Adding BGS filter to prepare_sim

[SBouchard01]

At Sandy's suggestion, I had a custom filter on the halos & particles subsamples to allow the BGS density to be reached in my branch of the package (it exists in the current file, but commented out). Given the recent changes in prepare_sim that need to re-run it, I propose an implementation of that filter as an option in the official pipeline.

This branch also replaces most of the print statements by logger calls for clarity and stdout buffering delays in prepare_sim and menv.

Open questions & remarks:

• 2 alternate functions of subsample_halos and submask_particles are commented out in the code. They are functionally the same as the current ones but with some magic values changed.
• By curiosity, submask_particles has an extra line when calling the MT (or BGS) filter: ntarget = np.minimum(ntarget, 100) that is not called for the ST usual filter. Is that intentional ? If yes, why ? If not, the statement can be moved before the submask initialization to avoid duplicate code.

[lgarrison]

Thanks @SBouchard01! This looks good to me, just some minor nits below.

Also, any reason we need a __init__.py file? Previously we had been using namespace packages. This is probably a harmless addition, but just confirming it was intentional.

[lgarrison]

I think this should use __name__ instead of __file__?

[lgarrison]

Same here

[lgarrison]

I think @SandyYuan can answer the questions about the different magic values and the ntarget maximum?