Merge efad43c197775eb3ef30b048c8d7b83c1ccbb57c 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/joj.py

@@ -2,6 +2,7 @@
 from __future__ import unicode_literals
 
 import re
+import json
 
 from .common import InfoExtractor
 from ..compat import compat_str
@@ -21,7 +22,8 @@ class JojIE(InfoExtractor):
                     (?P<id>[^/?#^]+)
                 '''
     _TESTS = [{
-        'url': 'https://media.joj.sk/embed/a388ec4c-6019-4a4a-9312-b1bee194e932',
+        'url':
+        'https://media.joj.sk/embed/a388ec4c-6019-4a4a-9312-b1bee194e932',
         'info_dict': {
             'id': 'a388ec4c-6019-4a4a-9312-b1bee194e932',
             'ext': 'mp4',
@@ -64,16 +66,20 @@ class JojIE(InfoExtractor):
                 r'(?s)(?:src|bitrates)\s*=\s*({.+?});', webpage, 'bitrates',
                 default='{}'),
             video_id, transform_source=js_to_json, fatal=False)
+        quality = self._search_regex(
+            r'var qualityLabels = ({.*?});', webpage, 'quality', default=None)
+        quality = json.loads(quality)
 
         formats = []
         for format_url in try_get(bitrates, lambda x: x['mp4'], list) or []:
             if isinstance(format_url, compat_str):
                 height = self._search_regex(
-                    r'(\d+)[pP]\.', format_url, 'height', default=None)
+                    r'-(\d+)p?\.', format_url, 'height', default=None)
                 formats.append({
                     'url': format_url,
                     'format_id': '%sp' % height if height else None,
                     'height': int(height),
+                    'resolution': quality.get(compat_str(height)+'p')
                 })
         if not formats:
             playlist = self._download_xml(
@@ -104,5 +110,5 @@ class JojIE(InfoExtractor):
             'title': title,
             'thumbnail': thumbnail,
             'duration': duration,
-            'formats': formats,
+            'formats': formats
         }

youtube_dl/extractor/markiza.py

@@ -1,21 +1,23 @@
 # coding: utf-8
 from __future__ import unicode_literals
 
+# update 15.01.2022 jastrab
 import re
+import json
 
 from .common import InfoExtractor
-from ..compat import compat_str
 from ..utils import (
     orderedSet,
-    parse_duration,
-    try_get,
+    url_or_none,
+    determine_ext
 )
 
 
 class MarkizaIE(InfoExtractor):
-    _VALID_URL = r'https?://(?:www\.)?videoarchiv\.markiza\.sk/(?:video/(?:[^/]+/)*|embed/)(?P<id>\d+)(?:[_/]|$)'
+    _VALID_URL = r'https:\/\/(?:www\.)?videoarchiv\.markiza\.sk\/(?:video\/(?:[^\/]+\/)*|embed\/)epizoda\/(?P<id>\d+)(?:[\_\/\-]|$)'
     _TESTS = [{
-        'url': 'http://videoarchiv.markiza.sk/video/oteckovia/84723_oteckovia-109',
+        'url': 'http://videoarchiv.markiza.sk/video/oteckovia/\
+            84723_oteckovia-109',
         'md5': 'ada4e9fad038abeed971843aa028c7b0',
         'info_dict': {
             'id': '139078',
@@ -26,54 +28,89 @@ class MarkizaIE(InfoExtractor):
             'duration': 2760,
         },
     }, {
-        'url': 'http://videoarchiv.markiza.sk/video/televizne-noviny/televizne-noviny/85430_televizne-noviny',
+        'url': ' https://videoarchiv.markiza.sk/video/laska-na-prenajom/epizoda/58779-seria-1-epizoda-14',
         'info_dict': {
             'id': '85430',
             'title': 'Televízne noviny',
         },
         'playlist_count': 23,
     }, {
-        'url': 'http://videoarchiv.markiza.sk/video/oteckovia/84723',
+        'url': 'https://videoarchiv.markiza.sk/video/oteckovia/84723',
         'only_matching': True,
     }, {
-        'url': 'http://videoarchiv.markiza.sk/video/84723',
+        'url': 'https://videoarchiv.markiza.sk/video/84723',
         'only_matching': True,
     }, {
-        'url': 'http://videoarchiv.markiza.sk/video/filmy/85190_kamenak',
+        'url': 'https://videoarchiv.markiza.sk/video/filmy/85190_kamenak',
         'only_matching': True,
     }, {
-        'url': 'http://videoarchiv.markiza.sk/video/reflex/zo-zakulisia/84651_pribeh-alzbetky',
+        'url': 'https://videoarchiv.markiza.sk/video/reflex/zo-zakulisia/84651_pribeh-alzbetky',
         'only_matching': True,
     }, {
-        'url': 'http://videoarchiv.markiza.sk/embed/85295',
+        'url': 'https://videoarchiv.markiza.sk/embed/85295',
         'only_matching': True,
     }]
 
     def _real_extract(self, url):
         video_id = self._match_id(url)
+        webpage = self._download_webpage(url, video_id)
+        embed = self._search_regex(
+            r'<iframe src="(https:\/\/media.*?)" style="',
+            webpage, 'embed', fatal=False)
+        webpage = self._download_webpage(embed, video_id)
+        data = self._search_regex(
+            r'processAdTagModifier\((\{.*)\), {"video":',
+            webpage, 'embed', fatal=False)
+        data2 = self._search_regex(
+            r'processAdTagModifier\(\{.*\), ({"video":.*)\);',
+            webpage, 'embed', fatal=False)
+        data = re.sub('\\\\/', '/', data)
+        data2 = re.sub('\\\\/', '/', data2)
+        info = json.loads(data)
+        info2 = json.loads(data2)
 
