Add support for per-atomic-config calculator in generic.calculator.

[bernstei]

Enables autoparallelization over calculator parameters by optionally constructing per-atom calculator in autoparallelized op wfl.calculators.generic.calculate(), using Atoms.info WFL_CALCULATOR_INITIALIZER, _ARGS, and _KWARGS.

Closes #253

[bernstei]

@Felixrccs if you are interested in using this, could I convince you to put together a pytest (could, but doesn't have to be based on a DFT calculator)?

[Felixrccs]

Yes, I can try to make a pytest. (I have no experience with that yet.)

[bernstei]

Yes, I can try to make a pytest. (I have no experience with that yet.)

If you want a hand, just ask. Relevant example tests are in tests/calculators/

[Felixrccs]

If you want a hand, just ask.

My pytest is in #261.

[bernstei]

@Felixrccs there's an issue with the vasp test. Do you need me to debug it, or have you figured out how to run the vasp tests locally on your machine to fix it?

[bernstei]

@Felixrccs any progress cleaning up the vasp test?

[Felixrccs]

Sry I had some busy weeks, I try to do it tomorrow.

[bernstei]

Thanks - just wanted to make sure that it wasn't forgotten

[Felixrccs]

Ok I still have calculations running and don't want to change my code meanwhile, I hope I get it done on the weekend.

[Felixrccs]

I fixed it in #265. This should work now, thank for your patience.

[bernstei]

@Felixrccs thanks for all the work. I'll merge this one once the CI tests pass. Then, hopefully, we can just update and merge the universal k-point spacing functionality.