Add whole-spine dataset

[sandrinebedard]

Description

This PR adds normative values of the whole-spine dataset for T2w images.

The spinal cord segmentations were generated with contrast-agnostic segmentation v3.1 and manually corrected. (for now manual corr are in our internal dataset on a branch, not yet on open-neuro)

This PR adds the CSV files and added the subjects to the participants.tsv file.

What's left

• One question remains, should we check for mild compression in the dataset? I noticed some subjects in the pathology column have n/a, do they have pathology??

• I added the script used to curate the csv files and participants.tsv under a new folder code, it should proably got under .gitignore afterwords

• We should recreate the plots of the normative values?

[valosekj]

• One question remains, should we check for mild compression in the dataset? I noticed some subjects in the pathology column have n/a, do they have pathology??

Yeah, I think it would be a good idea. To be consistent with the spine-generic dataset.

• I added the script used to curate the csv files and participants.tsv under a new folder code, it should proably got under .gitignore afterwords

Nice! So the idea is to have a single participants.tsv file containing subjects across different datasets, right? Alternatively, we could have one participants.tsv file for each dataset, e.g., participants_spine-generic.tsv and participants_whole-spine.tsv. But tbh, I do not know what the pros and cons of these two options are. So a single participants.tsv file is probably fine.

• We should recreate the plots of the normative values?

Yeah, this would be great. When would we put the updated figures? To the interactive preprint? Or somewhere else? Alternatively, we could build a simple readthedocs website for the figures (as done for spine-generic). This would allow us to recreate the figures automatically using GitHub actions with each new dataset.

[sandrinebedard]

Nice! So the idea is to have a single participants.tsv file containing subjects across different datasets, right? Alternatively, we could have one participants.tsv file for each dataset, e.g., participants_spine-generic.tsv and participants_whole-spine.tsv. But tbh, I do not know what the pros and cons of these two options are. So a single participants.tsv file is probably fine.

I think it will be easier to use in a single participants.tsv file, for like in sct_compute_compression to get the sex and age info for example, also considering that the columns names may differ across dataset

[valosekj]

What script did you use to process whole-spine? The same as for spine-generic?
https://github.com/sct-pipeline/dcm-metric-normalization/blob/r20250320/scripts/process_data_spine-generic.sh

[sandrinebedard]

This is the script: https://github.com/sct-pipeline/dcm-metric-normalization/blob/sb/add-canal/scripts/process_data_whole-spine.sh

But I want to clean it up and move to another branch before merging, (like this also includes T1w)