Merge 7ac17d4b4469d23e575c3483695a790a309b2ff8 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
-        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

@@ -1274,7 +1274,10 @@ from .telequebec import (
 from .teletask import TeleTaskIE
 from .telewebion import TelewebionIE
 from .tennistv import TennisTVIE
-from .tenplay import TenPlayIE
+from .tenplay import (
+    TenPlayIE,
+    TenPlaySeasonIE,
+)
 from .testurl import TestURLIE
 from .tf1 import TF1IE
 from .tfo import TFOIE

youtube_dl/extractor/tenplay.py

@@ -1,70 +1,203 @@
 # coding: utf-8
 from __future__ import unicode_literals
 
+import functools
+
 from .common import InfoExtractor
+from ..compat import (
+    compat_itertools_count,
+    compat_kwargs,
+)
 from ..utils import (
+    get_elements_by_attribute,
     HEADRequest,
-    parse_age_limit,
-    parse_iso8601,
-    # smuggle_url,
+    float_or_none,
+    int_or_none,
+    merge_dicts,
+    NO_DEFAULT,
+    T,
+    traverse_obj,
+    txt_or_none,
+    url_or_none,
+    urljoin,
 )
 
 
-class TenPlayIE(InfoExtractor):
+class TenPlayBase(InfoExtractor):
+    _GEO_COUNTRIES = ['AU']
+    _GEO_BYPASS = False
+
+    def raise_geo_restricted(self, *args, **kwargs):
+        countries = args[1] if len(args) > 1 else kwargs.get('countries', NO_DEFAULT)
+        if countries is NO_DEFAULT:
+            kwargs['countries'] = self._GEO_COUNTRIES
+            kwargs = compat_kwargs(kwargs)
+        super(TenPlayBase, self).raise_geo_restricted(*args, **kwargs)
+
+    def _download_webpage_handle(self, url_or_request, video_id, *args, **kwargs):
+        res = super(TenPlayBase, self)._download_webpage_handle(url_or_request, video_id, *args, **kwargs)
+        if res and any('Sorry, 10 play is not available in your region.' in e
+                       for e in get_elements_by_attribute('class', 'iserror__text', res[0])):
+            self.raise_geo_restricted()
+        return res
+
+
+class TenPlayIE(TenPlayBase):
     _VALID_URL = r'https?://(?:www\.)?10play\.com\.au/(?:[^/]+/)+(?P<id>tpv\d{6}[a-z]{5})'
+    _NETRC_MACHINE = '10play'
     _TESTS = [{
-        'url': 'https://10play.com.au/masterchef/episodes/season-1/masterchef-s1-ep-1/tpv190718kwzga',
+        'url': 'https://10play.com.au/neighbours/web-extras/season-41/heres-a-first-look-at-mischa-bartons-neighbours-debut/tpv230911hyxnz',
         'info_dict': {
-            'id': '6060533435001',
+            'id': '6336940246112',
             'ext': 'mp4',
-            'title': 'MasterChef - S1 Ep. 1',
-            'description': 'md5:4fe7b78e28af8f2d900cd20d900ef95c',
-            'age_limit': 10,
-            'timestamp': 1240828200,
-            'upload_date': '20090427',
+            'title': 'Here\'s A First Look At Mischa Barton\'s Neighbours Debut',
+            'alt_title': 'Here\'s A First Look At Mischa Barton\'s Neighbours Debut',
+            'description': 'Neighbours Premieres Monday, September 18 At 4:30pm On 10 And 10 Play And 6:30pm On 10 Peach',
+            'duration': 74,
+            'season': 'Season 41',
+            'season_number': 41,
+            'series': 'Neighbours',
+            'thumbnail': r're:https://.*\.jpg',
+            'uploader': 'Channel 10',
+            'age_limit': 15,
+            'timestamp': 1694386800,
+            'upload_date': '20230910',
             'uploader_id': '2199827728001',
         },
         'params': {
-            # 'format': 'bestvideo',
             'skip_download': True,
-        }
+        },
+        'skip': 'Only available in Australia',
+    }, {
+        'url': 'https://10play.com.au/neighbours/episodes/season-42/episode-9107/tpv240902nzqyp',
+        'info_dict': {
+            'id': '9000000000091177',
+            'ext': 'mp4',
+            'title': 'Neighbours - S42 Ep. 9107',
+            'alt_title': 'Thu 05 Sep',
+            'description': 'md5:37a1f4271be34b9ee2b533426a5fbaef',
+            'duration': 1388,
+            'episode': 'Episode 9107',
+            'episode_number': 9107,
+            'season': 'Season 42',
+            'season_number': 42,
+            'series': 'Neighbours',
+            'thumbnail': r're:https://.*\.jpg',
+            'age_limit': 15,
+            'timestamp': 1725517860,
+            'upload_date': '20240905',
+            'uploader': 'Channel 10',
+            'uploader_id': '2199827728001',
+        },
+        'params': {
+            'skip_download': True,
+        },
+        'skip': 'Only available in Australia',
     }, {
         'url': 'https://10play.com.au/how-to-stay-married/web-extras/season-1/terrys-talks-ep-1-embracing-change/tpv190915ylupc',
         'only_matching': True,
     }]
-    # BRIGHTCOVE_URL_TEMPLATE = 'https://players.brightcove.net/2199827728001/cN6vRtRQt_default/index.html?videoId=%s'
-    _GEO_BYPASS = False
-    _FASTLY_URL_TEMPL = 'https://10-selector.global.ssl.fastly.net/s/kYEXFC/media/%s?mbr=true&manifest=m3u&format=redirect'
+
+    _AUS_AGES = {
+        'G': 0,
+        'PG': 15,
+        'M': 15,
+        'MA': 15,
+        'MA15+': 15,
+        'R': 18,
+        'X': 18,
+    }
 
