Refactoring and improvements suggested

[levitsky]

Hi guys, I'm tasked with using IQMMA in my software and I'm looking at the code to see how to call it. It looks like it would need at least some changes to be able to be called externally from a Python function without CLI.

Necessary changes

• split the run function in two: one for Argparser stuff and one for actual processing, similar to other tools like Scavager (Split the run function to allow calling from Python #15)
• use logging in a way that allows external configuration (Logging changes #18)

Then I see some other changes that I can suggest, although they are optional, at least for me:

• simplify CLI arguments handling: right now there are too many entities: console_config, console_keys, args, users_config.
• simplify config file parsing - use configparser instead of custom code
• clean up unnecessary imports
• version requirements: is there any reason NumPy is pinned at 1.23.1? (Edit: this requirement is a workaround for a diffacto error which is fixed in Fix dtype spec for compatibility with latest numpy versions statisticalbiotechnology/diffacto#28)
• fix typos in the documentation
• resolve errors where some used variables are not defined: presumably, some of the optional features don't work but I have not tested anything yet. You can use a linter to see what places may have problems.

I am happy to help with all of the above but I'd like to coordinate this with you guys first, also I don't have any means of testing the integrity of the package to check if I break anything once I dig in. Can you suggest a benchmark to maximize the coverage of software features tested?