This is pretty neat. One suggestion though: for unknown file types, try to parse them as text files. Many might succeed.

Also, is it open source?