Merge 457f046c3ca82a2123eac2a80518d87287d828ad into 2b4fbfce25902d557b86b003cf48f738129efce4

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

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
+        assert m, '{0!r} should follow URL format'.format(url)
-        test_id = m.group('id')
+        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/dropout.py

@@ -0,0 +1,241 @@
+# coding: utf-8
+from __future__ import unicode_literals
+
+from .common import InfoExtractor
+from .vimeo import VHXEmbedIE, VimeoIE
+from ..utils import (
+    clean_html,
+    ExtractorError,
+    get_element_by_class,
+    get_element_by_id,
+    get_elements_by_class,
+    int_or_none,
+    unified_strdate,
+    urlencode_postdata
+)
+
+
+class DropoutIE(InfoExtractor):
+    _LOGIN_URL = 'https://www.dropout.tv/login'
+    _NETRC_MACHINE = 'dropout'
+
+    _VALID_URL = r'https?://(?:www\.)?dropout\.tv/(?:[^/]+/)*videos/(?P<id>[^/]+)/?$'
+    _TESTS = [
+        {
+            'url': 'https://www.dropout.tv/dimension-20-misfits-and-magic/season:2/videos/misfits-and-magic-holiday-special-trailer',
+            'note': 'No login required',
+            'md5': 'cafb8d704af8da70134d70a97d966799',
+            'info_dict': {
+                'id': '1893157',
+                'ext': 'mp4',
+                'title': 'Dimension 20: Misfits and Magic Holiday Special Trailer',
+                'description': 'Independent. Funny. Ad Free.',
+                'uploader': 'OTT Videos',
+                'uploader_id': 'user80538407',
+            },
+            'expected_warnings': ['No login information available']
+        },
+        {
+            'url': 'https://www.dropout.tv/game-changer/season:2/videos/yes-or-no',
+            'note': 'Episode in a series',
+            'md5': '5e000fdfd8d8fa46ff40456f1c2af04a',
+            'info_dict': {
+                'id': '738153',
+                'display_id': 'yes-or-no',
+                'ext': 'mp4',
+                'title': 'Yes or No',
+                'description': 'Ally, Brennan, and Zac are asked a simple question, but is there a correct answer?',
+                # 'release_date': '20200508',  # Release dates seem to have been removed from the website
+                'thumbnail': 'https://vhx.imgix.net/chuncensoredstaging/assets/351e3f24-c4a3-459a-8b79-dc80f1e5b7fd.jpg',
+                'series': 'Game Changer',
+                'season_number': 2,
+                'season': 'Season 2',
+                'episode_number': 6,
+                'episode': 'Yes or No',
+                'duration': 1180,
+                'uploader_id': 'user80538407',
+                'uploader_url': 'https://vimeo.com/user80538407',
+                'uploader': 'OTT Videos'
+            },
+            'expected_warnings': ['Ignoring subtitle tracks found in the HLS manifest'],
+            'skip': 'Username and password required',
+        },
+        {
+            'url': 'https://www.dropout.tv/dimension-20-fantasy-high/season:1/videos/episode-1',
+            'note': 'Episode in a series (missing release_date)',
+            'md5': '712caf7c191f1c47c8f1879520c2fa5c',
+            'info_dict': {
+                'id': '320562',
+                'display_id': 'episode-1',
+                'ext': 'mp4',
+                'title': 'The Beginning Begins',
+                'description': 'The cast introduces their PCs, including a neurotic elf, a goblin PI, and a corn-worshipping cleric.',
+                'thumbnail': 'https://vhx.imgix.net/chuncensoredstaging/assets/4421ed0d-f630-4c88-9004-5251b2b8adfa.jpg',
+                'series': 'Dimension 20: Fantasy High',
+                'season_number': 1,
+                'season': 'Season 1',
+                'episode_number': 1,
+                'episode': 'The Beginning Begins',
+                'duration': 6838,
+                'uploader_id': 'user80538407',
+                'uploader_url': 'https://vimeo.com/user80538407',
+                'uploader': 'OTT Videos'
+            },
+            'expected_warnings': ['Ignoring subtitle tracks found in the HLS manifest'],
+            'skip': 'Username and password required',
+        },
+        {
+            'url': 'https://www.dropout.tv/videos/misfits-magic-holiday-special',
+            'note': 'Episode not in a series',
+            'md5': '1cedb55910c0367c02d9d0aae524398e',
+            'info_dict': {
+                'id': '1915774',
+                'display_id': 'misfits-magic-holiday-special',
+                'ext': 'mp4',
+                'title': 'Misfits & Magic Holiday Special',
+                'description': 'The magical misfits spend Christmas break at Gowpenny, with an unwelcome visitor.',
+                # 'release_date': '20211215',
+                'thumbnail': 'https://vhx.imgix.net/chuncensoredstaging/assets/d91ea8a6-b250-42ed-907e-b30fb1c65176-8e24b8e5.jpg',
+                'duration': 11698,
+                'uploader_id': 'user80538407',
+                'uploader_url': 'https://vimeo.com/user80538407',
+                'uploader': 'OTT Videos'
+            },
+            'expected_warnings': ['Ignoring subtitle tracks found in the HLS manifest'],
+            'skip': 'Username and password required',
+        },
+    ]
+
+    def _get_authenticity_token(self, display_id):
+        signin_page = self._download_webpage(
+            self._LOGIN_URL, display_id, note='Getting authenticity token')
+        return self._html_search_regex(
+            r'name=["\']authenticity_token["\'] value=["\'](.+?)["\']',
+            signin_page, 'authenticity_token')
+
+    def _login(self, display_id):
+        username, password = self._get_login_info()
