Merge 9b828daa1a5bea7fe1dde12cb5963d223b800e94 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/ccma.py

@@ -1,16 +1,16 @@
 # coding: utf-8
 from __future__ import unicode_literals
 
-import calendar
-import datetime
 import re
 
 from .common import InfoExtractor
+
+from ..compat import compat_str
 from ..utils import (
     clean_html,
-    extract_timezone,
     int_or_none,
     parse_duration,
+    parse_iso8601,
     parse_resolution,
     try_get,
     url_or_none,
@@ -25,7 +25,7 @@ class CCMAIE(InfoExtractor):
         'info_dict': {
             'id': '5630208',
             'ext': 'mp4',
-            'title': 'L\'espot de La Marató de TV3',
+            'title': "L'espot de La Marató 2016: Ictus i les lesions medul·lars i cerebrals traumàtiques",
             'description': 'md5:f12987f320e2f6e988e9908e4fe97765',
             'timestamp': 1478608140,
             'upload_date': '20161108',
@@ -39,8 +39,8 @@ class CCMAIE(InfoExtractor):
             'ext': 'mp3',
             'title': 'El Consell de Savis analitza el derbi',
             'description': 'md5:e2a3648145f3241cb9c6b4b624033e53',
-            'upload_date': '20170512',
-            'timestamp': 1494622500,
+            'upload_date': '20161217',
+            'timestamp': 1482011700,
             'vcodec': 'none',
             'categories': ['Esports'],
         }
@@ -50,14 +50,29 @@ class CCMAIE(InfoExtractor):
         'info_dict': {
             'id': '6031387',
             'ext': 'mp4',
-            'title': 'Crims - Josep Talleda, l\'"Espereu-me" (capítol 1)',
+            'title': 'Josep Talleda, l\'"Espereu-me" (part 1)',
             'description': 'md5:7cbdafb640da9d0d2c0f62bad1e74e60',
-            'timestamp': 1582577700,
+            'timestamp': 1582577919,
             'upload_date': '20200224',
-            'subtitles': 'mincount:4',
+            'subtitles': 'mincount:1',
             'age_limit': 16,
             'series': 'Crims',
         }
+    }, {
+        'url': 'https://www.ccma.cat/tv3/sx3/la-tria-final/video/6178889/',
+        'md5': '835aecbda55ba7b70d147081cf4b61f2',
+        'info_dict': {
+            'id': '6178889',
+            'ext': 'mp4',
+            'title': 'La tria final (T1 - Capítol 4)',
+            'description': 'md5:7a66ef9802af0e44ce54304bb9b61eba',
+            'timestamp': 1665687816,
+            'episode_number': 4,
+            'upload_date': '20221013',
+            'categories': ['Ficció'],
+            'age_limit': 12,
+            'series': 'Guardians de la nit: Kimetsu no Yaiba',
+        }
     }]
 
     def _real_extract(self, url):
@@ -95,15 +110,7 @@ class CCMAIE(InfoExtractor):
         durada = informacio.get('durada') or {}
         duration = int_or_none(durada.get('milisegons'), 1000) or parse_duration(durada.get('text'))
         tematica = try_get(informacio, lambda x: x['tematica']['text'])
-
-        timestamp = None
-        data_utc = try_get(informacio, lambda x: x['data_emissio']['utc'])
-        try:
-            timezone, data_utc = extract_timezone(data_utc)
-            timestamp = calendar.timegm((datetime.datetime.strptime(
-                data_utc, '%Y-%d-%mT%H:%M:%S') - timezone).timetuple())
-        except TypeError:
-            pass
+        timestamp = parse_iso8601(try_get(informacio, lambda x: x['data_emissio']['utc']), compat_str)
 
         subtitles = {}
         subtitols = media.get('subtitols') or []

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)