IIIF integration (import/export)

[pleonard212]

Some thoughts on how PixPlot can integrate with IIIF repos. This is not connected to the annotation functionality - this is only about getting images in and out.

Import

• Parse a newline-delimited list of IIIF manifests instead of a folder of images and a csv.
• In case of multiple images per manifest: take the first image. Assume this is the "main" image of the set. (Surely a false assumption but good enough to prove IIIF functionality. We are shooting for a use case where the objects are likely to be well represented by the first image, and the second might be something like the back of the painting or an uncropped variant with color balance targets.)
• Take the label field of the manifest as the caption or title. (I believe label is the only field guaranteed to be present in a manifest.) Do not seek to create facets yet from the metadata (requires human judgement on a per-collection basis).

Export

• If the PixPlot was generated from IIIF manifests, show the IIIF logo image on the Lightbox view with a link either...
  • The raw IIIF manifest itself?
    • https://manifests.britishart.yale.edu/manifest/46796
  • The IIIF manifest as an argument to a Mirador instance?
    • http://mirador.britishart.yale.edu/?manifest=https://manifests.britishart.yale.edu/manifest/46796

The problem with the first option is that the manifest itself isn't useful unless you have a Mirador or other IIIF client running somewhere. The problem with the second is, there's no central, universal Mirador instance that I know of (besides something running on your localhost). It's not clear if for example YCBA would be cool with us directing all links to their Mirador...

This Export idea only lets you drag one IIIF manifest at a time, not the much-cooler export of whole subsets of the cluster.

[duhaime]

I suppose there wouldn't be harm in us hosting a mirador instance ourselves on GitHub or some other free static file server. Since the client would be doing the work it shouldn't tax the little server too much even if there were lots of traffic.

I'm thinking it will make sense to factor the data collection into its own module then just use that module in this codebase...