     def _real_extract(self, url):
         content_id = self._match_id(url)
         data = self._download_json(
-            'https://10play.com.au/api/video/' + content_id, content_id)
-        video = data.get('video') or {}
-        metadata = data.get('metaData') or {}
-        brightcove_id = video.get('videoId') or metadata['showContentVideoId']
-        # brightcove_url = smuggle_url(
-        #     self.BRIGHTCOVE_URL_TEMPLATE % brightcove_id,
-        #     {'geo_countries': ['AU']})
-        m3u8_url = self._request_webpage(HEADRequest(
-            self._FASTLY_URL_TEMPL % brightcove_id), brightcove_id).geturl()
+            'https://10play.com.au/api/v1/videos/' + content_id, content_id)
+
+        video_data = self._download_json(
+            'https://vod.ten.com.au/api/videos/bcquery?command=find_videos_by_id&video_id={0}'.format(data['altId']),
+            content_id, 'Downloading video JSON')
+        m3u8_url = self._request_webpage(
+            HEADRequest(video_data['items'][0]['HLSURL']),
+            content_id, 'Checking stream URL').url
         if '10play-not-in-oz' in m3u8_url:
-            self.raise_geo_restricted(countries=['AU'])
-        formats = self._extract_m3u8_formats(m3u8_url, brightcove_id, 'mp4')
+            self.raise_geo_restricted()
+        # Attempt to get a higher quality stream
+        m3u8_url = m3u8_url.replace(',150,75,55,0000', ',300,150,75,55,0000')
+        formats = self._extract_m3u8_formats(m3u8_url, content_id, 'mp4')
         self._sort_formats(formats)
 
-        return {
-            # '_type': 'url_transparent',
-            # 'url': brightcove_url,
+        return merge_dicts({
+            'id': content_id,
             'formats': formats,
-            'id': brightcove_id,
-            'title': video.get('title') or metadata.get('pageContentName') or metadata['showContentName'],
-            'description': video.get('description'),
-            'age_limit': parse_age_limit(video.get('showRatingClassification') or metadata.get('showProgramClassification')),
-            'series': metadata.get('showName'),
-            'season': metadata.get('showContentSeason'),
-            'timestamp': parse_iso8601(metadata.get('contentPublishDate') or metadata.get('pageContentPublishDate')),
-            'thumbnail': video.get('poster'),
+            'uploader': 'Channel 10',
             'uploader_id': '2199827728001',
-            # 'ie_key': 'BrightcoveNew',
-        }
+        }, traverse_obj(data, {
+            'subtitles': ('captionUrl', T(lambda x: None if x is None
+                                          else {'en': [{'url': x}]})),
+            'id': ('altId', T(txt_or_none)),
+            'duration': ('duration', T(float_or_none)),
+            'title': ('subtitle', T(txt_or_none)),
+            'alt_title': ('title', T(txt_or_none)),
+            'description': ('description', T(txt_or_none)),
+            'age_limit': ('classification', T(self._AUS_AGES.get)),
+            'series': ('tvShow', T(txt_or_none)),
+            'season_number': ('season', T(int_or_none)),
+            'episode_number': ('episode', T(int_or_none)),
+            'timestamp': ('published', T(int_or_none)),
+            'thumbnail': ('imageUrl', T(url_or_none)),
+        }), rev=True)
+
+
+class TenPlaySeasonIE(TenPlayBase):
+    _VALID_URL = r'https?://(?:www\.)?10play\.com\.au/(?P<show>[^/?#]+)/episodes/(?P<season>[^/?#]+)/?(?:$|[?#])'
+    _TESTS = [{
+        'url': 'https://10play.com.au/masterchef/episodes/season-14',
+        'info_dict': {
+            'title': 'Season 14',
+            'id': 'MjMyOTIy',
+        },
+        'playlist_mincount': 64,
+        'skip': 'Only available in Australia',
+    }, {
+        'url': 'https://10play.com.au/the-bold-and-the-beautiful-fast-tracked/episodes/season-2022',
+        'info_dict': {
+            'title': 'Season 2022',
+            'id': 'Mjc0OTIw',
+        },
+        'playlist_mincount': 256,
+        'skip': 'Only available in Australia',
+    }]
+
+    def _entries(self, load_more_url, display_id=None):
+        skip_ids = []
+        for page in compat_itertools_count(1):
+            episodes_carousel = self._download_json(
+                load_more_url, display_id, query={'skipIds[]': skip_ids},
+                note='Fetching episodes page {0}'.format(page))
+
+            episodes_chunk = episodes_carousel['items']
+            skip_ids.extend(ep['id'] for ep in episodes_chunk)
+
+            for ep in episodes_chunk:
+                yield ep['cardLink']
+            if not episodes_carousel.get('hasMore'):
+                break
+
+    def _real_extract(self, url):
+        show, season = self._match_valid_url(url).group('show', 'season')
+        season_info = self._download_json(*(s.format(show=show, season=season) for s in (
+            'https://10play.com.au/api/shows/{show}/episodes/{season}', '{show}/{season}')
+        ))
+
+        episodes_carousel = traverse_obj(season_info, (
+            'content', 0, 'components', (
+                lambda _, v: v['title'].lower() == 'episodes',
+                (Ellipsis, T(dict)),
+            )), any) or {}
+
+        playlist_id = episodes_carousel['tpId']
+
+        return self.playlist_from_matches(
+            self._entries(urljoin(url, episodes_carousel['loadMoreUrl']), playlist_id),
+            playlist_id, traverse_obj(season_info, ('content', 0, 'title', T(txt_or_none))),
+            getter=functools.partial(urljoin, 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'/(?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(',')],