Figure naming incorrect when t2s-coreg option used in v20.0.7

[mnarayan]

This report pertains to a bug that appeared between 20.0.0 and 20.0.7. It seems the figures are incorrectly named when processing multi-echo datasets.

• I expected figures to contain only the combined multi-echo data since t2s-coreg option was used. In version 20.0.0 this is how the files were named.

 3.0M May 15 15:59 sub-CCxxxx_task-movie_run-1_desc-bbregister_bold.svg
159K May 15 17:05 sub-CCxxxx_task-movie_run-1_desc-carpetplot_bold.svg
 100K May 15 16:02 sub-CCxxxx_task-movie_run-1_desc-compcorvar_bold.svg
 438K May 15 16:03 sub-CCxxxx_task-movie_run-1_desc-confoundcorr_bold.svg
 495K May 15 16:03 sub-CCxxx_task-movie_run-1_desc-rois_bold.svg

• Instead now in version 20.0.7, I see the following, labeled echo-1.

May 18 11:32 sub-CCxxx_task-movie_run-1_echo-1_desc-aroma_bold.svg
May 18 00:55 sub-CCxxx_task-movie_run-1_echo-1_desc-bbregister_bold.svg
May 18 01:59 sub-CCxxx_task-movie_run-1_echo-1_desc-carpetplot_bold.svg
May 18 00:55 sub-CCxxx_task-movie_run-1_echo-1_desc-compcorvar_bold.svg
May 18 00:55 sub-CCxxx_task-movie_run-1_echo-1_desc-confoundcorr_bold.svg
May 18 00:55 sub-CCxxx_task-movie_run-1_echo-1_desc-rois_bold.svg

My suspicion is this has to do with naming conventions for the fmriprep working directory. This uses the naming convention of task-xxxx_run-1_echo-1 regardless of whether an echo-specific option is used --echo-id 1 or whether the joint echo processing option is being done --t2s-coreg is used.

[oesteban]

Thanks much - this is a use case I didn't capture while I was at #2114. The new dismiss_entities feature (nipreps/niworkflows@73f54c1) will come handy here to drop the echo tag.