Improving autobpm concurrency

[jcjordyn130]

So I'm currently running autobpm after importing my library of ~4000 tracks (93.5GiB) and after looking at the log, my system stats, and the source code, it seems to be processing the tracks one at a time.

I feel that this would be a good place to add concurrency as my data source can read much faster than a track at a time.