv'
+    _VALID_URL = r'https?://(?:www\.)?tubitv\.com/(?P<type>video|movies|tv-shows)/(?P<id>\d+)'
     _LOGIN_URL = 'http://tubitv.com/login'
     _NETRC_MACHINE = 'tubitv'
-    _GEO_COUNTRIES = ['US']
     _TESTS = [{
+        'url': 'https://tubitv.com/movies/100004539/the-39-steps',
+        'info_dict': {
+            'id': '100004539',
+            'ext': 'mp4',
+            'title': 'The 39 Steps',
+            'description': 'md5:bb2f2dd337f0dc58c06cb509943f54c8',
+            'uploader_id': 'abc2558d54505d4f0f32be94f2e7108c',
+            'release_year': 1935,
+            'thumbnail': r're:^https?://.+\.(jpe?g|png)$',
+            'duration': 5187,
+        },
+        'params': {'skip_download': 'm3u8'},
+        'skip': 'This content is currently unavailable',
+    }, {
+        'url': 'https://tubitv.com/tv-shows/554628/s01-e01-rise-of-the-snakes',
+        'info_dict': {
+            'id': '554628',
+            'ext': 'mp4',
+            'title': 'S01:E01 - Rise of the Snakes',
+            'description': 'md5:ba136f586de53af0372811e783a3f57d',
+            'episode': 'Rise of the Snakes',
+            'episode_number': 1,
+            'season': 'Season 1',
+            'season_number': 1,
+            'uploader_id': '2a9273e728c510d22aa5c57d0646810b',
+            'release_year': 2011,
+            'thumbnail': r're:^https?://.+\.(jpe?g|png)$',
+            'duration': 1376,
+        },
+        'params': {'skip_download': 'm3u8'},
+        'skip': 'This content is currently unavailable',
+    }, {
         'url': 'http://tubitv.com/video/283829/the_comedian_at_the_friday',
         'md5': '43ac06be9326f41912dc64ccf7a80320',
         'info_dict': {
@@ -27,6 +66,7 @@ class TubiTvIE(InfoExtractor):
             'description': 'A stand up comedian is forced to look at the decisions in his life while on a one week trip to the west coast.',
             'uploader_id': 'bc168bee0d18dd1cb3b86c68706ab434',
         },
+        'skip': 'Content Unavailable',
     }, {
         'url': 'http://tubitv.com/tv-shows/321886/s01_e01_on_nom_stories',
         'only_matching': True,
@@ -34,24 +74,42 @@ class TubiTvIE(InfoExtractor):
         'url': 'http://tubitv.com/movies/383676/tracker',
         'only_matching': True,
     }, {
-        'url': 'https://tubitv.com/movies/560057/penitentiary?start=true',
+        'url': 'https://tubitv.com/tv-shows/200141623/s01-e01-episode-1',
         'info_dict': {
-            'id': '560057',
+            'id': '200141623',
             'ext': 'mp4',
-            'title': 'Penitentiary',
-            'description': 'md5:8d2fc793a93cc1575ff426fdcb8dd3f9',
-            'uploader_id': 'd8fed30d4f24fcb22ec294421b9defc2',
-            'release_year': 1979,
+            'title': 'Shameless S01:E01 - Episode 1',
+            'description': 'Having her handbag stolen proves to be a blessing in disguise for Fiona when handsome stranger Steve comes to her rescue.',
+            'timestamp': 1725148800,
+            'upload_date': '20240901',
+            'uploader': 'all3-media',
+            'uploader_id': '9b8e3a8d789b1c843f4b680c025a1853',
+            'release_year': 2004,
+            'episode': 'Episode 1',
+            'episode_number': 1,
+            'season': 'Season 1',
+            'season_number': 1,
+            'series': 'Shameless',
+            'cast': list,
+            'age_limit': 17,
         },
         'params': {
-            'skip_download': True,
+            'format': 'best/bestvideo',
+            'skip_download': 'm3u8'
         },
     }]
 
+    # DRM formats are included only to raise appropriate error
+    _UNPLAYABLE_FORMATS = ('hlsv6_widevine', 'hlsv6_widevine_nonclearlead', 'hlsv6_playready_psshv0',
+                           'hlsv6_fairplay', 'dash_widevine', 'dash_widevine_nonclearlead')
+
     def _login(self):
         username, password = self._get_login_info()
         if username is None:
             return
+        self._perform_login(username, password)
+
+    def _perform_login(self, username, password):
         self.report_login()
         form_data = {
             'username': username,
@@ -62,7 +120,7 @@ class TubiTvIE(InfoExtractor):
         request.add_header('Content-Type', 'application/x-www-form-urlencoded')
         login_page = self._download_webpage(
             request, None, False, 'Wrong login info')
-        if not re.search(r'id="tubi-logout"', login_page):
+        if get_element_by_id('tubi-logout', login_page) is None:
             raise ExtractorError(
                 'Login failed (invalid username/password)', expected=True)
 
