Does CuBIDS account for inheritance with variants' associated files?

[tsalo]

Summary

I was thinking of cases where the datasets leverage inheritance with associated files (e.g., events, bvals, bvecs) and some of the source data files end up being variants (especially ones that are shortened).

As an explicit example, what if we have a film-viewing events file at the top level of the dataset. This events file says that the stimulus starts at 5 seconds and ends at 300 seconds. Since everyone is presented with the same stimulus and participant responses are not logged, the same events file applies to all fMRI runs in the dataset (hence why it's at the top level of the dataset). However, we discover with CuBIDS that one fMRI run only contains 150 seconds of data (maybe the participant got out to go to the bathroom?). In that scenario, the events file applies to all runs in the dataset except that run, for which it is inaccurate. What we would want is to account for that in CuBIDS.

The answer might be as simple as adding a step to disable the inheritance principle before CuBIDS curation (see #387) and keep the rest of the workflow as-is, but it's worth discussing.