yt_data_extract: normalize thumbnail and author urls

for instance, urls that start with // become https://

adjustment required in comments.py because the url was left as a
relative url in yt_data_extract by mistake and was using URL_ORIGIN
prefix as fix.

see #31

youtube/comments.py

@@ -90,7 +90,7 @@ def single_comment_ctoken(video_id, comment_id):
 def post_process_comments_info(comments_info):
     for comment in comments_info['comments']:
         comment['author_url'] = concat_or_none(
-            util.URL_ORIGIN, comment['author_url'])
+            '/', comment['author_url'])
         comment['author_avatar'] = concat_or_none(
             settings.img_prefix, comment['author_avatar'])