Only copy URLs from known classes

[roachcord3]

My use case here is fanbox posts, where each archive file will have three URLs, two of which are from the post (and are part of known URL classes if you use the community fanbox parser here,) and one of which is just a CDN link for the archive file. It's pretty useless to associate this CDN link to a bunch of files, and it's kind of a hassle to clean up later, too. It would be nice if hyextract only copied URLs if they're from known URL classes.

[floogulinc]

You could just have hydrus not store those URLs in the original archive files but this is possible and something I'll consider.