Merge ec8cc20805d856dd4b6e8e098bf007c12e4ec904 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/YoutubeDL.py

@@ -214,6 +214,7 @@ class YoutubeDL(object):
     writeannotations:  Write the video annotations to a .annotations.xml file
     writethumbnail:    Write the thumbnail image to a file
     write_all_thumbnails:  Write all thumbnail formats to files
+    writestoryboards:  Write all storyboards (grid of video frames) to a file
     writesubtitles:    Write the video subtitles to a file
     writeautomaticsub: Write the automatically generated subtitles to a file
     allsubtitles:      Downloads all the subtitles of the video
@@ -310,6 +311,7 @@ class YoutubeDL(object):
                        [sleep_interval; max_sleep_interval].
     listformats:       Print an overview of available video formats and exit.
     list_thumbnails:   Print a table of all thumbnails and exit.
+    list_storyboards:  Print a table of all storyboards and exit.
     match_filter:      A function that gets called with the info_dict of
                        every video.
                        If it returns a message, the video is ignored.
@@ -1693,6 +1695,10 @@ class YoutubeDL(object):
             self.list_thumbnails(info_dict)
             return
 
+        if self.params.get('list_storyboards'):
+            self.list_thumbnails(info_dict, item_name='storyboards')
+            return
+
         thumbnail = info_dict.get('thumbnail')
         if thumbnail:
             info_dict['thumbnail'] = sanitize_url(thumbnail)
@@ -2466,17 +2472,27 @@ class YoutubeDL(object):
             '[info] Available formats for %s:\n%s' %
             (info_dict['id'], render_table(header_line, table)))
 
-    def list_thumbnails(self, info_dict):
-        thumbnails = info_dict.get('thumbnails')
+    def list_thumbnails(self, info_dict, item_name='thumbnails'):
+        thumbnails = info_dict.get(item_name)
         if not thumbnails:
-            self.to_screen('[info] No thumbnails present for %s' % info_dict['id'])
+            self.to_screen('[info] No %s present for %s' % (item_name, info_dict['id']))
             return
 
         self.to_screen(
-            '[info] Thumbnails for %s:' % info_dict['id'])
-        self.to_screen(render_table(
-            ['ID', 'width', 'height', 'URL'],
-            [[t['id'], t.get('width', 'unknown'), t.get('height', 'unknown'), t['url']] for t in thumbnails]))
+            '[info] %s for %s:' % (item_name.title(), info_dict['id']))
+
+        columns = ['ID', 'width', 'height']
+        if item_name == 'storyboards':
+            columns += ['cols', 'rows', 'frames']
+        columns += ['URL']
+
+        table = []
+        for t in thumbnails:
+            table.append([])
+            for column in columns:
+                table[-1].append(t.get(column.lower(), 'unknown'))
+
+        self.to_screen(render_table(columns, table))
 
     def list_subtitles(self, video_id, subtitles, name='subtitles'):
         if not subtitles:
@@ -2679,12 +2695,16 @@ class YoutubeDL(object):
                 return
 
     def _write_thumbnails(self, info_dict, filename):
+        item_name = 'thumbnail'
         if self.params.get('writethumbnail', False):
             thumbnails = info_dict.get('thumbnails')
             if thumbnails:
                 thumbnails = [thumbnails[-1]]
         elif self.params.get('write_all_thumbnails', False):
             thumbnails = info_dict.get('thumbnails')
+        elif self.params.get('writestoryboards', False):
+            thumbnails = info_dict.get('storyboards')
+            item_name = 'storyboard'
         else:
             return
 
@@ -2694,22 +2714,28 @@ class YoutubeDL(object):
 
         for t in thumbnails:
             thumb_ext = determine_ext(t['url'], 'jpg')
-            suffix = '_%s' % t['id'] if len(thumbnails) > 1 else ''
+            if item_name == 'thumbnails':
+                suffix = '_%s' % t['id'] if len(thumbnails) > 1 else ''
+            else:
+                suffix = '_%s_%s' % (item_name, t['id'])
             thumb_display_id = '%s ' % t['id'] if len(thumbnails) > 1 else ''
             t['filename'] = thumb_filename = replace_extension(filename + suffix, thumb_ext, info_dict.get('ext'))
 
             if self.params.get('nooverwrites', False) and os.path.exists(encodeFilename(thumb_filename)):
-                self.to_screen('[%s] %s: Thumbnail %sis already present' %
-                               (info_dict['extractor'], info_dict['id'], thumb_display_id))
+                self.to_screen('[%s] %s: %s %sis already present' %
+                               (info_dict['extractor'], info_dict['id'],
+                                item_name.title(), thumb_display_id))
             else:
-                self.to_screen('[%s] %s: Downloading thumbnail %s...' %
-                               (info_dict['extractor'], info_dict['id'], thumb_display_id))
+                self.to_screen('[%s] %s: Downloading %s %s...' %
+                               (info_dict['extractor'], info_dict['id'],
+                                item_name, thumb_display_id))
                 try:
                     uf = self.urlopen(t['url'])
                     with open(encodeFilename(thumb_filename), 'wb') as thumbf:
                         shutil.copyfileobj(uf, thumbf)
-                    self.to_screen('[%s] %s: Writing thumbnail %sto: %s' %
-                                   (info_dict['extractor'], info_dict['id'], thumb_display_id, thumb_filename))
+                    self.to_screen('[%s] %s: Writing %s %sto: %s' %
+                                   (info_dict['extractor'], info_dict['id'],
+                                    item_name, thumb_display_id, thumb_filename))
                 except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
-                    self.report_warning('Unable to download thumbnail "%s": %s' %
-                                        (t['url'], error_to_compat_str(err)))
+                    self.report_warning('Unable to download %s "%s": %s' %
+                                        (t['url'], item_name, error_to_compat_str(err)))

