search

Sveino / Inst4CIM-KG

Instance of CIM Knowledge Graph
Apache License 2.0
4 stars 1 forks source link

prof:hasArtifact use of xsd:anyURI #89

Open Sveino opened 1 month ago

Sveino commented 1 month ago

Due to the repose on https://github.com/opengeospatial/ogc-geosparql/issues/553 should we update:

https://github.com/Sveino/Inst4CIM-KG/blob/8a1846e8f766dd89a2ff5142d597344326b343ac/source/CGMES-NC/r2.3/ap/rdf/AssessedElement-AP-RDFS2020.rdf#L34

prof:hasArtifact https://ap-voc.cim4.eu/AssessedElement/2.3/AssessedElement-AP-Voc-RDFS2020.rdf^^xsd:anyURI ;

Bit related to https://github.com/Sveino/Inst4CIM-KG/issues/63

VladimirAlexiev commented 1 month ago

Nicholas Carr wrote

hasArtifact is expected to indicate a non-RDF resource

But the spec has Usage Note: "A property to link from a Resource Descriptor to an actual information resource (rdfs:Resource; an individual) that implements it": rdfs:Resources are nodes in an RDF graph

Nicholas gives an example where a resource descriptor links to a semantic resource:

ex:pr-01 a prof:ResourceDescriptor ;
    rdf:value ex:shapes-02

I think that CIM profiles shouldn't point only to a specific RDF rendition

Sveino commented 3 weeks ago

I have to say that I did not understand everything. However, I think Nicholas argument for using xsd:anyURI make a lot of sense. However, I also agree with Vladimir. Let me start the latest point first:

I think that CIM profiles shouldn't point only to a specific RDF rendition

  • They should point to all renditions (rdfxml, turtle, json) using anyURI literal (or URL, depending on preference)

  • They should also use rdf:value to link to the corresponding semantic URL Agree, but as part of importing the artifact we only need one reference. Yes, this could be referring to the dataset - but in the end it need to be a distribution that is imported. We should create a dct:Catalog that list all the profile distribution. So I am thinking that in rdfxml distribution of AP we will refere to the rdfxml distribution of all the relevant prof:ResourceDescriptor.

I do not understand the reference to a semantic resource as Valdimir is pointing out. Does this mean that we should use rdf:value rather than hasArtificats when it is an RDF resource?

I think it make sense to use xsd:anyURI when it is non-RDF resource as the case is for: https://github.com/Sveino/Inst4CIM-KG/blob/cecaf7495d71e72fa634596ca5559a7c6f129f10/source/CGMES-NC/r2.3/ap/rdf/AssessedElement-AP-RDFS2020.rdf#L67C1-L74C20

griddigit-ci commented 3 weeks ago

I think we need to discuss this. I am having a bit of a trouble to follow :)

VladimirAlexiev commented 3 weeks ago