[telegram] Fix metadata extraction

Closes #3528

yt_dlp/extractor/telegram.py

@@ -1,4 +1,5 @@
 from .common import InfoExtractor
+from ..utils import clean_html, get_element_by_class
 
 
 class TelegramEmbedIE(InfoExtractor):
@@ -17,8 +18,8 @@ class TelegramEmbedIE(InfoExtractor):
 
     def _real_extract(self, url):
         video_id = self._match_id(url)
-        webpage = self._download_webpage(url, video_id)
-        webpage_embed = self._download_webpage(f'{url}?embed=1', video_id)
+        webpage = self._download_webpage(url, video_id, query={'embed': 0})
+        webpage_embed = self._download_webpage(url, video_id, query={'embed': 1}, note='Downloading ermbed page')
 
         formats = [{
             'url': self._proto_relative_url(self._search_regex(
@@ -29,9 +30,12 @@ class TelegramEmbedIE(InfoExtractor):
 
         return {
             'id': video_id,
-            'title': self._html_search_meta(['og:title', 'twitter:title'], webpage, fatal=True),
-            'description': self._html_search_meta(['og:description', 'twitter:description'], webpage, fatal=True),
-            'thumbnail': self._search_regex(r'tgme_widget_message_video_thumb"[^>]+background-image:url\(\'([^\']+)\'\)',
+            'title': self._html_search_meta(['og:title', 'twitter:title'], webpage, default=None),
+            'description': self._html_search_meta(
+                ['og:description', 'twitter:description'], webpage,
+                default=clean_html(get_element_by_class('tgme_widget_message_text', webpage_embed))),
+            'thumbnail': self._search_regex(
+                r'tgme_widget_message_video_thumb"[^>]+background-image:url\(\'([^\']+)\'\)',
                 webpage_embed, 'thumbnail'),
             'formats': formats,
         }