Validation query for hasCommunicationAddress

[Jamie-SA]

I was working on queries to validate the new restrictions on various changes made in this release but after working on them a bit I realized that these types of checks would be better to do in SHACL. If OntoTool can do SHACL validation as part of the migration process I could create some shapes to enforce the new restrictions. Shapes could also be added later.

This PR has a couple of queries that can be included.

[sa-bpelakh]

onto-tool can handle SHACL on local execution (see https://github.com/semanticarts/gist/blob/develop/bundle.yaml#L59), but since there is no uniform way to invoke SHACL across different triple stores, there is no support for endpoint SHACL validation. For that reason I have gone with SPARQL for the migration, so that the queries can be reused.

Good catch not assuming that triple store queries the union of all named graphs, I need to add a version of my detect query that looks inside graphs.

I would simplify the detection query as follows:

CONSTRUCT {
    ?report a sh:ValidationReport ;
        sh:conforms false ;
        sh:result
            [
                a sh:ValidationResult ;
                sh:focusNode ?failedEntity ;
                sh:resultMessage "Domain restriction on gist:hasCommunicationAddress violated" ;
                sh:resultSeverity sh:Warning ;
                sh:sourceConstraintComponent <urn:constraint:communicatonAddress-domain>
            ] .
}
WHERE {
    { ?failedEntity gist:hasCommunicationAddress ?addr . }
    FILTER NOT EXISTS {
        { ?failedEntity a gist:Person } UNION { ?failedEntity a gist:Organization }
    }
    bind(<urn:new-domain-validation-report> as ?report)
}

There is no reason to include the URI in the error message, since it's automatically included in the report as the focusNode.

Since the detection queries are run via the following section of the script, if you prefix the file names with detect_, you won't need to make any additional changes.

- action: "verify"
  message: "Find remaining problems"
  type: "construct"
  inference: "none"
  stopOnFail: false
  target: '{report}'
  endpoint:
    query_uri: "{endpoint}"
    user: "{user}"
    password: "{password}"
  queries:
    source: '{migration}/v10.0'
    includes:
      - 'detect_*.rq'