Merge 34556cd67dcce647ca9d08a72afc797eb731f702 into da7223d4aa42ff9fc680b0951d043dd03cec2d30

* improve extraction of global "useful data" Array from player JS
* also handle tv-player and add tests: thx seproDev (yt-dlp/yt-dlp#12684)

Co-Authored-By: sepro <sepro@sepr0.com>

test/test_youtube_signature.py

@@ -232,8 +232,32 @@ _NSIG_TESTS = [
         'W9HJZKktxuYoDTqW', 'jHbbkcaxm54',
     ),
     (
-        'https://www.youtube.com/s/player/91201489/player_ias_tce.vflset/en_US/base.js',
-        'W9HJZKktxuYoDTqW', 'U48vOZHaeYS6vO',
+        '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',
+    ),
+    (
+        'https://www.youtube.com/s/player/d50f54ef/player_ias_tce.vflset/en_US/base.js',
+        'Ha7507LzRmH3Utygtj', 'XFTb2HoeOE5MHg',
+    ),
+    (
+        'https://www.youtube.com/s/player/074a8365/player_ias_tce.vflset/en_US/base.js',
+        'Ha7507LzRmH3Utygtj', 'ufTsrE0IVYrkl8v',
+    ),
+    (
+        'https://www.youtube.com/s/player/643afba4/player_ias.vflset/en_US/base.js',
+        'N5uAlLqm0eg1GyHO', 'dCBQOejdq5s-ww',
+    ),
+    (
+        'https://www.youtube.com/s/player/69f581a5/tv-player-ias.vflset/tv-player-ias.js',
+        '-qIP447rVlTTwaZjY', 'KNcGOksBAvwqQg',
+    ),
+    (
+        'https://www.youtube.com/s/player/643afba4/tv-player-ias.vflset/tv-player-ias.js',
+        'ir9-V6cdbCiyKxhr', '2PL7ZDYAALMfmA',
     ),
 ]
 

youtube_dl/extractor/extractors.py

@@ -568,6 +568,7 @@ from .khanacademy import (
     KhanAcademyIE,
     KhanAcademyUnitIE,
 )
+from .kick import KickIE
 from .kickstarter import KickStarterIE
 from .kinja import KinjaEmbedIE
 from .kinopoisk import KinoPoiskIE

youtube_dl/extractor/kick.py

@@ -0,0 +1,60 @@
+# coding: utf-8
+from __future__ import unicode_literals
+
+from .common import InfoExtractor
+from ..utils import (
+    float_or_none,
+    int_or_none,
+    parse_iso8601,
+    strip_or_none,
+    traverse_obj,
+    url_or_none,
+)
+
+
+class KickIE(InfoExtractor):
+    _VALID_URL = r'https?://(?:www\.)?kick\.com/video/(?P<id>[0-9a-zA-Z-]+)'
+    _TESTS = [{
+        'url': 'https://kick.com/video/82a3c11d-7a17-4747-aecb-2e61413eb11f',
+        'md5': 'f052bc1046cd9ca6751925dd12420010',
+        'info_dict': {
+            'id': '82a3c11d-7a17-4747-aecb-2e61413eb11f',
+            'ext': 'mp4',
+            'title': 'Weekly Stake Stream',
+            'uploader': 'Eddie',
+            'thumbnail': r're:^https?://.*\.jpg.*$',
+            'timestamp': 1676890314,
+            'upload_date': '20230220',
+        }
+    }]
+
+    def _real_extract(self, url):
+        id = self._match_id(url)
+
+        headers = {
+            'Accept': 'application/json',
+        }
+
+        data = self._download_json('https://kick.com/api/v1/video/%s' % id, id, headers=headers)
+
+        formats = self._extract_m3u8_formats(
+            data['source'], id, 'mp4')
+        self._sort_formats(formats)
+        livestream = data['livestream']
+        strip_lambda = lambda x: strip_or_none(x) or None
+
+        return {
+            'id': id,
+            'formats': formats,
+            'title': livestream.get('session_title'),
+            'uploader': traverse_obj(livestream, ('channel', 'user', 'username'), expected_type=strip_lambda),
+            'thumbnail': url_or_none(livestream.get('thumbnail')),
+            'duration': float_or_none(livestream.get('duration'), scale=1000),
+            'timestamp': traverse_obj(data, 'updated_at', 'created_at', expected_type=parse_iso8601),
+            'release_timestamp': parse_iso8601(data.get('created_at')),
+            'view_count': int_or_none(data.get('views')),
+            'is_live': livestream.get('is_live'),
+            'channel': traverse_obj(livestream, ('channel', 'slug'), expected_type=strip_lambda),
+            'categories': traverse_obj(data, ('categories', Ellipsis, 'name'), expected_type=strip_lambda) or None,
+            'tags': traverse_obj(data, ('categories', Ellipsis, 'tags', Ellipsis), expected_type=strip_lambda) or None,
+        }