+        if not (username and password):
+            # self.raise_login_required()
+            self.report_warning('No login information available', display_id)
+            return False
+
+        response = self._download_webpage(
+            self._LOGIN_URL, display_id, note='Logging in', data=urlencode_postdata({
+                'email': username,
+                'password': password,
+                'authenticity_token': self._get_authenticity_token(display_id),
+                'utf8': True
+            }))
+
+        user_has_subscription = self._search_regex(
+            r'user_has_subscription:\s*["\'](.+?)["\']', response, 'subscription status', default='none')
+        if user_has_subscription.lower() == 'true':
+            return response
+        elif user_has_subscription.lower() == 'false':
+            raise ExtractorError('Account is not subscribed')
+        else:
+            raise ExtractorError('Incorrect username/password')
+
+    def _real_extract(self, url):
+        display_id = self._match_id(url)
+        try:
+            logged_in = self._login(display_id)
+            webpage = self._download_webpage(url, display_id, note='Downloading video webpage')
+        except ExtractorError:
+            if logged_in is False:
+                self.raise_login_required()
+            raise
+        finally:
+            if logged_in is not False:
+                self._download_webpage('https://www.dropout.tv/logout', display_id, note='Logging out')
+
+        embed_url = self._search_regex(r'embed_url:\s*["\'](.+?)["\']', webpage, 'embed url')
+        thumbnail = self._og_search_thumbnail(webpage)
+        watch_info = get_element_by_id('watch-info', webpage) or ''
+
+        title = clean_html(get_element_by_class('video-title', watch_info))
+        season_episode = get_element_by_class(
+            'site-font-secondary-color', get_element_by_class('text', watch_info))
+        episode_number = int_or_none(self._search_regex(
+            r'Episode (\d+)', season_episode or '', 'episode', default=None))
+
+        return {
+            '_type': 'url_transparent',
+            'ie_key': VHXEmbedIE.ie_key(),
+            'url': VimeoIE._smuggle_referrer(embed_url, 'https://www.dropout.tv'),
+            'id': self._search_regex(r'embed.vhx.tv/videos/(.+?)\?', embed_url, 'id'),
+            'display_id': display_id,
+            'title': title,
+            'description': self._html_search_meta('description', webpage, fatal=False),
+            'thumbnail': thumbnail.split('?')[0] if thumbnail else None,  # Ignore crop/downscale
+            'series': clean_html(get_element_by_class('series-title', watch_info)),
+            'episode_number': episode_number,
+            'episode': title if episode_number else None,
+            'season_number': int_or_none(self._search_regex(
+                r'Season (\d+),', season_episode or '', 'season', default=None)),
+            'release_date': unified_strdate(self._search_regex(
+                r'data-meta-field-name=["\']release_dates["\'] data-meta-field-value=["\'](.+?)["\']',
+                watch_info, 'release date', default=None)),
+        }
+
+
+class DropoutSeasonIE(InfoExtractor):
+    _VALID_URL = r'https?://(?:www\.)?dropout\.tv/(?P<id>[^\/$&?#]+)(?:/?$|/season:[0-9]+/?$)'
+    _TESTS = [
+        {
+            'url': 'https://www.dropout.tv/dimension-20-fantasy-high/season:1',
+            'note': 'Multi-season series with the season in the url',
+            'playlist_count': 17,
+            'info_dict': {
+                'id': 'dimension-20-fantasy-high-season-1',
+                'title': 'Dimension 20 Fantasy High - Season 1'
+            }
+        },
+        {
+            'url': 'https://www.dropout.tv/dimension-20-fantasy-high',
+            'note': 'Multi-season series with the season not in the url',
+            'playlist_count': 17,
+            'info_dict': {
+                'id': 'dimension-20-fantasy-high-season-1',
+                'title': 'Dimension 20 Fantasy High - Season 1'
+            }
+        },
+        {
+            'url': 'https://www.dropout.tv/dimension-20-shriek-week',
+            'note': 'Single-season series',
+            'playlist_count': 4,
+            'info_dict': {
+                'id': 'dimension-20-shriek-week-season-1',
+                'title': 'Dimension 20 Shriek Week - Season 1'
+            }
+        }
+    ]
+
+    def _real_extract(self, url):
+        season_id = self._match_id(url)
+        season_title = season_id.replace('-', ' ').title()
+        webpage = self._download_webpage(url, season_id)
+
+        entries = [
+            self.url_result(
+                url=self._search_regex(r'<a href=["\'](.+?)["\'] class=["\']browse-item-link["\']',
+                                       item, 'item_url'),
+                ie=DropoutIE.ie_key()
+            ) for item in get_elements_by_class('js-collection-item', webpage)
+        ]
+
+        seasons = (get_element_by_class('select-dropdown-wrapper', webpage) or '').strip().replace('\n', '')
+        current_season = self._search_regex(r'<option[^>]+selected>([^<]+)</option>',
+                                            seasons, 'current_season', default='').strip()
+
+        def join_nonempty(*args, **kwargs):
+            delim = kwargs.get('delim', '-')
+            return delim.join(x for x in args if x)
+
+        return {
+            '_type': 'playlist',
+            'id': join_nonempty(season_id, current_season.lower().replace(' ', '-')),
+            'title': join_nonempty(season_title, current_season, delim=' - '),
+            'entries': entries
+        }

