[ie/youtube:tab] Improve shorts title extraction (#11991) (#11997)

Closes #11991
Authored by: d3d9, bashonly

Co-authored-by: bashonly <88596187+bashonly@users.noreply.github.com>

yt_dlp/extractor/youtube.py

@@ -5370,10 +5370,12 @@ class YoutubeTabBaseInfoExtractor(YoutubeBaseInfoExtractor):
             yield self.url_result(
                 f'https://www.youtube.com/shorts/{video_id}',
                 ie=YoutubeIE, video_id=video_id,
-                **traverse_obj(renderer, ('overlayMetadata', {
-                    'title': ('primaryText', 'content', {str}),
-                    'view_count': ('secondaryText', 'content', {parse_count}),
-                })),
+                **traverse_obj(renderer, {
+                    'title': ((
+                        ('overlayMetadata', 'primaryText', 'content', {str}),
+                        ('accessibilityText', {lambda x: re.fullmatch(r'(.+), (?:[\d,.]+(?:[KM]| million)?|No) views? - play Short', x)}, 1)), any),
+                    'view_count': ('overlayMetadata', 'secondaryText', 'content', {parse_count}),
+                }),
                 thumbnails=self._extract_thumbnails(renderer, 'thumbnail', final_key='sources'))
             return