Add extra token cleanup functionality

Currently the only token cleanup functionality is to convert to lower case, this is not enough.

Spaces, certain characters, certain words, etc should all be filtered similar to the way Pathauto already offers. If possible it would be good to actually hook directly into pathauto if it is present, but I will have to investigate if this is even possible. Certain filtering should still be offered if pathauto is not present if this is the path to take.