youtube_dl/extractor/extractors.py

@@ -335,6 +335,10 @@ from .discoveryvr import DiscoveryVRIE
 from .disney import DisneyIE
 from .dispeak import DigitallySpeakingIE
 from .dropbox import DropboxIE
+from .dropout import (
+    DropoutSeasonIE,
+    DropoutIE
+)
 from .dw import (
     DWIE,
     DWArticleIE,

youtube_dl/extractor/vimeo.py

@@ -1171,8 +1171,15 @@ class VHXEmbedIE(VimeoBaseInfoExtractor):
         return unescapeHTML(mobj.group(1)) if mobj else None
 
     def _real_extract(self, url):
+        url, data = unsmuggle_url(url, {})
+        headers = std_headers.copy()
+        if 'http_headers' in data:
+            headers.update(data['http_headers'])
+        if 'Referer' not in headers:
+            headers['Referer'] = url
+
         video_id = self._match_id(url)
-        webpage = self._download_webpage(url, video_id)
+        webpage = self._download_webpage(url, video_id, headers=headers)
         config_url = self._parse_json(self._search_regex(
             r'window\.OTTData\s*=\s*({.+})', webpage,
             'ott data'), video_id, js_to_json)['config_url']

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'/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$',
+        r'/(?P<id>[a-zA-Z0-9_-]{8,})/player(?:_ias(?:_tce)?\.vflset(?:/[a-zA-Z]{2,3}_[a-zA-Z]{2,3})?|-plasma-ias-(?:phone|tablet)-[a-z]{2}_[A-Z]{2}\.vflset)/base\.js$',
-        r'\b(?P<id>vfl[a-zA-Z0-9_-]+)\b.*?\.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(self, player_url):
-    def _extract_player_info(cls, player_url):
+        try:
-        for player_re in cls._PLAYER_INFO_RE:
+            return self._search_regex(
-            id_m = re.search(player_re, player_url)
+                self._PLAYER_INFO_RE, player_url, 'player info', group='id')
-            if id_m:
+        except ExtractorError as e:
-                break
+            raise ExtractorError(
-        else:
+                'Cannot identify player %r' % (player_url,), cause=e)
-            raise ExtractorError('Cannot identify player %r' % player_url)
-        return id_m.group('id')
 
     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)
+        func_code = self._extract_sig_fn(jsi, func_name)
-        if var_ay:
-            func_code = (func_code[0], ';\n'.join((var_ay, func_code[1])))
 
         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,
+            (player_response, api_player_response),
-            ('captions', 'playerCaptionsTracklistRenderer', T(dict)))
+            (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)
@@ -1204,7 +1208,8 @@ 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(
+                    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':
@@ -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(',')],