Refactor scoring pipeline

[rvhonorato]

This PR is a small refactoring to the scoring pipeline that implements #37 and adds a Setup python module, in which we should centralize the validations related to the .toml file.

I also re-introduced the haddock-style begin/ and data/ directories.

[joaomcteixeira]

How do you keep the link between the new begin_mol and the original mol_identifier path that is now in data_dir?

[rvhonorato]

You don't, the data is there just for safekeeping the pre-processed inputs, it's how its done in production (2.4)

[joaomcteixeira]

Adding my suggestions at: https://github.com/rvhonorato/haddock3/pull/1

[joaomcteixeira]

Okay. I think I finished reviewing your PR; my reviews and suggestions at: https://github.com/rvhonorato/haddock3/pull/1, if you accept them they will become part of this PR automatically.

It is a great idea @rvhonorato to firstly check the configuration file and prepare the files and folders for run. I like that idea. My suggestions in your branch are basically pythonic and architecture refactoring. I dislike the OOP-everywhere paradigm. It might be worthwhile in other languages but in python not really. It is useful where it is useful, but creates a lot of headaches otherwise. I haven't implemented any tests, but you'll see tests are easy to write with the changes I propose.

You'll notice the new folder gear/. We can call it differently if you wish. But to my experience, I am organizing my code in three main folders: clis/, libs/ and gear/. libs have all the general and abstract implementations, those that can serve anywhere in the project. While gear/ is dedicated to specific logic to serve specific implementations. To my experience, this gives very nice results. I rename haddock.setup.py because setup.py has a very specific meaning in the python world, and it may confuse other developers or us in the future.

Let me know,
cheers!

[rvhonorato]

Thanks for the review and the edits @joaomcteixeira !