Support two different makers for DoubleRelaxMaker

[Andrew-S-Rosen]

Summary

Addresses #30, wherein the DoubleRelaxMaker flow can now take in two separate BaseVaspMaker args in case the user wants the first or second relax job to have different settings. A common use-case could be to set user_kpoint_settings in the first relax maker to be half the usual value. Ideally, there'd be an optional parameter called half_kpts_first_relax that could do this automatically, but I had some trouble figuring out the best way to modify the k-point density of the first job in a clean way that accounts for KSPACING vs. the other k-point generation schemes.

Edit: In hindsight, I guess the user could always have just defined the flow first and then edited the desired job.maker.input_set_generator afterwards... hmm...

[utf]

I think this is a reasonable approach to take. Am I ok to merge?

[utf]

I actually think relax_maker1 and relax_maker2 could be better than relax_maker and relax_maker2. Just for style and consistency.

[Andrew-S-Rosen]

Got it! I'll make that consistent throughout (eg in the tests too).

[Andrew-S-Rosen]

Also, sorry about the hacky commits here. I'm coding in quick bursts at the airport 😅 I'll get it sorted out and ready to merge soon though.

[utf]

Sorry, you might still be editing but noticed a couple of things.

[Andrew-S-Rosen]

@utf, yup I'm still editing. I pushed the change because I have some questions. The tests fail due to the LREAL assertion. Do you know why this might be? I'm having trouble pinpoint it.

Definitely a (useful) learning process!

Edit: And don't worry, I'll squash all the commits before the PR. I mainly wanted to prematurely push here so you could see where I was struggling.

[Andrew-S-Rosen]

@utf, okay the tests finally pass now. However, could I ask you to take care of the symlinking? I am on Windows and am worried that it's going to wreak havoc on things. The only files that need to be kept as an actual file are the .../inputs/INCAR in the Si_double_relax_swaps subdirectories. The rest can be symlinks.

[codecov-commenter]

Codecov Report

Merging #32 (03dec13) into main (fe5d8a6) will decrease coverage by 0.01%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##             main      #32      +/-   ##
==========================================
- Coverage   70.39%   70.38%   -0.02%     
==========================================
  Files          45       45              
  Lines        3804     3809       +5     
  Branches      576      576              
==========================================
+ Hits         2678     2681       +3     
- Misses        979      981       +2     
  Partials      147      147              

Impacted Files	Coverage Δ
src/atomate2/vasp/flows/amset.py	0.00% <ø> (ø)
src/atomate2/vasp/flows/elastic.py	94.73% <ø> (ø)
src/atomate2/vasp/flows/elph.py	84.09% <ø> (ø)
src/atomate2/vasp/flows/core.py	93.91% <100.00%> (+0.27%)	⬆️
src/atomate2/utils/path.py	88.23% <0.00%> (-11.77%)	⬇️

[utf]

Fantastic, thank you!