Harmonize ontology identifier with Interlex

[aeidi89]

Suggestion to change "preferredOntologyIdentifier" to a list of all ontologyIdentifiers, and thus harmonize with how it is displayed in Interlex (example link: https://neuinfo.org/interlex/view/ilx_0751626?searchTerm=alzheimer).

[apdavison]

I don't understand the suggestion. Interlex only shows a single "preferred identifier". It has a list of all identifiers, but only one of them is marked as preferred.

[aeidi89]

That is true @apdavison, the idea is to have a list of 'ontologyIdentifiers', instead of or in addition to the 'preferredOntologyIdentifier'. This came from a case where openMINDS, Interlex and Knowledge Space all referred to different preferred ontology identifiers.

[apdavison]

I agree it would be a good idea to add an array property ontologyIdentifier in addition to preferredOntologyIdentifier.

[lzehl]

Fine with me. But this issue needs to move then to the repo where the schema is defined (openMINDS_controlledTerms).

[lzehl]

TBD (@apdavison @tgbugs @UlrikeS91 @spieschnik @Raphael-Gazzotti ):

old

• preferredOntologyIdentifier
• InterLexIdentifier (THIS ONTOLOGY ID)
• KnowledgeSpaceLink (IRI)

new?

• ontologyIdentfier (one or list?)
• KSLink (IRI) [or is this service link?]

[UlrikeS91]

I think it is generally good to reconsider how we organize these property names. I just need to make this a bit more complicated 😉

We need to additionally consider:
1. SANDS atlas schemas - They already have the property "ontologyIdentifier", which allows a single IRI. I'm not sure if these atlas elements could receive mulitple ontology IDs, but I guess they can? If it is generally the case, that an entity can have multiple ontology IDs assigned, we should always allow an array of IRIs. We can restrict it to a single one under "ontologyIdentifier" as well (instructions would specify that e.g., the preferred ID should be stated), but I do not see a got reason to mix this (i.e., allowing sometimes an array and somtimes only one IRI).

2. ControlledTerms terminologies linking to other resources as their "preferredOntologyIdentifier" - I just remembered this because of issue #174. There are several terminologies that link to Wikidata entries (found another example in OpertingDevices. Maybe we could consider these for the renaming as well? Either change the name of the property to include it somehow add a property to the ControlledTerms schemas?

Otherwise, I think it is accurate to consider KS links as service links. But I have a general question about this: Where do we draw the line for something to be considered a service? I feel like InterLex may also be considered a service. Is it not?

Sidenote, if we decide to use ServiceLinks, the schema needs updates to fit this purpose.

[lzehl]

@openMetadataInitiative/openminds-developers

We may want to keep a single property for "ontologyIdentifier" which is an embedded schema providing the ID and the ontology service to use to extract that term from the ontology (entry point for queries in the ontology)?