@@ -70,41 +128,146 @@ class TubiTvIE(InfoExtractor):
         self._login()
 
     def _real_extract(self, url):
-        video_id = self._match_id(url)
-        video_data = self._download_json(
-            'http://tubitv.com/oz/videos/%s/content' % video_id, video_id)
-        title = video_data['title']
+        video_id, video_type = self._match_valid_url(url).group('id', 'type')
+        webpage = self._download_webpage('https://tubitv.com/{0}/{1}/'.format(video_type, video_id), video_id)
+        data = self._search_json(
+            r'window\.__data\s*=', webpage, 'data', video_id,
+            transform_source=js_to_json)['video']['byId']
+        video_data = data[video_id]
+        info = self._search_json_ld(webpage, video_id, expected_type='VideoObject', default={})
+        title = strip_or_none(info.get('title'))
+        info['title'] = title or strip_or_none(video_data['title'])
 
-        formats = self._extract_m3u8_formats(
-            self._proto_relative_url(video_data['url']),
-            video_id, 'mp4', 'm3u8_native')
+        formats = []
+        drm_formats = 0
+
+        for resource in traverse_obj(video_data, ('video_resources', lambda _, v: v['type'] and v['manifest']['url'])):
+            manifest_url = url_or_none(resource['manifest']['url'])
+            if not manifest_url:
+                continue
+            resource_type = resource['type']
+            if resource_type == 'dash':
+                formats.extend(self._extract_mpd_formats(manifest_url, video_id, mpd_id=resource_type, fatal=False))
+            elif resource_type in ('hlsv3', 'hlsv6'):
+                formats.extend(self._extract_m3u8_formats(manifest_url, video_id, 'mp4', m3u8_id=resource_type, fatal=False))
+            elif resource_type in self._UNPLAYABLE_FORMATS:
+                drm_formats += 1
+            else:
+                self.report_warning('Skipping unknown resource type "{0}"'.format(resource_type))
+
+        if not formats and drm_formats > 0:
+            self.report_drm(video_id)
+        elif not formats and not video_data.get('policy_match'):  # policy_match is False if content was removed
+            raise ExtractorError('This content is currently unavailable', expected=True)
         self._sort_formats(formats)
 
-        thumbnails = []
-        for thumbnail_url in video_data.get('thumbnails', []):
-            if not thumbnail_url:
-                continue
-            thumbnails.append({
-                'url': self._proto_relative_url(thumbnail_url),
-            })
-
         subtitles = {}
-        for sub in video_data.get('subtitles', []):
-            sub_url = sub.get('url')
+        for sub in traverse_obj(video_data, ('subtitles', lambda _, v: v['url'])):
+            sub_url = self._proto_relative_url(sub['url'])
             if not sub_url:
                 continue
             subtitles.setdefault(sub.get('lang', 'English'), []).append({
-                'url': self._proto_relative_url(sub_url),
+                'url': sub_url,
             })
 
