From b5fea53f2099bed41ba1b17ab0ac87c8dba5a5ec Mon Sep 17 00:00:00 2001
From: bashonly <88596187+bashonly@users.noreply.github.com>
Date: Sat, 12 Jul 2025 18:12:05 -0500
Subject: [PATCH] [ie] Rework `_search_nextjs_v13_data` helper (#13711)

Fix 5245231e4a39ecd5595d4337d46d85e150e2430a

Authored by: bashonly
---
 test/test_InfoExtractor.py | 31 ++++++++++++++++++-------------
 yt_dlp/extractor/common.py | 18 ++++++++++++------
 2 files changed, 30 insertions(+), 19 deletions(-)

diff --git a/test/test_InfoExtractor.py b/test/test_InfoExtractor.py
index 7c3825f779..40dd05e136 100644
--- a/test/test_InfoExtractor.py
+++ b/test/test_InfoExtractor.py
@@ -1969,21 +1969,26 @@ def test_search_nextjs_v13_data(self):
             <script>self.__next_f.push([1,"df:[\"$undefined\",[\"$\",\"div\",null,{\"children\":[\"$\",\"$L17\",null,{}],\"do_not_include_this_field\":\"fail\"}],[\"$\",\"div\",null,{\"children\":[[\"$\",\"$L19\",null,{\"duplicated_field_name\":{\"x\":1}}],[\"$\",\"$L20\",null,{\"duplicated_field_name\":{\"y\":2}}]]}],\"$undefined\"]\n"])</script>
             <script>self.__next_f.push([3,"MzM6WyIkIiwiJEwzMiIsbnVsbCx7ImRlY29kZWQiOiJzdWNjZXNzIn1d"])</script>
             '''
-        EXPECTED = [{
-            'foo': 'bar',
-        }, {
-            'meta': {
-                'dateCreated': 1730489700,
-                'uuid': '40cac41d-8d29-4ef5-aa11-75047b9f0907',
+        EXPECTED = {
+            '18': {
+                'foo': 'bar',
             },
-        }, {
-            'duplicated_field_name': {'x': 1},
-        }, {
-            'duplicated_field_name': {'y': 2},
-        }]
+            '16': {
+                'meta': {
+                    'dateCreated': 1730489700,
+                    'uuid': '40cac41d-8d29-4ef5-aa11-75047b9f0907',
+                },
+            },
+            '19': {
+                'duplicated_field_name': {'x': 1},
+            },
+            '20': {
+                'duplicated_field_name': {'y': 2},
+            },
+        }
         self.assertEqual(self.ie._search_nextjs_v13_data(HTML, None), EXPECTED)
-        self.assertEqual(self.ie._search_nextjs_v13_data('', None, fatal=False), [])
-        self.assertEqual(self.ie._search_nextjs_v13_data(None, None, fatal=False), [])
+        self.assertEqual(self.ie._search_nextjs_v13_data('', None, fatal=False), {})
+        self.assertEqual(self.ie._search_nextjs_v13_data(None, None, fatal=False), {})
 
     def test_search_nuxt_json(self):
         HTML_TMPL = '<script data-ssr="true" id="__NUXT_DATA__" type="application/json">[{}]</script>'
diff --git a/yt_dlp/extractor/common.py b/yt_dlp/extractor/common.py
index a3ff5a1c0b..d601e17514 100644
--- a/yt_dlp/extractor/common.py
+++ b/yt_dlp/extractor/common.py
@@ -1785,7 +1785,7 @@ def _search_nextjs_data(self, webpage, video_id, *, fatal=True, default=NO_DEFAU
 
     def _search_nextjs_v13_data(self, webpage, video_id, fatal=True):
         """Parses Next.js app router flight data that was introduced in Next.js v13"""
-        nextjs_data = []
+        nextjs_data = {}
         if not fatal and not isinstance(webpage, str):
             return nextjs_data
 
@@ -1797,9 +1797,9 @@ def flatten(flight_data):
                 if not isinstance(data, dict):
                     return
                 children = data.pop('children', None)
-                if data and isinstance(name, str) and name.startswith('$'):
+                if data and isinstance(name, str) and re.fullmatch(r'\$L[0-9a-f]+', name):
                     # It is useful hydration JSON data
-                    nextjs_data.append(data)
+                    nextjs_data[name[2:]] = data
                 flatten(children)
                 return
             for f in flight_data:
@@ -1823,10 +1823,16 @@ def flatten(flight_data):
                 flight_text += chunk
 
         for f in flight_text.splitlines():
-            prefix, _, body = f.partition(':')
-            if body.startswith('[') and body.endswith(']') and re.fullmatch(r'[0-9a-f]{1,3}', prefix.lstrip()):
-                # The body isn't necessarily valid JSON, so this should always be non-fatal
+            prefix, _, body = f.lstrip().partition(':')
+            if not re.fullmatch(r'[0-9a-f]+', prefix):
+                continue
+            # The body still isn't guaranteed to be valid JSON, so parsing should always be non-fatal
+            if body.startswith('[') and body.endswith(']'):
                 flatten(self._parse_json(body, video_id, fatal=False, errnote=False))
+            elif body.startswith('{') and body.endswith('}'):
+                data = self._parse_json(body, video_id, fatal=False, errnote=False)
+                if data is not None:
+                    nextjs_data[prefix] = data
 
         return nextjs_data