youtube_dl/extractor/youtube.py

@@ -91,12 +91,12 @@ class YoutubeBaseInfoExtractor(InfoExtractor):
             'INNERTUBE_CONTEXT': {
                 'client': {
                     'clientName': 'IOS',
-                    'clientVersion': '19.45.4',
+                    'clientVersion': '20.10.4',
                     'deviceMake': 'Apple',
                     'deviceModel': 'iPhone16,2',
-                    'userAgent': 'com.google.ios.youtube/19.45.4 (iPhone16,2; U; CPU iOS 18_1_0 like Mac OS X;)',
+                    'userAgent': 'com.google.ios.youtube/20.10.4 (iPhone16,2; U; CPU iOS 18_3_2 like Mac OS X;)',
                     'osName': 'iPhone',
-                    'osVersion': '18.1.0.22B83',
+                    'osVersion': '18.3.2.22D82',
                 },
             },
             'INNERTUBE_CONTEXT_CLIENT_NAME': 5,
@@ -109,7 +109,7 @@ class YoutubeBaseInfoExtractor(InfoExtractor):
             'INNERTUBE_CONTEXT': {
                 'client': {
                     'clientName': 'MWEB',
-                    'clientVersion': '2.20241202.07.00',
+                    'clientVersion': '2.20250311.03.00',
                     # mweb previously did not require PO Token with this UA
                     'userAgent': 'Mozilla/5.0 (iPad; CPU OS 16_7_10 like Mac OS X) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/16.6 Mobile/15E148 Safari/604.1,gzip(gfe)',
                 },
@@ -122,7 +122,7 @@ class YoutubeBaseInfoExtractor(InfoExtractor):
             'INNERTUBE_CONTEXT': {
                 'client': {
                     'clientName': 'TVHTML5',
-                    'clientVersion': '7.20250120.19.00',
+                    'clientVersion': '7.20250312.16.00',
                     'userAgent': 'Mozilla/5.0 (ChromiumStylePlatform) Cobalt/Version',
                 },
             },
@@ -133,7 +133,7 @@ class YoutubeBaseInfoExtractor(InfoExtractor):
             'INNERTUBE_CONTEXT': {
                 'client': {
                     'clientName': 'WEB',
-                    'clientVersion': '2.20241126.01.00',
+                    'clientVersion': '2.20250312.04.00',
                 },
             },
             'INNERTUBE_CONTEXT_CLIENT_NAME': 1,
@@ -692,7 +692,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
         'invidious': '|'.join(_INVIDIOUS_SITES),
     }
     _PLAYER_INFO_RE = (
-        r'/s/player/(?P<id>[a-zA-Z0-9_-]{8,})/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$',
         r'\b(?P<id>vfl[a-zA-Z0-9_-]+)\b.*?\.js$',
     )
@@ -1857,7 +1857,7 @@ 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\("\{"\))(?=\s*[,;])',
+            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)