-        data = self._download_json(
-            'http://videoarchiv.markiza.sk/json/video_jwplayer7.json',
-            video_id, query={'id': video_id})
-
-        info = self._parse_jwplayer_data(data, m3u8_id='hls', mpd_id='dash')
-
-        if info.get('_type') == 'playlist':
-            info.update({
-                'id': video_id,
-                'title': try_get(
-                    data, lambda x: x['details']['name'], compat_str),
-            })
-        else:
-            info['duration'] = parse_duration(
-                try_get(data, lambda x: x['details']['duration'], compat_str))
-        return info
+        formats = []
+        for format_id, format_list in info['tracks'].items():
+            if not isinstance(format_list, list):
+                format_list = [format_list]
+            for format_dict in format_list:
+                if not isinstance(format_dict, dict):
+                    continue
+                format_url = url_or_none(format_dict.get('src'))
+                format_type = format_dict.get('type')
+                ext = determine_ext(format_url)
+                if (format_type == 'application/x-mpegURL'
+                        or format_id == 'HLS' or ext == 'm3u8'):
+                    formats.extend(self._extract_m3u8_formats(
+                        format_url, video_id, 'mp4',
+                        entry_protocol='m3u8_native', m3u8_id='hls',
+                        fatal=False))
+                elif (format_type == 'application/dash+xml'
+                      or format_id == 'DASH' or ext == 'mpd'):
+                    formats.extend(self._extract_mpd_formats(
+                        format_url, video_id, mpd_id='dash', fatal=False))
+                else:
+                    formats.append({
+                        'url': format_url,
+                    })
+        thumbnail = info.get('plugins').get('thumbnails').get('url')
+        thumbnail = re.sub('$Num$', '001', thumbnail)
+        duration = info.get('duration')
+        title2 = info2.get('video').get('title')
+        title = info2.get('video').get('custom').get('show_title')
+        return {
+            'id': video_id,
+            'title': title + ' - ' + title2,
+            'thumbnail': thumbnail,
+            'duration': duration,
+            'formats': formats
+        }
 
 
 class MarkizaPageIE(InfoExtractor):
     _VALID_URL = r'https?://(?:www\.)?(?:(?:[^/]+\.)?markiza|tvnoviny)\.sk/(?:[^/]+/)*(?P<id>\d+)_'
     _TESTS = [{
-        'url': 'http://www.markiza.sk/soubiz/zahranicny/1923705_oteckovia-maju-svoj-den-ti-slavni-nie-su-o-nic-menej-rozkosni',
+        'url': 'https://www.markiza.sk/soubiz/zahranicny/1923705_oteckovia-maju-svoj-den-ti-slavni-nie-su-o-nic-menej-rozkosni',
         'md5': 'ada4e9fad038abeed971843aa028c7b0',
         'info_dict': {
             'id': '139355',
@@ -87,29 +124,30 @@ class MarkizaPageIE(InfoExtractor):
             'skip_download': True,
         },
     }, {
-        'url': 'http://dajto.markiza.sk/filmy-a-serialy/1774695_frajeri-vo-vegas',
+        'url': 'https://dajto.markiza.sk/filmy-a-serialy/1774695_frajeri-vo-vegas',
         'only_matching': True,
     }, {
-        'url': 'http://superstar.markiza.sk/aktualne/1923870_to-je-ale-telo-spevacka-ukazala-sexy-postavicku-v-bikinach',
+        'url': 'https://superstar.markiza.sk/aktualne/\
+        1923870_to-je-ale-telo-spevacka-ukazala-sexy-postavicku-v-bikinach',
         'only_matching': True,
     }, {
-        'url': 'http://hybsa.markiza.sk/aktualne/1923790_uzasna-atmosfera-na-hybsa-v-poprade-superstaristi-si-prve-koncerty-pred-davom-ludi-poriadne-uzili',
+        'url': 'https://hybsa.markiza.sk/aktualne/1923790_uzasna-atmosfera-na-hybsa-v-poprade-superstaristi-si-prve-koncerty-pred-davom-ludi-poriadne-uzili',
         'only_matching': True,
     }, {
-        'url': 'http://doma.markiza.sk/filmy/1885250_moja-vysnivana-svadba',
+        'url': 'https://doma.markiza.sk/filmy/1885250_moja-vysnivana-svadba',
         'only_matching': True,
     }, {
-        'url': 'http://www.tvnoviny.sk/domace/1923887_po-smrti-manzela-ju-cakalo-poriadne-prekvapenie',
+        'url': 'https://www.tvnoviny.sk/domace/1923887_po-smrti-manzela-ju-cakalo-poriadne-prekvapenie',
         'only_matching': True,
     }]
 
