Merge a2fd63ce22280cef3856d39d795ec91303152ff6 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_InfoExtractor.py

@@ -187,10 +187,10 @@ class TestInfoExtractor(unittest.TestCase):
         self.assertEqual(search['track']['id'], 'testid')
 
     def test_search_json_ld_realworld(self):
-        # https://github.com/ytdl-org/youtube-dl/issues/23306
-        expect_dict(
-            self,
-            self.ie._search_json_ld(r'''<script type="application/ld+json">
+        _TESTS = [
+            # https://github.com/ytdl-org/youtube-dl/issues/23306
+            (
+                r'''<script type="application/ld+json">
 {
 "@context": "http://schema.org/",
 "@type": "VideoObject",
@@ -223,17 +223,86 @@ class TestInfoExtractor(unittest.TestCase):
 "name": "Kleio Valentien",
 "url": "https://www.eporner.com/pornstar/kleio-valentien/"
 }]}
-</script>''', None),
-            {
-                'title': '1 On 1 With Kleio',
-                'description': 'Kleio Valentien',
-                'url': 'https://gvideo.eporner.com/xN49A1cT3eB/xN49A1cT3eB.mp4',
-                'timestamp': 1449347075,
-                'duration': 743.0,
-                'view_count': 1120958,
-                'width': 1920,
-                'height': 1080,
-            })
+                </script>''',
+                {
+                    'title': '1 On 1 With Kleio',
+                    'description': 'Kleio Valentien',
+                    'url': 'https://gvideo.eporner.com/xN49A1cT3eB/xN49A1cT3eB.mp4',
+                    'timestamp': 1449347075,
+                    'duration': 743.0,
+                    'view_count': 1120958,
+                    'width': 1920,
+                    'height': 1080,
+                },
+                {},
+            ),
+            (
+                r'''<script type="application/ld+json">
+      {
+      "@context": "https://schema.org",
+      "@graph": [
+      {
+      "@type": "NewsArticle",
+      "mainEntityOfPage": {
+      "@type": "WebPage",
+      "@id": "https://www.ant1news.gr/Society/article/620286/symmoria-anilikon-dikigoros-thymaton-ithelan-na-toys-apoteleiosoyn"
+      },
+      "headline": "Συμμορία ανηλίκων – δικηγόρος θυμάτων: ήθελαν να τους αποτελειώσουν",
+      "name": "Συμμορία ανηλίκων – δικηγόρος θυμάτων: ήθελαν να τους αποτελειώσουν",
+      "description": "Τα παιδιά δέχθηκαν την επίθεση επειδή αρνήθηκαν να γίνουν μέλη της συμμορίας, ανέφερε ο Γ. Ζαχαρόπουλος.",
+      "image": {
+      "@type": "ImageObject",
+      "url": "https://ant1media.azureedge.net/imgHandler/1100/a635c968-be71-447c-bf9c-80d843ece21e.jpg",
+      "width": 1100,
+      "height": 756            },
+      "datePublished": "2021-11-10T08:50:00+03:00",
+      "dateModified": "2021-11-10T08:52:53+03:00",
+      "author": {
+      "@type": "Person",
+      "@id": "https://www.ant1news.gr/",
+      "name": "Ant1news",
+      "image": "https://www.ant1news.gr/images/logo-e5d7e4b3e714c88e8d2eca96130142f6.png",
+      "url": "https://www.ant1news.gr/"
+      },
+      "publisher": {
+      "@type": "Organization",
+      "@id": "https://www.ant1news.gr#publisher",
+      "name": "Ant1news",
+      "url": "https://www.ant1news.gr",
+      "logo": {
+      "@type": "ImageObject",
+      "url": "https://www.ant1news.gr/images/logo-e5d7e4b3e714c88e8d2eca96130142f6.png",
+      "width": 400,
+      "height": 400                },
+      "sameAs": [
+      "https://www.facebook.com/Ant1news.gr",
+      "https://twitter.com/antennanews",
+      "https://www.youtube.com/channel/UC0smvAbfczoN75dP0Hw4Pzw",
+      "https://www.instagram.com/ant1news/"
+      ]
+      },
+
+      "keywords": "μαχαίρωμα,συμμορία ανηλίκων,ΕΙΔΗΣΕΙΣ,ΕΙΔΗΣΕΙΣ ΣΗΜΕΡΑ,ΝΕΑ,Κοινωνία - Ant1news",
+
+
+      "articleSection": "Κοινωνία"
+      }
+      ]
+      }
+                </script>''',
+                {
+                    'timestamp': 1636523400,
+                    'title': 'md5:91fe569e952e4d146485740ae927662b',
+                },
+                {'expected_type': 'NewsArticle'},
+            ),
+        ]
+        for html, expected_dict, search_json_ld_kwargs in _TESTS:
+            expect_dict(
+                self,
+                self.ie._search_json_ld(html, None, **search_json_ld_kwargs),
+                expected_dict
+            )
 
     def test_download_json(self):
         uri = encode_data_uri(b'{"foo": "blah"}', 'application/json')

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/common.py

@@ -1438,8 +1438,16 @@ class InfoExtractor(object):
             })
             extract_interaction_statistic(e)
 
-        for e in json_ld:
-            if '@context' in e:
+        def traverse_json_ld(json_ld, info, at_top_level=True):
+            for e in json_ld:
+                if at_top_level and '@context' not in e:
+                    continue
+                if at_top_level and all(k in ('@context', '@graph') for k in e):
+                    graph = e['@graph']
+                    if isinstance(graph, dict):
+                        graph = [graph]
+                    traverse_json_ld(graph, info, at_top_level=False)
+                    break
                 item_type = e.get('@type')
                 if expected_type is not None and expected_type != item_type:
                     continue
@@ -1472,7 +1480,7 @@ class InfoExtractor(object):
                     info.update({
                         'timestamp': parse_iso8601(e.get('datePublished')),
                         'title': unescapeHTML(e.get('headline')),
-                        'description': unescapeHTML(e.get('articleBody')),
+                        'description': unescapeHTML(e.get('articleBody') or e.get('description')),
                     })
                 elif item_type == 'VideoObject':
                     extract_video_object(e)
@@ -1487,6 +1495,8 @@ class InfoExtractor(object):
                     continue
                 else:
                     break
+        traverse_json_ld(json_ld, info)
+
         return dict((k, v) for k, v in info.items() if v is not None)
 
     def _search_nextjs_data(self, webpage, video_id, **kw):

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)