[BUG] Complete subject ids in 08MQ group level analysis

[bclenet]

Describe the bug

The group level analysis of 08MQ's pipeline might not sort the subject ids properly between equalRange and equalIndifference groups.

One needs to check that using the elements_in_string with only the short ids (i.e.: "001", "002", ...) is sufficient to identify the right copes and varcopes. For example we nee to be sure that files containing these patterns for other things than subject id do not become false positives.

Reproduction

Run group level analysis of 08MQ and analyse reports from nodes responsible for group sorting.

Expected behavior

This is what has been done in pipeline C88N to solve the issue :

# Create a function to complete the subject ids out from the get_equal_*_subjects nodes
#   If not complete, subject id '001' in search patterns
#   would match all contrast files with 'con_0001.nii'.
complete_subject_ids = lambda l : [f'_subject_id_{a}' for a in l]

NARPS Open Pipeline version

commit 6e48177

Operating System

Softwares versions

Screenshots

No response