buda-base / buda-iiif-presentation

iiif presentation (collections and manifests) generation
Apache License 2.0
0 stars 0 forks source link

Relocate repository for ImageInfoList service #115

Open jimk-bdrc opened 3 years ago

jimk-bdrc commented 3 years ago

As part of researching how to fix archive-ops 506, I realize we should think about using a different repository from the images to store information about them. There are two reasons for this:

We could just use a different folder under the same work (meta/ or a new one, perhaps manifest The work would be straight forward, just simultaneously changing /images/ with /manifest (or equivalent) in:

  1. buda-iiif-presentation.ImageInfoListService.java
    return "Works/" + md5firsttwo + "/" + imageInstanceLocalName + "/images/" + imageInstanceLocalName + "-" + imageGroupLocalName + "/dimensions.json";
  1. change the Python volume_manifest_builders generators' ImageRepositoryBase.getPathFromLocators from
        return f'Works/{two}/{work_Rid}/images/{work_Rid}-{image_group_folder_name}/'

There's a little more work on the Python side, because v_m_b assumes its input and output folders are the same, but that's not big deal.

There will be a short migration process while we move all the existing dimensions.json and 'fileList.json` to their new homes.

wdyt?

TBRC-Travis commented 3 years ago

yes, I think this is a good idea. i'm in favor of storing them in the meta folder.

eroux commented 3 years ago

sure, works for me

jimk-bdrc commented 3 years ago

Juste realized, we'll have to add them all before Elie can use it. Scheduling is the issue.

jimk-bdrc commented 2 years ago

A better way would be to leave them out of the images stream, and only recombine them when DIPping to the s3://archive.tbrc.org/Works stream.

See buda-base/volume-manifest-builder/48