Define an approach for enforcing ontologies in the different shapes

[MFSY]

The usage of controlled vocabularies (instead of plain text) as values of properties is a highly recommended best practice. The reasons are multiple but the most important ones are:

• Consistent (way less ambiguous) annotation of dataset between organizations and within communities => they can speak the same language (use the same vocabulary) so to say.

• The vocabulary used in a given domain of application to describe entities, agents, activities,... becomes an explicit and normalised artefact that can be shared and maintained.

Now how do we enforce the usage of ontologies/vocabularies in the data models defined in neuroshapes ?

Well let first recognise the fact that it is unlikely one can identified upfront ontologies/vocabularies that will cover all possible used cases users may have. Then a good approach would be to provide a set of defaults ontologies/vocabularies that can be enforced but allow users to enforce other ontologies/vocabularies (with better coverage for example) and/or to just provide text in first place when no ontology/vocabulary is available for a given property value.

Currently, some ontologies/vocabularies needed are referenced in the typedlabeledontologyterm schema. Clearly it needs to be extended to support the above approach.

How then ? Let assume a shape for enforcing the usage of a brain region ontology needs to be created:

• let give it a (fragment) name: BrainRegionShape
• The BrainRegionShape can be defined as follows:

 {
      "@id": "this:BrainRegionShape",
      "@type": "sh:NodeShape",
      "label": "A brain region shape.",
      "or": [
        {
          "label": "The expected brain region value is an ontology/vocabulary term: identified by an IRI.",
          "node": "{{base}}/schemas/neurosciencegraph/commons/labeledontologyentity/v0.1.0/shapes/LabeledOntologyEntityShape"
        },{
          "label": "The expected brain region value is from a specific brain region ontology: identified by an IRI.",
          "node": "{{base}}/schemas/neurosciencegraph/commons/typedlabeledontologyterm/v0.1.0/shapes/BrainRegionOntologyTermShape"
        },{
         "label": "The expected brain region value is a free plain text.",
          "datatype": "xsd:string"
        }
      ]
    }

• and a property (nsg:brainRegion for example) value can be constrained in the following way:

{
          "path": "nsg:brainRegion",
          "name": "Brain region",
          "description": "A brain region shape.",
          "node": "{{base}}/schemas/neurosciencegraph/commons/typedlabeledontologyterm/v0.1.0/shapes/BrainRegionShape"
}

Related #42

[MFSY]

NIFSTD ontologies are clearly relevant here for:

• taxonomies (species, strain)
• brain regions/ brain parcels. See the issue here
• cell types

I think most of the defaults we'll provide should come from their.

@tgbugs can help may be ?

[tgbugs]

I will take a look.

[MFSY]

Great.

[tgbugs]

@MFSY is there a way that I can set up a dev server that will serve proper json-ld instead of the templates that are in this repo, and/or are there endpoints that I can access to get see it? Thanks!

[MFSY]

Hi @tgbugs. I'm working on generating a documentation for these schemas as github pages. Also they will be deployed through the BBP staging Nexus instance (a REST API) for now. I'll let you know when it's done.