fix logic error in NEB vacancy supercell generation

[jacksund]

This pull request address the bug in #245 where the incorrect supercell is generated with vac_mode=True

[shyuep]

Thanks. Can you add a simple unittest pls?

[coveralls]

Pull Request Test Coverage Report for Build 1139703774

• 7 of 7 (100.0%) changed or added relevant lines in 1 file are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage increased (+0.09%) to 79.941%

---

Totals
Change from base Build 1076204842:	0.09%
Covered Lines:	1614
Relevant Lines:	2019

---

💛 - Coveralls

[jacksund]

Sure thing. Looks like I need to run black formatting too.

Before you accept my changes, I wanted to point out that this changes how some supercell structures are generated when vac_mode=False. In reality, I think there are actually three modes that should be considered when making supercells:

1. vac_mode=True: where all diffusing species are kept in the supercell structure except for 1 (so a single vacancy)
2. vac_mode=False (v1): where all diffusing species are removed in the supercell structure except for 1 (so a single diffusion ion plus the host lattice). This is the current implementation of your package.
3. vac_mode=False (v2): where all symmetrically equivalent sites of the diffusing atom are removed (so sites of the same species that are not part of the diffusing pathway are still kept). This is the logic that this pull request currently implements.

The difference between 2 and 3 won't be seen in the majority of structures, but it may be important for lower-symmetry ones.

[shyuep]

There are still some linting errors.

[jacksund]

@shyuep for future reference, what commands do you use to run all linting locally? I'm only familiar with black and don't typically use pylint. Is it just something like pylint --rcfile=pylintrc in the base directory?

[shyuep]

The commands are listed in this Github workflow: https://github.com/materialsvirtuallab/pymatgen-analysis-diffusion/blob/master/.github/workflows/linting.yml

[jacksund]

Awesome, thank you!