Create a git-annex repo for data from basel/6900_nifti_20

[plbenveniste]

The data is stored in duke/mri/basel/6900_nifti_20 and the segmentations in duke/mri/basel/6900_20_lesions-PD
The steps are the following:

• A git-annex repo should be created :
• The data should be Bidsfied
• The data should be pushed to the repo
• A PR should be opened
• The PR should be merged.

@jcohenadad What name should be used for the git-annex repo ? ms-basel-pd-6900 ?

Another strategy could be to merge this data with the other dataset under construction from basel (issue #19)

[plbenveniste]

In this case, the data is stored in duke/mri/basel/6900_nifti_20. The files are already .nii.gz files. Unfortunately, the folder doesn't contain the images' json sidecar. The data spans from 2017 until 2020. Therefore we chose to use the git annex repo name ms-basel-2020 (discussed here). Each subject contains multiple acquisitions for two chunks (cervical = chunk I ; thoracic = chunk II).

As for the segmentation, the segmentations also don't have the a json sidecar. They were done for both thoratic and cervical chunks but only for sag pd images. The segmentations are also stored as nii.gz files.

@jcohenadad Should I consider that the contacts for this data is the same as the one for ms-basel-2020 ? (You should note that this time, the number is 6900) And, was it again Haris Tsagkas who segmented the the lesions ?

@mguaypaq Could you create the git-annex repo ms-basel-2020 for this dataset?

[jcohenadad]

@jcohenadad Should I consider that the contacts for this data is the same as the one for ms-basel-2020 ? (You should note that this time, the number is 6900) And, was it again Haris Tsagkas who segmented the the lesions ?

yes and yes. Please also refer to this discussion in the README.

[plbenveniste]

The script bidsification.py was wrote to do the following:

• create the necessary files and folders to have the data in the BIDS format.
• deal with the segmentation files and create the json sidecar files.
• create the participants.tsv and participants.json files.
• create the dataset_description.json file.
• fix the header of the segmentation files that are not in the same orientation as the image files.
• convert the files to uint8 if they are not already.

The README.md was completed.

Ready to be pushed when repo is created.

[plbenveniste]

Code was updated to use sct_register_multimodal similarly to what was done in #19 and to update the json sidecars.

The code was run using:

python ms-basel-2020/code/bidsification.py --i 6900_nifti_20/ --o ms-basel-2020/ --s 6900_20_lesions-PD/

Ready to be pushed when repo is ready.

[mguaypaq]

The repo is created, and @plbenveniste has write access:
https://data.neuro.polymtl.ca/datasets/ms-basel-2020

[plbenveniste]

PR was opened and it is ready for review !

[plbenveniste]

The naming conventions were reviewed (by me) and everything looked fine !
PR is ready for review