Raise ValueError in SpacegroupAnalyzer.get_symmetrized_structure() if spglib returns no symmetries

[janosh]

Closes #2723.

[shyuep]

I think we need to be clear what spglib not returning symmetries mean. If it is a problem assigning any symmetry at all, then it should be an error. If it is a P1 structure, then it should return a single symmetry operation - the identity.

[shyuep]

Also, let's try to keep linting fixes separate from code fixes unless they are in the same file. This PR has 24 files changed and it is not possible to see what is relevant easily.

[janosh]

I think we need to be clear what spglib not returning symmetries mean. If it is a problem assigning any symmetry at all, then it should be an error. If it is a P1 structure, then it should return a single symmetry operation - the identity.

It's the former. I think the structure should have spacegroup 225 (cF4). That's what I get when I decrease symprec=1e-1 to e.g. 1e-2. What type of error? ValueError?

Also, let's try to keep linting fixes separate from code fixes unless they are in the same file. This PR has 24 files changed and it is not possible to see what is relevant easily.

Good point. I'll try to accumulate them over time into separate PRs.

Although you can see the relevant changes by clicking on the commits with non-formatting messages like fb2fc72.

[shyuep]

If it is the former, you can have a ValueError that says spglib cannot determine symmetry and you may need to change the symprec. Unfortunately symmetry determination is always precision specific. Nothing much we can do on that front.

For linting fixes, you can directly push to master branch and there is no need to create PRs for it.