[FIX] ASR: asr_calibrate must pass step_sizes to fit_eeg_distribution

[sappelhoff]

fit_eeg_distribution's step_sizes default was [0.0220, 0.6000], an accidental copy of the fit_quantiles default. That value collapses the scale grid search to essentially a single point, so asr_calibrate (which called fit_eeg_distribution without step_sizes) inherited a degenerate grid, biasing the fitted mu/sig and hence the threshold matrix T. The function's own docstring already documents default=[0.01, 0.01], matching the value clean_windows passes explicitly.

Correct the signature default to match the docstring, and have asr_calibrate pass step_sizes=[0.01, 0.01] explicitly so its behavior no longer depends on the default.

Testing

Added a regression test that builds a windowed-RMS vector from the eeg_raw.npy fixture and asserts the default fit now matches an explicit [0.01, 0.01] call while diverging materially from the old [0.022, 0.6] default. Full tests/test_asr.py passes; ruff clean.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 83.11%. Comparing base (107cd1b) to head (d987300).

Additional details and impacted files

@@           Coverage Diff           @@
##           master     #109   +/-   ##
=======================================
  Coverage   83.11%   83.11%           
=======================================
  Files          25       25           
  Lines        2825     2825           
=======================================
  Hits         2348     2348           
  Misses        477      477           

☔ View full report in Codecov by Harness.
📢 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.

[sappelhoff]

Note: includes the nonlinear_eigenspace reproducibility fix from #118. test_asr_class[*-riemann] compares incremental-vs-bulk output with a loose tolerance and is flaky on CI because that path used an unseeded RNG; this can trip on any PR regardless of what it changes. #118 is the root fix — once it merges first, this commit drops out on rebase.