Collection name missing in object records

[staplegun]

id: "113163",
type: "object",
title: "The Stool Journals",
collection: {
  id: "4193",
  type: "Collection"
},

[Conal-Tuohy]

Snipped to exclude a lot of other properties, but the "included_in" property is there now:

{
  "@context": "/context.json",
  "id": "http://nma-dev.conaltuohy.com/object/113163#",
  "type": "PhysicalObject",
  "included_in": {
    "id": "http://nma-dev.conaltuohy.com/collection/4193#",
    "type": "http://www.cidoc-crm.org/cidoc-crm/E78_Collection",
    "identified_by": "http://nma-dev.conaltuohy.com/collection/4193#repositorynumber",
    "documented_in": "http://nma-dev.conaltuohy.com/collection/4193",
    "label": "Turner and Valentine Families collection",
    "see_also": "http://nma-dev.conaltuohy.com/object?collection=%22Turner%20and%20Valentine%20Families%20collection%22"
  },
  "label": "The Stool Journals"
}

The "simple" JSON is still missing details though:

{
      "collection": {
        "id": "4193",
        "type": "Collection"
      }
}

[Conal-Tuohy]

ready to deploy to production

[f27wood]

Record updated to test tomorrow https://data.nma.gov.au/object/235906

BTW, this record should have an educational significance but it doesn't - it was added on Friday last week (the 17th)... so also check tomorrow if the educational significance comes across.

[f27wood]

I also just updated the above record: 113163 so we have two test records. Note to self - remember to change them back

[f27wood]

Testing this, it looks like the collection info is missing altogether from the simple format. Looks OK for LD format.

[Conal-Tuohy]

... which points the finger at the trix-to-json.xsl stylesheet.

[staplegun]

I think this is our first major victim of following a schema that is still evolving.

In March, the Linked Art website didn't specify how to model objects being inside a curated collection. However there was a Getty Cookbook example using included_in (crm:P106i_forms_part_of). This is the version we modeled in our mapping emu-to-crm and trix-description-to-dc. https://linked.art/cookbook/getty/photoarchive/

In April, the Collections page on the Linked Art website was updated to prefer aggregated_by (ore:isAggregatedBy). However we did not notice this change. https://linked.art/model/collection/

Recently collection titles disappeared (this issue). In an attempt to fix this we mistakenly changed the mapping in emu-to-crm to part_of (crm:P46i_forms_part_of), but not in trix-description-to-dc, which is why collection details now appear in the JSON-LD but not at all in the simple format.

We will need to test this a bit further. If we revert to crm:P106i_forms_part_of we still need to determine why the titles are missing. If we change to the now preferred ore:isAggregatedBy, we will need to check other parts of the mapping are not affected as this predicate is also used elsewhere (we need to double-check the aggregator found is a collection and not something else).

[Conal-Tuohy]

I vote we stick with crm:P46i_forms_part_of and just fix the trix-description-to-dc.xsl.

[Conal-Tuohy]

Done. See http://nma-dev.conaltuohy.com/object/31726#

[f27wood]

Tested AOK in dev, moving to done for final testing once deployed to prod.