[facebook] extract better titles (fixes #14156)

This now tries to find the primary title, which is displayed in bold
above the previously extracted full text. This is the title which is
also displayed - if given by the uploader - to the top left when
displaying a video fullscreen.

For videos posted to groups, the extractor still doesn't use the
given comment text as a title, and now extracts e.g. "[group name]
Public Group" instead of "[group name] has 481 members", which may
or may not be better.

youtube_dl/extractor/facebook.py

@@ -410,6 +410,10 @@ class FacebookIE(InfoExtractor):
         video_title = self._html_search_regex(
             r'<h2\s+[^>]*class="uiHeaderTitle"[^>]*>([^<]*)</h2>', webpage,
             'title', default=None)
+        if not video_title:
+            video_title = self._html_search_regex(
+                r'(?s)<title id="pageTitle"[^>]*>([^<]*)(?: \| Facebook)</title>',
+                webpage, 'title', default=None)
         if not video_title:
             video_title = self._html_search_regex(
                 r'(?s)<span class="fbPhotosPhotoCaption".*?id="fbPhotoPageCaption"><span class="hasCaption">(.*?)</span>',