Fix neighborlist for skewed cells

[pfebrer]

This PR fixes a bug that @ialcon brought up for non-orthogonal cells with the neighbor finder.

Essentially, the algorithm assumes that if you are exactly at the center of a bin you can only access atoms within the same bin. This is clearly not the case in skewed cells when not being careful. Essentially, something like this was happening:

@zerothi could you finish the PR (e.g. pre-commit stuff)? I'm abroad on a limited data plan and I don't want to use it to download pre-commit and friends 😄

[pfebrer]

Ah, I'm just thinking that probably it is not needed to change bin_size to start with 1. We can substitute the 2 here for the bin_size argument for axis i.

Although personally I think it is unnecessary to explain to the users why it starts with 2, and starting with 1 seems simpler to explain/understand. This is something that you added, so up to you to decide :)

[zerothi]

I agree, simpler. Needed another fix in one of the tests.

[codecov]

Codecov Report

❌ Patch coverage is 51.72414% with 14 lines in your changes missing coverage. Please review.
✅ Project coverage is 86.66%. Comparing base (947b52b) to head (cfd3a14).
⚠️ Report is 4 commits behind head on main.

Files with missing lines	Patch %	Lines
src/sisl/geom/_neighbors/tests/test_finder.py	0.00%	14 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #976      +/-   ##
==========================================
- Coverage   86.67%   86.66%   -0.02%     
==========================================
  Files         412      412              
  Lines       54375    54368       -7     
==========================================
- Hits        47131    47117      -14     
- Misses       7244     7251       +7     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.