     @classmethod
     def suitable(cls, url):
-        return False if MarkizaIE.suitable(url) else super(MarkizaPageIE, cls).suitable(url)
+        return False if MarkizaIE.suitable(url) else \
+            super(MarkizaPageIE, cls).suitable(url)
 
     def _real_extract(self, url):
         playlist_id = self._match_id(url)
-
         webpage = self._download_webpage(
             # Downloading for some hosts (e.g. dajto, doma) fails with 500
             # although everything seems to be OK, so considering 500
@@ -117,7 +155,8 @@ class MarkizaPageIE(InfoExtractor):
             url, playlist_id, expected_status=500)
 
         entries = [
-            self.url_result('http://videoarchiv.markiza.sk/video/%s' % video_id)
+            self.url_result(
+                'http://videoarchiv.markiza.sk/video/%s' % video_id)
             for video_id in orderedSet(re.findall(
                 r'(?:initPlayer_|data-entity=["\']|id=["\']player_)(\d+)',
                 webpage))]

youtube_dl/extractor/rtvs.py

@@ -1,7 +1,13 @@
 # coding: utf-8
 from __future__ import unicode_literals
 
+import re
+
 from .common import InfoExtractor
+from ..utils import (
+    url_or_none,
+    determine_ext
+)
 
 
 class RTVSIE(InfoExtractor):
@@ -26,7 +32,8 @@ class RTVSIE(InfoExtractor):
             'id': '63118',
             'ext': 'mp4',
             'title': 'Amaro Džives - Náš deň',
-            'description': 'Galavečer pri príležitosti Medzinárodného dňa Rómov.'
+            'description':
+            'Galavečer pri príležitosti Medzinárodného dňa Rómov.'
         },
         'params': {
             'skip_download': True,
@@ -36,12 +43,66 @@ class RTVSIE(InfoExtractor):
     def _real_extract(self, url):
         video_id = self._match_id(url)
 
-        webpage = self._download_webpage(url, video_id)
+        if url.find('/radio/') != -1:
+            a2 = url.split('/')[-1]
+            a1 = url.split('/')[-2]
+            embed = self._download_webpage(
+                "https://www.rtvs.sk/embed/radio/archive/%s/%s" % (a1, a2),
+                video_id)
+            audio_id = re.search('audio5f.json?id=(?P<id>[^\"]+)', embed)
+            audio_id = audio_id.group('id')
+            info = self._download_json(
+                "https://www.rtvs.sk/json/audio5f.json?id=%s" % audio_id,
+                audio_id)
 
-        playlist_url = self._search_regex(
-            r'playlist["\']?\s*:\s*(["\'])(?P<url>(?:(?!\1).)+)\1', webpage,
-            'playlist url', group='url')
+            formats = []
+            formats.append({
+                'url': info['playlist'][0]['sources'][0]['src'],
+                'format_id':  None,
+                'height': 0
+                })
+            info = info['playlist'][0]
+            return {
+                'id': audio_id,
+                'title': info.get('title'),
+                'thumbnail': info.get('image'),
+                'formats': formats
+            }
+        else:
+            info = self._download_json(
+                "https://www.rtvs.sk/json/archive5f.json?id=%s" % video_id,
+                video_id)
+            info = info.get('clip')
 
-        data = self._download_json(
-            playlist_url, video_id, 'Downloading playlist')[0]
-        return self._parse_jwplayer_data(data, video_id=video_id)
+            formats = []
+            for format_id, format_list in info.items():
+                if not isinstance(format_list, list):
+                    format_list = [format_list]
+                for format_dict in format_list:
+                    if not isinstance(format_dict, dict):
+                        continue
+                    format_url = url_or_none(format_dict.get('src'))
+                    format_type = format_dict.get('type')
+                    ext = determine_ext(format_url)
+                    if (format_type == 'application/x-mpegURL'
+                            or format_id == 'HLS' or ext == 'm3u8'):
+                        formats.extend(self._extract_m3u8_formats(
+                            format_url, video_id, 'mp4',
+                            entry_protocol='m3u8_native', m3u8_id='hls',
+                            fatal=False))
+                    elif (format_type == 'application/dash+xml'
+                          or format_id == 'DASH' or ext == 'mpd'):
+                        pass
+                    else:
+                        formats.append({
+                            'url': format_url,
+                        })
+            formats = sorted(formats, key=lambda i: i['tbr'])
+            dt = info.get('datetime_create')
+            return {
+                'id': video_id,
+                'title': info.get('title') + '-' + dt[:10],
+                'thumbnail': info.get('image'),
+                'description': info.get('description'),
+                'formats': formats
+            }

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(',')],