[nbc] fix NBCNews/Today/MSNBC extraction

whosa · Nov 16, 2020 · 2ea9c97 · 2ea9c97
1 parent efc589b
commit 2ea9c97
Showing 1 changed file with 2 additions and 3 deletions.
diff --git a/youtube_dl/extractor/nbc.py b/youtube_dl/extractor/nbc.py
@@ -10,7 +10,6 @@
 from ..compat import compat_urllib_parse_unquote
 from ..utils import (
     int_or_none,
-    js_to_json,
     parse_duration,
     smuggle_url,
     try_get,
@@ -394,8 +393,8 @@ def _real_extract(self, url):
         webpage = self._download_webpage(url, video_id)
 
         data = self._parse_json(self._search_regex(
-            r'window\.__data\s*=\s*({.+});', webpage,
-            'bootstrap json'), video_id, js_to_json)
+            r'<script[^>]+id="__NEXT_DATA__"[^>]*>({.+?})</script>',
+            webpage, 'bootstrap json'), video_id)['props']['initialState']
         video_data = try_get(data, lambda x: x['video']['current'], dict)
         if not video_data:
             video_data = data['article']['content'][0]['primaryMedia']['video']