Feature Request: Support server-side zipping of select files to reduce number of files in Dataset

Overview of the Feature Request

In some cases, the number of files in a dataset might become huge. Either by a huge onetime deposit or if files are continuously added for long running measurements etc.

To reduce the number of files again, it would be good to allow a (API-only?) request by a curator or higher of a dataset to do a server-side compression of a selection of files into a ZIP file, and replace the files in the dataset with the single ZIP file.

What kind of user is the feature intended for?
Curator, Depositor, Superuser, Sysadmin

What inspired the request?

We have folks from https://www.iagos.org/ over here and they are asking about how to properly archive the data portal or continuous snapshots of the (to date) 17.000 flights they did. Being able to keep on depositing and wrap a number of files into a ZIP at some time interval would be a great benefit.

Also, with regards to the new feature of file count quotas per dataset (#11275) this would enable to drive large datasets back into quota limits.

What existing behavior do you want changed?
There is no way to replace files in a dataset with a ZIP bundle except when downloading the files and reuploading as ZIP. This roundtrip can be saved when doing it on the server side.

Any brand new behavior do you want to add to Dataverse?
Kind of - zipping server side was before used to create downloads of datasets, but it's buggy. Replacing files with a ZIP bundle is different to that.

Any open or closed issues related to this feature request?

• #11275
• #8029
• #10723
• #10591

Are you thinking about creating a pull request for this feature?
Maybe. This is mostly a braindump.

No implementation files or tests are named. Start by reviewing the existing server-side zipping used for dataset downloads and related issues #11275, #8029, #10723, and #10591; clarify the API, authorization, file-selection, and replacement requirements. Done means a defined and tested workflow can replace selected dataset files with one ZIP without a download-and-reupload round trip.