Fix generated SHACL to allow subclasses to satisfy constraints

[JPEWdev]

The model that is produced by the spec parser is not allowing sub classes to satisfy the requirements of a property.

For example, core:CreationInfo.createdBy should allow core:Agent or any subclass derived from it (e.g. core:Organization), but instead it only allows an exact core:Agent

This is a problem with both the generated TTL and JSON-LD model file.

[zvr]

Confirming the issue.

With an example test data file:

@prefix core: <https://spdx.org/rdf/v3/Core/> .
@prefix ex: <http://zvr.invalid/rdf/> .

ex:someOrganization a core:Organization .
ex:someCreationInfo a core:CreationInfo .
ex:someCreationInfo core:createdBy ex:someOrganization .

checking the PySHACL validation with pyshacl -s model.ttl example-data.ttl produces:

Constraint Violation in ClassConstraintComponent (http://www.w3.org/ns/shacl#ClassConstraintComponent):
        Severity: sh:Violation
        Source Shape: [ sh:class core:Agent ; sh:minCount Literal("1", datatype=xsd:integer) ; sh:name Literal("createdBy") ; sh:path core:createdBy ]
        Focus Node: ex:someCreationInfo
        Value Node: ex:someOrganization
        Result Path: core:createdBy
        Message: Value does not have class core:Agent

[zvr]

So the problem is not in the rdfs:range declaration, but in the generated SHACL.

[zvr]

The corresponding SHACL part of the generated model is:

core:CreationInfo a sh:NodeShape ;
    sh:property   [ sh:class core:Agent ;
            sh:minCount 1 ;
            sh:name "createdBy" ;
            sh:path core:createdBy ],
...

[kestewart]

@zvr - any update on when this is resolved. Getting questions from @JPEWdev on the call as he has too much SPDX to manually validate.

[zvr]

This refers to the old parser output, no longer relevant.

Nevertheless, I leave this open to remind me to check this on the new generated output, just in case.