Multiprocessing pool is breaking long shot scripts

[Jean-Reinhold]

On a Fedora system, after about two hours of operation, the pool used to extract features starts hanging open processes, resulting in a significant increase in memory usage.

The solution was to force the pool to close with pool.close(); pool.join():

I know it is super strange because the context manager should already do this...

[david-waterworth]

My understanding is the context manager calls terminate() so close() and join() are still required. See https://stackoverflow.com/a/55035444/2981639

[mjanson]

we also ran into an issue with multiprocessing from within a flask web service.

Would the project be open to a PR that allows the threading/multiprocessing to be disabled? Our current workaround is to directly invoke _get_feats which is undesirable for several reasons.