remove URL regex caret anchor

[missinglink]

This PR adjusts the URL detection regex by removing the leading ^ caret anchor.
This allows us to detect URLs within strings containing other characters.

Found via: #279 (comment)

Originally added in: d4c5305

[missinglink]

I've added a second commit which reworks the URL detection, so instead of discarding names containing URLs, we now strip the URL component and then try to recover the rest of the name.

[orangejulius]

Yeah good idea. Might as well use what information is there

[orangejulius]

We should probably port this to the OSM importer too, we get so many questions about these errors.

[Joxit]

Why don't you use this logic in model repository? Instead of throw an error try to trim the http / ftp and if the name is empty throw the error with the original name? 🤔

[missinglink]

Why don't you use this logic in model repository?

Hmm yeah, good idea, I was originally thinking that model would only validate things and the individual importers would be better suited to cleaning their data, but you're right, it would make it much easier to maintain and avoid all these GH issues with people complaining about the warnings.