This is what I would do:
Analyze the html code of that page. If it were very simple, I would use wget to get the page, and then a combination of sed, awk, or grep inside a bash script.
If it were a complex page, I would use python with a library called BeautifulSoup that is designed to parse html code.
Once I got the *ubuntu*torrents links, I would use transmission-daemon as a back end, and load the torrents using transmission-remote:
Code:
$ transmission-remote -n user:password -a url
Just some thoughts,
Regards.
Bookmarks