Fix CuPy batched solve shape mismatch in _estimate_transfer_function

[tomHep10]

Fixes CuPy GPU crash in _estimate_transfer_function where a non-batched identity matrix (M, M) was passed to xp.linalg.solve with a batched a of shape (..., 1, M, M).

Fix: broadcast identity to match H_0.shape before solving.

Confirmed working on NVIDIA B200 GPU with LFP dataset.

[emilyps14]

I checked this and it looks reasonable to me. I'm fairly certain that all of the annoying assert changes are just formatting. Want @edeno to doublecheck the CuPy main edits.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 88.35%. Comparing base (a5d2983) to head (02eca6d).

Additional details and impacted files

@@            Coverage Diff             @@
##           master      #76      +/-   ##
==========================================
- Coverage   88.65%   88.35%   -0.30%     
==========================================
  Files           8        8              
  Lines        1789     1786       -3     
==========================================
- Hits         1586     1578       -8     
- Misses        203      208       +5     

☔ 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.

[edeno]

Hi @tomHep10,

Thank you for catching this and for your contribution. I hope you don't mind, but I reverted some changes the ruff formatter would have complained about, added a test, and fixed another spot this would be an issue.

[tomHep10]

@edeno Not at all! Thank you, glad this helped.