Closed bkossows closed 5 years ago
For the ongoing project affected by this issue (which has two sessions with same subject names but different sets of tasks) we can mitigate it by trying to parse task names & run numbers from the dicom series name. I can do it this week.
edit: but then if we have a "rigid" assignment of task & run, we must set some rules for handling repeated runs; otherwise Heudiconv will crash with a runtime error in convert.py:
RuntimeError: was asked to convert into /full/path/sub-<>_task-<>_run-<>_bold.nii.gz but destination already exists
To solve it in general - I think we would need to check if a given subject is already present in the reports folder and change subject's name during conversion?
Good job! Now we can deal with different fMRI sessions of the same subjects. We can leave other cases for further discussion and/or pay attention to unique ids for longitudinal studies.
It looks like we still need to find a workaround for multiple sessions. I'd opt for sub-XXX/ses-DATE structure. In dicominfo.tsv there is already a 'date' field available.
Currently we do not separate subsequent sessions of the same subject but overwrites previous one. One should pay attention to use unique subjects ids.