[extractor] Add `_search_nextjs_data` (#1386)

Authored by: selfisekai
2021-10-23 04:02:23 +02:00 · 2021-10-23 04:02:23 +02:00 · f98709af31
parent c586f9e8de
commit f98709af31
1 changed files with 7 additions and 0 deletions
--- a/yt_dlp/extractor/common.py
+++ b/yt_dlp/extractor/common.py
@ -1496,6 +1496,13 @@ class InfoExtractor(object):
                    break
        return dict((k, v) for k, v in info.items() if v is not None)

+    def _search_nextjs_data(self, webpage, video_id, **kw):
+        return self._parse_json(
+            self._search_regex(
+                r'(?s)<script[^>]+id=[\'"]__NEXT_DATA__[\'"][^>]*>([^<]+)</script>',
+                webpage, 'next.js data', **kw),
+            video_id, **kw)
+
    @staticmethod
    def _hidden_inputs(html):
        html = re.sub(r'<!--(?:(?!<!--).)*-->', '', html)