[BUG] In adni-to-bids, empty sessions files are not cleaned up

[AliceJoubert]

While working on PR #1188 I noticed that even if all images of a session are deleted for a modality the folder specific for the modality is still kept. As an example, for the ADNI subject 002_S_1261 :

(clinicaEnv) alice.joubert@ICM-COLLI-MP013 testing % tree
.
├── README
├── conversion_info
│   └── v0
│       └── fmap_paths.tsv
├── dataset_description.json
├── participants.tsv
├── sub-ADNI002S1261
│   ├── ses-M096
│   │   ├── fmap
│   │   └── sub-ADNI002S1261_ses-M096_scans.tsv
│   ├── ses-M120
│   │   ├── fmap
│   │   │   ├── sub-ADNI002S1261_ses-M120_fmap_e1.json
│   │   │   ├── sub-ADNI002S1261_ses-M120_fmap_e1.nii.gz
│   │   │   ├── sub-ADNI002S1261_ses-M120_fmap_e2.json
│   │   │   ├── sub-ADNI002S1261_ses-M120_fmap_e2.nii.gz
│   │   │   ├── sub-ADNI002S1261_ses-M120_fmap_e2_ph.json
│   │   │   └── sub-ADNI002S1261_ses-M120_fmap_e2_ph.nii.gz
│   │   └── sub-ADNI002S1261_ses-M120_scans.tsv
    └── sub-ADNI002S1261_sessions.tsv

The files got deleted for session M096 but the fmap directory was kept. This might be something we want to clean up.

[NicolasGensollen]

Thanks for opening this @AliceJoubert !

I agree that we should cleanup those empty folders. Here you show an example at the modality level, but I suppose the problem could also happen at the session, or subject level.
However, we would need to make sure that the different files have consistent metadata after removal.

Ideally, this should be a generic function taking a BIDS dataset as input and doing the necessary cleaning steps. We could then have this function called at the end of each converter to post-process the generated BIDS dataset.

[github-actions]

This issue is considered stale because it has not received further activity for the last 14 days. You may remove the inactive label or add a comment, otherwise it will be closed after the next 14 days.

[NicolasGensollen]

@AliceJoubert is this something that you are planning to work on in the coming days / weeks or should we move it to the 0.10.0 milestone ? (no problem to do that if we have more urgent things do do)

[AliceJoubert]

@NicolasGensollen Definitely not in the incoming days ! I want to finish the documentation first I think