-        return {
+        season_number, episode_number, episode_title = self._search_regex(
+            r'\bS(\d+):E(\d+) - (.+)', info['title'], 'episode info', fatal=False, group=(1, 2, 3), default=(None, None, None))
+
+        return merge_dicts({
             'id': video_id,
-            'title': title,
             'formats': formats,
             'subtitles': subtitles,
-            'thumbnails': thumbnails,
-            'description': video_data.get('description'),
-            'duration': int_or_none(video_data.get('duration')),
-            'uploader_id': video_data.get('publisher_id'),
-            'release_year': int_or_none(video_data.get('year')),
-        }
+            'season_number': int_or_none(season_number),
+            'episode_number': int_or_none(episode_number),
+            'episode': episode_title
+        }, traverse_obj(video_data, {
+            'description': ('description', T(strip_or_none)),
+            'duration': ('duration', T(int_or_none)),
+            'uploader': ('import_id', T(strip_or_none)),
+            'uploader_id': ('publisher_id', T(strip_or_none)),
+            'release_year': ('year', T(int_or_none)),
+            'thumbnails': ('thumbnails', Ellipsis, T(self._proto_relative_url), {'url': T(url_or_none)}),
+            'cast': ('actors', Ellipsis, T(strip_or_none)),
+            'categories': ('tags', Ellipsis, T(strip_or_none)),
+            'age_limit': ('ratings', 0, 'value', T(parse_age_limit)),
+        }), traverse_obj(data, (lambda _, v: v['type'] == 's', {
+            'series': ('title', T(strip_or_none)),
+            # 'series_id': ('id', T(compat_str)),
+        }), get_all=False), info)
+
+
+class TubiTvShowIE(InfoExtractor):
+    IE_NAME = 'tubitv:series'
+    _VALID_URL = r'https?://(?:www\.)?tubitv\.com/series/\d+/(?P<show_name>[^/?#]+)(?:/season-(?P<season>\d+))?'
+    _TESTS = [{
+        'url': 'https://tubitv.com/series/3936/the-joy-of-painting-with-bob-ross?start=true',
+        'playlist_mincount': 390,
+        'info_dict': {
+            'id': 'the-joy-of-painting-with-bob-ross',
+        },
+    }, {
+        'url': 'https://tubitv.com/series/3936/the-joy-of-painting-with-bob-ross/season-1',
+        'playlist_count': 13,
+        'info_dict': {
+            'id': 'the-joy-of-painting-with-bob-ross-season-1',
+        },
+    }, {
+        'url': 'https://tubitv.com/series/3936/the-joy-of-painting-with-bob-ross/season-3',
+        'playlist_count': 13,
+        'info_dict': {
+            'id': 'the-joy-of-painting-with-bob-ross-season-3',
+        },
+    }]
+
+    def _real_extract(self, url):
+        playlist_id, selected_season = self._match_valid_url(url).group(
+            'show_name', 'season')
+
+        def entries(s_url, playlist_id, selected_season_num):
+
+            def get_season_data(s_num, fatal=False):
+                if s_num is None:
+                    url, s_id = s_url, playlist_id
+                else:
+                    url = '%s/season-%d' % (s_url, s_num)
+                    s_id = '%s-season-%d' % (playlist_id, s_num)
+                webpage = self._download_webpage(url, s_id, fatal=fatal)
+                data = self._search_json(
+                    r'window\s*\.\s*__data\s*=', webpage or '', 'data', s_id,
+                    transform_source=js_to_json, default={})
+                return data['video'] if fatal else data.get('video', {})
+
+            data = get_season_data(None, fatal=True)
+            # The {series_id}.seasons JSON may lack some episodes that are available
+            # Iterate over the season numbers instead [1]
+            # 1. https://github.com/yt-dlp/yt-dlp/issues/11170#issuecomment-2399918777
+            seasons = (
+                traverse_obj(data, (
+                    'byId', lambda _, v: v['type'] == 's', 'seasons', Ellipsis,
+                    'number', T(int_or_none)))
+                if selected_season is None
+                else [selected_season])
+
+            unavail_cnt = 0
+            select_episodes = lambda _, v: v['type'] == 'v'
+            for season_number in seasons:
+                if not data:
+                    data = get_season_data(season_number)
+
+                unavail_cnt += len(traverse_obj(data, ('byId', select_episodes, 'policy_match', T(lambda m: (not m) or None))))
+
+                for episode_id, episode in traverse_obj(data, ('byId', select_episodes, T(lambda e: (e['id'], e)))):
+                    yield merge_dicts(self.url_result(
+                        'https://tubitv.com/tv-shows/{0}/'.format(episode_id), TubiTvIE.ie_key(), episode_id), {
+                            'season_number': season_number,
+                            'episode_number': int_or_none(episode.get('num')),
+                    })
+
+                data = None
+
+            if unavail_cnt > 0:
+                self.report_warning('%d items were marked as unavailable: check that the desired content is available or provide login parameters if needed' % unavail_cnt)
+
+        return self.playlist_result(
+            entries(url, playlist_id, int_or_none(selected_season)),
+            join_nonempty(playlist_id, selected_season, delim='-season-'))

youtube_dl/extractor/youtube.py

@@ -692,9 +692,9 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
         'invidious': '|'.join(_INVIDIOUS_SITES),
     }
     _PLAYER_INFO_RE = (
-        r'/s/player/(?P<id>[a-zA-Z0-9_-]{8,})//(?:tv-)?player',
-        r'/(?P<id>[a-zA-Z0-9_-]{8,})/player(?:_ias\.vflset(?:/[a-zA-Z]{2,3}_[a-zA-Z]{2,3})?|-plasma-ias-(?:phone|tablet)-[a-z]{2}_[A-Z]{2}\.vflset)/base\.js$',
-        r'\b(?P<id>vfl[a-zA-Z0-9_-]+)\b.*?\.js$',
+        r'/s/player/(?P<id>[a-zA-Z0-9_-]{8,})/(?:tv-)?player',
+        r'/(?P<id>[a-zA-Z0-9_-]{8,})/player(?:_ias(?:_tce)?\.vflset(?:/[a-zA-Z]{2,3}_[a-zA-Z]{2,3})?|-plasma-ias-(?:phone|tablet)-[a-z]{2}_[A-Z]{2}\.vflset)/base\.js$',
+        r'\b(?P<id>vfl[a-zA-Z0-9_-]{6,})\b.*?\.js$',
     )
     _SUBTITLE_FORMATS = ('json3', 'srv1', 'srv2', 'srv3', 'ttml', 'vtt')
 
