couldn't download neurodesk-freesurfer--8.0.0.simg

[yibeichan]

Hi,

I'm trying to use neurodesk-freesurfer--8.0.0.simg from here via datalad.

here is my command

datalad run -m "Freeze neurodesk freesurfer container version" \
  code/containers/scripts/freeze_versions --save-dataset=. neurodesk-freesurfer=8.0.0

datalad containers-run \
    --explicit \
    -n neurodesk-freesurfer \
    --input "${SUBJECT_INPUT}" \
    --output "${SUBJECT_OUTPUT}" \
    '{inputs}' '{outputs}' participant \
    --participant_label "${SUBJECT_ID##*-}" \
    -w "workdir/${SUBJECT_ID}"

but then I get the following error

[https://d15yxasja65rk8.cloudfront.net/freesurfer_8.0.0_20250127.simg download failed: Forbidden
downloading from all 1 known url(s) failed
https://d15yxasja65rk8.cloudfront.net/freesurfer_8.0.0_20250127.simg download failed: Forbidden
downloading from all 1 known url(s) failed
https://d15yxasja65rk8.cloudfront.net/freesurfer_8.0.0_20250127.simg download failed: Forbidden
downloading from all 1 known url(s) failed]

If I directly do wget https://d15yxasja65rk8.cloudfront.net/freesurfer_8.0.0_20250127.simg, it fails too.

The above error appears on a cluster (Rocky 8); the same error also occurs on my laptop (MacOS, intel).

Thank you!

[stebo85]

i am looking into it

[stebo85]

freesurfer_8.0.0_20250127.simg had a bug and is deactivated. You need to use the latest release, which is freesurfer_8.0.0_20250210.

wget https://d15yxasja65rk8.cloudfront.net/freesurfer_8.0.0_20250210.simg

[stebo85]

@yarikoptic - do you see a reason why the datalad repository is not updating when we push a new release?

I guess it's because it's technically the same version of freesurfer (8.0.0) - we had to rebuild it because it was missing libraries. Any idea how we could update these in datalad in a clean way?

[yarikoptic]

ATM we are not verifying anyhow that URLs changed and we skip an image if already has image known: https://github.com/ReproNim/containers/blob/master/scripts/create_singularities#L167 .

Do you think we should strive in general keep the list of registered URLs "in sync"? ATM we have hardcoded location -- https://d15yxasja65rk8.cloudfront.net so we would then filter/manage only URLs which start with that , and add new/remove new and remove no longer valid. Right?