Subfolders of omex archives should not be listed in manifest, but containing files should be

[bilalshaikh42]

@jonrkarr could you confirm that my interpretation is correct?

Hello,
It seems that if there are subfolders in a project, they are being zipped into their own combine archive, and then the archive is being listed in the manifest with a format type of "None". Instead, the contained files should be included in the top-level manifest directly.

I spotted this issue in the following exposure:

https://staging.physiomeproject.org/e/5f3/download_generated_omex

[jonrkarr]

Generally, I think COMBINE archives shouldn't recursively contain more archives. Instead, archives can contain subdirectories, in which case their contents should be described in the manifest (with locations that begin with paths to their parent directories).

Convention is that directories shouldn't be directly described in manifests; rather directories are implicit in the locations of the files. I don't think there's a clear way to directly describe directories in manifests because there's no appropriate format URI.

That said, I wouldn't necessarily consider a nested COMBINE archive to be invalid. Nested archives could be described in manifests with format="http://identifiers.org/combine.specifications/omex". I wouldn't encourage a nested design because the benefits aren't obvious to me, but I think it would be fine if there was a good reason.

For Physiome, we could flatten this out or encourage the PMR team to do so.

[jonrkarr]

One way this might be arising with Physiome is if the author submitted a COMBINE archive, which wasn't unpacked into the Git repo, and then PMR's software packages that submitted archive into the archive that the software automatically generates.

The same issue appears in BioModels. In that case, we've similarly encouraged BioModels to flatten out these archives that authors submitted.

[bilalshaikh42]

Looking at the archive more closely, that appears to be exactly what happened. The archive must have been included as its own file in the workspace.

…

On Wed, Dec 1, 2021 at 1:32 PM Jonathan Karr ***@***.***> wrote: One way this might be arising with Physiome is if the author submitted a COMBINE archive, which wasn't unpacked into the Git repo, and then PMR's software packages that submitted archive into the archive that the software automatically generates. The same appears in BioModels. In that case, we've similarly encouraged BioModels to flatten out these archives that authors submitted. — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#8 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AHX4FID6ZTGWFSFC3PDS2FTUOZS2XANCNFSM5JFEUJRA> .

[jonrkarr]

I think its up to the PMR curators to decide whether its ok for authors to do this, or to flatten this out.

If archives are recursively including in archives, their formats should be properly annotated in COMBINE manifests.

[nickerso]

Thanks for flagging this @bilalshaikh42. Just to confirm that in this specific example, the archive was mistakenly added to the workspace. (The archive is just a copy of the workspace content.) In general, this shouldn't happen any more and we are looking to remove any instances where that has been done, especially now that we have this feature to generate archives for all exposures.