@@ -1626,15 +1626,13 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
         """ Return a string representation of a signature """
         return '.'.join(compat_str(len(part)) for part in example_sig.split('.'))
 
-    @classmethod
-    def _extract_player_info(cls, player_url):
-        for player_re in cls._PLAYER_INFO_RE:
-            id_m = re.search(player_re, player_url)
-            if id_m:
-                break
-        else:
-            raise ExtractorError('Cannot identify player %r' % player_url)
-        return id_m.group('id')
+    def _extract_player_info(self, player_url):
+        try:
+            return self._search_regex(
+                self._PLAYER_INFO_RE, player_url, 'player info', group='id')
+        except ExtractorError as e:
+            raise ExtractorError(
+                'Cannot identify player %r' % (player_url,), cause=e)
 
     def _load_player(self, video_id, player_url, fatal=True, player_id=None):
         if not player_id:
@@ -1711,6 +1709,23 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
                 '    return %s\n') % (signature_id_tuple, expr_code)
         self.to_screen('Extracted signature function:\n' + code)
 
+    def _extract_sig_fn(self, jsi, funcname):
+        var_ay = self._search_regex(
+            r'''(?x)
+                (?:\*/|\{|\n|^)\s*(?:'[^']+'\s*;\s*)
+                    (var\s*[\w$]+\s*=\s*(?:
+                        ('|")(?:\\\2|(?!\2).)+\2\s*\.\s*split\(\s*('|")\W+\3\s*\)|
+                        \[\s*(?:('|")(?:\\\4|(?!\4).)*\4\s*(?:(?=\])|,\s*))+\]
+                    ))(?=\s*[,;])
+            ''', jsi.code, 'useful values', default='')
+
+        sig_fn = jsi.extract_function_code(funcname)
+
+        if var_ay:
+            sig_fn = (sig_fn[0], ';\n'.join((var_ay, sig_fn[1])))
+
+        return sig_fn
+
     def _parse_sig_js(self, jscode):
         # Examples where `sig` is funcname:
         # sig=function(a){a=a.split(""); ... ;return a.join("")};
@@ -1736,8 +1751,12 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
             jscode, 'Initial JS player signature function name', group='sig')
 
         jsi = JSInterpreter(jscode)
-        initial_function = jsi.extract_function(funcname)
-        return lambda s: initial_function([s])
+
+        initial_function = self._extract_sig_fn(jsi, funcname)
+
+        func = jsi.extract_function_from_code(*initial_function)
+
+        return lambda s: func([s])
 
     def _cached(self, func, *cache_id):
         def inner(*args, **kwargs):
@@ -1856,15 +1875,9 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
 
     def _extract_n_function_code_jsi(self, video_id, jsi, player_id=None):
 
-        var_ay = self._search_regex(
-            r'(?:[;\s]|^)\s*(var\s*[\w$]+\s*=\s*"(?:\\"|[^"])+"\s*\.\s*split\("\W+"\))(?=\s*[,;])',
-            jsi.code, 'useful values', default='')
-
         func_name = self._extract_n_function_name(jsi.code)
 
-        func_code = jsi.extract_function_code(func_name)
-        if var_ay:
-            func_code = (func_code[0], ';\n'.join((var_ay, func_code[1])))
+        func_code = self._extract_sig_fn(jsi, func_name)
 
         if player_id:
             self.cache.store('youtube-nsig', player_id, func_code)
@@ -2136,7 +2149,8 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
                     video_details = merge_dicts(*traverse_obj(
                         (player_response, api_player_response),
                         (Ellipsis, 'videoDetails', T(dict))))
-                    player_response.update(api_player_response or {})
+                    player_response.update(filter_dict(
+                        api_player_response or {}, cndn=lambda k, _: k != 'captions'))
                     player_response['videoDetails'] = video_details
 
         def is_agegated(playability):
@@ -2566,8 +2580,8 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
         }
 
