Detrend in place in NoisyChannels.__init__ to avoid full-array copies

[sappelhoff]

Summary

Detrend the EEG signal in place during NoisyChannels.__init__, avoiding two transient full-size copies of the recording during construction.

• removeTrend gains a copy=True keyword argument that is passed through to mne.filter.filter_data. The default (copy=True) preserves the exact current behavior for every existing caller.
• NoisyChannels.__init__ now calls removeTrend(self.raw_mne._data, ..., copy=False) instead of removeTrend(self.raw_mne.get_data(), ...).

Why

The previous code allocated two avoidable full-array copies while detrending:

1. get_data() always returns a copy (verified: it does not share memory with _data).
2. mne.filter.filter_data(copy=True) allocates a separate output buffer.

Filtering in place removes both. On an 18.5-min, 99-channel @ 500 Hz recording (~439 MB), filter_data peaks at +887 MB → +443 MB (one full-array transient removed), and __init__ drops from ~1.94 s to ~1.77 s.

The matlab_strict path already filters in place, so copy only affects the default (MNE high-pass) branch.

Correctness

• mne.filter.filter_data(copy=False) is bit-identical to copy=True (verified np.array_equal); copy only controls whether MNE allocates a new array.
• Full test suite passes (47 passed), including tests/test_matprep_compare.py.
• get_bads(as_dict=True) and all _extra_info arrays unchanged on the real recording (reject "omit"/None) and the eegbci fixture (matlab_strict False/True).

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 97.85%. Comparing base (24530cb) to head (ca8708d).
⚠️ Report is 1 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #196   +/-   ##
=======================================
  Coverage   97.85%   97.85%           
=======================================
  Files           7        7           
  Lines         839      839           
=======================================
  Hits          821      821           
  Misses         18       18           

☔ 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.