Enable bad word filter per channel

Currently the bad word filter works in every channel and every direct message. I think this should be configurable. In most of the systems, direct messages aren’t moderated.