The documentation for the sister repo spine-generic currently relies on an archive called results.zip attached to data-multi-subject releases. However, recent releases do not have a results.zip file, which has caused spine-generic documentation failures for 2+ years (see https://github.com/spine-generic/spine-generic/issues/268).
Given the contents of results.zip, it seems like the archive was created manually by bundling up files produced by Spine Generic's Analysis Pipeline run on data-multi-subject. So, if we want to keep this process up-to-date, I think the most useful way to eliminate this manual effort would be to write down the steps required to generate results.zip, then automate them by putting them in a GitHub Actions workflow.
Some notes:
data-multi-subject has changed quite a bit since the last time results.zip was generated (renamed file suffixes, etc.). So, the old way of generating results.zip may not work right away, and could require some changes to the spine-generic repo.
Running a processing script, selecting output files, adding them to a zip, and uploading them as a release asset are all things we've done before (see: sct_tutorial_data/run_script_and_create_release.yml). So, it shouldn't be too difficult to re-use some of the logic from that workflow here.
The documentation for the sister repo
spine-genericcurrently relies on an archive calledresults.zipattached todata-multi-subjectreleases. However, recent releases do not have aresults.zipfile, which has causedspine-genericdocumentation failures for 2+ years (see https://github.com/spine-generic/spine-generic/issues/268).Given the contents of
results.zip, it seems like the archive was created manually by bundling up files produced by Spine Generic's Analysis Pipeline run ondata-multi-subject. So, if we want to keep this process up-to-date, I think the most useful way to eliminate this manual effort would be to write down the steps required to generateresults.zip, then automate them by putting them in a GitHub Actions workflow.Some notes:
data-multi-subjecthas changed quite a bit since the last timeresults.zipwas generated (renamed file suffixes, etc.). So, the old way of generatingresults.zipmay not work right away, and could require some changes to thespine-genericrepo.sct_tutorial_data/run_script_and_create_release.yml). So, it shouldn't be too difficult to re-use some of the logic from that workflow here.