Fix for fir_delays issue in Convolve

[jmumford]

Overview

When users correctly specify fir_delays (FIRDelays in a model spec), they provide delays in TR units. However, in compute_regressor within the Convolve class, the resulting FIR regressors were constructed incorrectly, which showed up as unexpected temporal shifts in the regressors.

This happens because compute_regressor assumes that the provided frame_times are defined at the same temporal resolution as the fir_delays specification (typically TR units). In pybids, frame_times is set to resample_frames, which may have a different resolution than the TR. As a result, the FIR shifts were misspecified, and the regressors did not reflect the intended delays.

Changes

1. Create fir_delays_resample_adjusted and use it in compute_regressor to correctly align delays with the resampled frame times.
2. Repair the regressor names returned by compute_regressor (issue caused by Change 1). The names embed the delay value input (fir_delays_resample_adjusted), so the resampled delays are replaced with the original fir_delays to ensure consistency with model/contrast specification.

Things an expert should review closely

• I am assuming that tr is always available via tr = var.run_info[0].tr. I’m fairly confident this is the case, but it may be worth confirming.
• I am assuming that resample_frames contains uniformly spaced frame times. This may not hold for inputs that are not SparseRunVariable objects. This matters because the sampling interval is computed as dt = resample_frames[1] - resample_frames[0], and any non-uniform spacing could affect the FIR delay adjustment.

[codecov]

Codecov Report

❌ Patch coverage is 22.22222% with 7 lines in your changes missing coverage. Please review.
✅ Project coverage is 89.25%. Comparing base (8895e28) to head (f244774).
⚠️ Report is 5 commits behind head on master.

Files with missing lines	Patch %	Lines
src/bids/modeling/transformations/compute.py	22.22%	5 Missing and 2 partials ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #1164      +/-   ##
==========================================
- Coverage   89.33%   89.25%   -0.09%     
==========================================
  Files          66       66              
  Lines        7259     7268       +9     
  Branches     1149     1152       +3     
==========================================
+ Hits         6485     6487       +2     
- Misses        565      570       +5     
- Partials      209      211       +2     

☔ 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.

[jmumford]

Bah! Something is off (other than my missing tests). Michael tested this and found it was including a lag0 by default (even though that wasn't specified), so I'll look into that issue next.