youtube_dl/__init__.py

@@ -374,6 +374,7 @@ def _real_main(argv=None):
         'writeinfojson': opts.writeinfojson,
         'writethumbnail': opts.writethumbnail,
         'write_all_thumbnails': opts.write_all_thumbnails,
+        'writestoryboards': opts.writestoryboards,
         'writesubtitles': opts.writesubtitles,
         'writeautomaticsub': opts.writeautomaticsub,
         'allsubtitles': opts.allsubtitles,
@@ -421,6 +422,7 @@ def _real_main(argv=None):
         'max_sleep_interval': opts.max_sleep_interval,
         'external_downloader': opts.external_downloader,
         'list_thumbnails': opts.list_thumbnails,
+        'list_storyboards': opts.list_storyboards,
         'playlist_items': opts.playlist_items,
         'xattr_set_filesize': opts.xattr_set_filesize,
         'match_filter': match_filter,

youtube_dl/extractor/common.py

@@ -239,6 +239,12 @@ class InfoExtractor(object):
                                         deprecated)
                         * "filesize" (optional, int)
     thumbnail:      Full URL to a video thumbnail image.
+    storyboards:    A list of dictionaries representing storyboards.
+                    A storyboard is an image grid made of frames from the video.
+                    This has the same structure as the thumbnails list, plus:
+                        * "cols" (optional, int)
+                        * "rows" (optional, int)
+                        * "frames" (optional, int)
     description:    Full video description.
     uploader:       Full name of the video uploader.
     license:        License name the video is licensed under.

youtube_dl/extractor/youtube.py

@@ -12,6 +12,7 @@ import re
 import string
 import time
 import traceback
+import math
 
 from .common import InfoExtractor, SearchInfoExtractor
 from ..compat import (
@@ -692,9 +693,9 @@ 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 +1627,13 @@ 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 +1710,23 @@ 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 +1752,12 @@ 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 +1876,9 @@ 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 +2150,8 @@ 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):
@@ -2514,6 +2529,58 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
             if thumbnail:
                 thumbnails = [{'url': thumbnail}]
 
+        storyboards = []
+        sb_spec = try_get(player_response,
+                          lambda x: x['storyboards']['playerStoryboardSpecRenderer']['spec'],
+                          compat_str)
+        if sb_spec:
+            s_parts = sb_spec.split('|')
+            base_url = s_parts[0]
+            for i, params in enumerate(s_parts[1:]):
+                storyboard_attrib = params.split('#')
+                if len(storyboard_attrib) != 8:
+                    self._downloader.report_warning('Unable to extract storyboard')
+                    continue
+
+                frame_width = int_or_none(storyboard_attrib[0])
+                frame_height = int_or_none(storyboard_attrib[1])
+                total_frames = int_or_none(storyboard_attrib[2])
+                cols = int_or_none(storyboard_attrib[3])
+                rows = int_or_none(storyboard_attrib[4])
+                filename = storyboard_attrib[6]
+                sigh = storyboard_attrib[7]
+
+                if frame_width and frame_height and cols and rows and total_frames:
+                    frames = cols * rows
+                    width, height = frame_width * cols, frame_height * rows
+                    n_images = int(math.ceil(total_frames / float(cols * rows)))
+                else:
+                    self._downloader.report_warning('Unable to extract storyboard')
+                    continue
+
+                storyboards_url = base_url.replace('$L', compat_str(i)) + '&'
+                for j in range(n_images):
+                    url = storyboards_url.replace('$N', filename).replace('$M', compat_str(j)) + 'sigh=' + sigh
+                    if j == n_images - 1:
+                        remaining_frames = total_frames % (cols * rows)
+                        if remaining_frames != 0:
+                            frames = remaining_frames
+                            rows = int(math.ceil(float(remaining_frames) / rows))
+                            height = rows * frame_height
+                            if rows == 1:
+                                cols = remaining_frames
+                                width = cols * frame_width
+
+                    storyboards.append({
+                        'id': 'L{0}-M{1}'.format(i, j),
+                        'width': width,
+                        'height': height,
+                        'cols': cols,
+                        'rows': rows,
+                        'frames': frames,
+                        'url': url
+                    })
+
         category = microformat.get('category') or search_meta('genre')
         channel_id = self._extract_channel_id(
             webpage, videodetails=video_details, metadata=microformat)
@@ -2563,11 +2630,12 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
             'categories': [category] if category else None,
             'tags': keywords,
             'is_live': is_live,
+            'storyboards': storyboards,
         }
 
         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 +2652,21 @@ 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(',')],

youtube_dl/options.py

@@ -789,6 +789,14 @@ def parseOpts(overrideArguments=None):
         '--list-thumbnails',
         action='store_true', dest='list_thumbnails', default=False,
         help='Simulate and list all available thumbnail formats')
+    thumbnail.add_option(
+        '--write-storyboards',
+        action='store_true', dest='writestoryboards', default=False,
+        help='Write all storyboards (grid of video frames) to disk')
+    thumbnail.add_option(
+        '--list-storyboards',
+        action='store_true', dest='list_storyboards', default=False,
+        help='Simulate and list all available storyboards')
 
     postproc = optparse.OptionGroup(parser, 'Post-processing Options')
     postproc.add_option(