         pctr = traverse_obj(
-            player_response,
-            ('captions', 'playerCaptionsTracklistRenderer', T(dict)))
+            (player_response, api_player_response),
+            (Ellipsis, 'captions', 'playerCaptionsTracklistRenderer', T(dict)))
         if pctr:
             def process_language(container, base_url, lang_code, query):
                 lang_subs = []
@@ -2584,20 +2598,21 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
             def process_subtitles():
                 subtitles = {}
                 for caption_track in traverse_obj(pctr, (
-                        'captionTracks', lambda _, v: v.get('baseUrl'))):
+                        Ellipsis, 'captionTracks', lambda _, v: (
+                            v.get('baseUrl') and v.get('languageCode')))):
                     base_url = self._yt_urljoin(caption_track['baseUrl'])
                     if not base_url:
                         continue
+                    lang_code = caption_track['languageCode']
                     if caption_track.get('kind') != 'asr':
-                        lang_code = caption_track.get('languageCode')
-                        if not lang_code:
-                            continue
                         process_language(
                             subtitles, base_url, lang_code, {})
                         continue
                     automatic_captions = {}
+                    process_language(
+                        automatic_captions, base_url, lang_code, {})
                     for translation_language in traverse_obj(pctr, (
-                            'translationLanguages', lambda _, v: v.get('languageCode'))):
+                            Ellipsis, 'translationLanguages', lambda _, v: v.get('languageCode'))):
                         translation_language_code = translation_language['languageCode']
                         process_language(
                             automatic_captions, base_url, translation_language_code,

youtube_dl/jsinterp.py

@@ -678,7 +678,7 @@ class JSInterpreter(object):
             return len(obj)
         try:
             return obj[int(idx)] if isinstance(obj, list) else obj[compat_str(idx)]
-        except (TypeError, KeyError, IndexError) as e:
+        except (TypeError, KeyError, IndexError, ValueError) as e:
             # allow_undefined is None gives correct behaviour
             if allow_undefined or (
                     allow_undefined is None and not isinstance(e, TypeError)):
@@ -1038,6 +1038,10 @@ class JSInterpreter(object):
                     left_val = self._index(left_val, idx)
             if isinstance(idx, float):
                 idx = int(idx)
+            if isinstance(left_val, list) and len(left_val) <= int_or_none(idx, default=-1):
+                # JS Array is a sparsely assignable list
+                # TODO: handle extreme sparsity without memory bloat, eg using auxiliary dict
+                left_val.extend((idx - len(left_val) + 1) * [JS_Undefined])
             left_val[idx] = self._operator(
                 m.group('op'), self._index(left_val, idx) if m.group('op') else None,
                 m.group('expr'), expr, local_vars, allow_recursion)
@@ -1204,9 +1208,10 @@ class JSInterpreter(object):
                 elif member == 'join':
                     assertion(isinstance(obj, list), 'must be applied on a list')
                     assertion(len(argvals) <= 1, 'takes at most one argument')
-                    return (',' if len(argvals) == 0 else argvals[0]).join(
-                        ('' if x in (None, JS_Undefined) else _js_toString(x))
-                        for x in obj)
+                    return (',' if len(argvals) == 0 or argvals[0] in (None, JS_Undefined)
+                            else argvals[0]).join(
+                                ('' if x in (None, JS_Undefined) else _js_toString(x))
+                                for x in obj)
                 elif member == 'reverse':
                     assertion(not argvals, 'does not take any arguments')
                     obj.reverse()
@@ -1364,19 +1369,21 @@ class JSInterpreter(object):
         code, _ = self._separate_at_paren(func_m.group('code'))  # refine the match
         return self.build_arglist(func_m.group('args')), code
 
-    def extract_function(self, funcname):
+    def extract_function(self, funcname, *global_stack):
         return function_with_repr(
-            self.extract_function_from_code(*self.extract_function_code(funcname)),
+            self.extract_function_from_code(*itertools.chain(
+                self.extract_function_code(funcname), global_stack)),
             'F<%s>' % (funcname,))
 
     def extract_function_from_code(self, argnames, code, *global_stack):
         local_vars = {}
 
+        start = None
         while True:
-            mobj = re.search(r'function\((?P<args>[^)]*)\)\s*{', code)
+            mobj = re.search(r'function\((?P<args>[^)]*)\)\s*{', code[start:])
             if mobj is None:
                 break
-            start, body_start = mobj.span()
+            start, body_start = ((start or 0) + x for x in mobj.span())
             body, remaining = self._separate_at_paren(code[body_start - 1:])
             name = self._named_object(local_vars, self.extract_function_from_code(
                 [x.strip() for x in mobj.group('args').split(',')],