[
https://issues.apache.org/jira/browse/RYA-293?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16120123#comment-16120123
]
ASF GitHub Bot commented on RYA-293:
------------------------------------
Github user jessehatfield commented on the issue:
https://github.com/apache/incubator-rya/pull/180
So this is a case where a few different schema terms (rdfs:subClassOf,
owl:unionOf, and eventually owl:equivalentClass) end up being represented by
just one term internally (subclass/superclass relationships). We could
plausibly organize by either; my intuition is to use the internal
representation, since the internal graph being complete and accurate can matter
to the other rules. That would mean pulling the logic introduced here, plus the
logic introduced in [https://github.com/apache/incubator-rya/pull/184](PR 184)
, plus the logic for rdfs:subClassOf (which 184 incidentally simplifies to a
method call anyway) into some "refreshSubClassGraph" method. Thoughts on that
approach? Would probably do the same with subPropertyOfGraph for consistency
(though it's simpler because there's no equivalent to union).
> Implement owl:unionOf inference
> -------------------------------
>
> Key: RYA-293
> URL: https://issues.apache.org/jira/browse/RYA-293
> Project: Rya
> Issue Type: Sub-task
> Components: sail
> Reporter: Jesse Hatfield
> Assignee: Jesse Hatfield
>
> An *{{owl:unionOf}}* expression defines one type to be equivalent to the
> union of another set of types. If the ontology states that {{:Parent}} is the
> union of {{:Mother}} and {{:Father}}, then the inference engine should
> rewrite statement patterns of the form {{?x rdf:type :Parent}} to check for
> resources that are stated to be any of the types {{:Mother}}, {{:Father}}, or
> {{:Parent}}.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
