[WIP] Create adni json from xml metadata

[omar-rifai]

Module to allow integration of xml metadata to the adni-to-bids clinical data converter

[pep8speaks]

Hello @omar-rifai! Thanks for updating this PR. We checked the lines you've touched for PEP 8 issues, and found:

• In the file clinica/iotools/converters/adni_to_bids/adni_utils.py:

Line 580:23: W605 invalid escape sequence '\d'

Comment last updated at 2022-02-22 09:10:05 UTC

[omar-rifai]

Hello @emaheux. Please let me know if this would satisfy the requirements for leaspy.
The few modifications I made over your code is:

• Running the xml metadata extractor when the argument --xml_path is given
• Removing the use of the multiprocessing library (in the objective of switching to a multithreading/asyncio library).

[omar-rifai]

I have kept the resulting file with all the data extracted separate from the BIDs datastructure but could have the write function add the data next to the images if necessary.

[emaheux]

Thank you Omar :)

I have kept the resulting file with all the data extracted separate from the BIDs datastructure but could have the write function add the data next to the images if necessary.

It would have been nice to have this metadata as part of the BIDS structure (metadata of scans tsv? - or at least in merged tsv BIDS.tsv?) but if it is not possible / too complicated it will be fair enough to load this extra json file!

[omar-rifai]

Thank you Omar :)

I have kept the resulting file with all the data extracted separate from the BIDs datastructure but could have the write function add the data next to the images if necessary.

It would have been nice to have this metadata as part of the BIDS structure (metadata of scans tsv? - or at least in merged tsv BIDS.tsv?) but if it is not possible / too complicated it will be fair enough to load this extra json file!

It should be doable! I was just wondering what would be best. I'll keep you posted.
(for merge tsv it should appear automatically if added to the scan tsv file).

[omar-rifai]

Closed as no consensus found currently as to where this data should ideally be integrated. Merge-tsv should also see a refactoring soon (c.f issue 559) which goes against integrating the data there.

[omar-rifai]

Hi @emaheux, since @NicolasGensollen is working on similar topics, we were wondering if this is still relevant. I did not have the time to pursue the task but it can be an opportunity to pick up the work. Let us know.

[emaheux]

Indeed @omar-rifai, this is still relevant (though obviously not blocking / top-priority since we circumvented the issue long time ago by adding ADNI imaging metadata downstream by ourselves with the codebase I gave you).

Yet if a Leaspyer needs to conduct experiments with a newer version of ADNI, I think it would be really nice / more consistent to have those imaging metadata directly accessible in Clinica outputs!