Make compatible with Python 3

[j-wags]

I'm working on this in my free time, if for no other reason than to better understand python 2-to-3 migrations.

Basically all real changes have been in pymolecule.py. The rest of the files are just vanilla output from running 2to3.py.

This PR aims to

• Make tests pass on python 3
• black the whole codebase (thanks for the tip, @Seanny123)

Future PRs will

• Migrate tests to pytest
• Replace pymolecule with MDAnalysis, which is better in literally every way
• Get this up on conda-forge

[j-wags]

I attended a talk by Jacob Durrant (original POVME author!) a few months ago, and in chatting afterwards, he pointed out that he'd upgraded POVME2 to Python 3. There may be some good tips on the migration here: https://git.durrantlab.pitt.edu/jdurrant/POVME

[Seanny123]

Looking at Durrant's code, he chose to keep pymolecule.py and update it. So the easiest way forward would probably be copying over his code.

That being said, I'm currently trying to learn how to leverage MDAnalysis in various projects. If you could point me towards the simplest test-case you can find, I'd like to give a shot at replacing pymolecule with MDAnalysis.

[j-wags]

Sorry to drop this conversation -- Totally lost track of this!

A drop-in replacement of pymolecule from Durrant's repo might work here. I can't recall if I'd made modifications to it that modified the behavior or API in reverse-incompatible ways, but it'd be worth a try.

The MDAnalysis documentation is really top-notch, and the simplest POVME test is probably defined_region_inclusion.

IIRC, when I'd looked at incorporating MDAnalysis earlier, there were some deep differences between the MDAnalysis Universe and and pymolecule Molecule object models that would make it tricky. Unfortunately I think that the complexity of the behaviors that need to be replicated might make this a somewhat challenging task.