Thanks Holger! That worked! I had to rebalance some ]'s but then it performed perfectly!
I'm still getting used to the syntax of the sh: operators. If I could trouble you for one more piece of advice. I realized that there are some instances of qudt:PhysicalConstant that also point to instances of qudt:QuantityKind, so I want to filter out anything that is not an instance of qudt:Unit. After exploring the documentation at https://w3c.github.io/shacl/shacl-af, I thought sh:filterShape might do the trick. The amended code is shown below. It isn't working - in fact it is weirdly now returning just the value of $this, i.e. the subject QuantityKind. Am I using sh:filterShape correctly? I'm trying to say "Look to see if anything is pointing to you via the hasQuantityKind property, and filter those down to just instances of qudt:Unit. If you find some, return those. If not, follow the skos:broader relation and try again with that QuantityKind." Here is the functioning code: qudt:QuantityKind-appropriateUnit rdf:type sh:NodeShape ; sh:property [ rdf:type sh:PropertyShape ; sh:path qudt:appropriateUnit ; sh:group qudt:AppropriateUnitsGroup ; sh:order "15"^^xsd:decimal ; sh:values [ sh:distinct [ sh:else [ sh:path qudt:appropriateUnit ; sh:nodes [ sh:path skos:broader ; ] ; ] ; sh:if [ sh:exists [ sh:path [ sh:inversePath qudt:hasQuantityKind ; ] ; sh:filterShape qudt:Unit ; ] ; ] ; sh:then [ sh:path [ sh:inversePath qudt:hasQuantityKind ; ] ; sh:filterShape qudt:Unit ; ] ; ] ; ] ; ] ; sh:targetClass qudt:QuantityKind ; . I realized that I need to Steve On Wed, Oct 28, 2020 at 7:39 PM Holger Knublauch <[email protected]> wrote: > > On 10/29/2020 11:51 AM, Steve Ray wrote: > > Hi, > I am trying to recreate a recursive magic property, using a SHACL property > shape. However, I suspect that the reason the recursion is not working is > because the property identified in the sh:path statement isn't really a > function in the same sense as a magic property is a function. > > My shape code is as follows: > > qudt:QuantityKind-appropriateUnit > > rdf:type sh:NodeShape ; > > sh:property [ > > rdf:type sh:PropertyShape ; > > sh:path qudt:dappropriateUnit ; > > sh:values [ > > sh:prefixes <http://qudt.org/2.1/schema/shacl/overlay/qudt> ; > > sh:select """SELECT DISTINCT ?unit > > WHERE { > > { > > ?unit qudt:hasQuantityKind $this . > > } > > UNION > > { > > NOT EXISTS { > > ?unit qudt:hasQuantityKind $this . > > } . > > $this skos:broader ?parent1 . > > ?parent1 qudt:dappropriateUnit ?unit . > > } . > > } """ ; > > ] ; > > ] ; > > sh:targetClass qudt:QuantityKind ; > > . > > The non-recursive part works (the first clause of the union). But the > second union clause does not work. Is there a way to make the > qudt:dappropriateUnit property know that there's this associated sh:values > clause? > > As you have found out, SPARQL itself doesn't "see" the additional inferred > triples. SHACL node expressions do, if you can reformulate this query, > using sh:path. > > In TopBraid there is a "native" way of querying inferred values within > SPARQL too: > > Try > > ( ?parent1 qudt:dappropriateUnit ) tosh:values ?unit . > > which is a magic property that will look at both inferred and asserted > values at query time. Make sure to avoid infinite loops though, i.e. no > cycles in the graph. > > As this would only work within TopBraid, a more "standard" approach would > be to use node expressions. Something like > > sh:values [ > sh:distinct [ > sh:if [ > sh:exists [ > sh:path [ sh:inversePath qudt:hasQuantityKind ] > > ] > ] ; > sh:then [ > sh:path [ sh:inversePath qudt:hasQuantityKind ] > ] > ] ; > sh:else [ > sh:path qudt:dappropriateUnit ; > sh:nodes [ > sh:path skos:broader ; > ] > ] > ] > ] > > which due to the geekiness and my lack of example data I haven't tested. > It's another syntax for "If this has a subject for hasQuantityKind then use > that, otherwise recurse into the broader parent". The sh:path near the > sh:else would do the recursion automatically, I believe. > > https://w3c.github.io/shacl/shacl-af/#node-expressions-path > > (I just noticed the example there was wrong but this will hopefully be > fixed soon). > > Holger > > > > Steve > > > -- > You received this message because you are subscribed to the Google Groups > "TopBraid Suite Users" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to [email protected]. > To view this discussion on the web visit > https://groups.google.com/d/msgid/topbraid-users/CAGUep86_5U4TAKX-Jjb5erVdzZZZ9Xrfvv5%3DOzNJi-69gCjdzg%40mail.gmail.com > <https://groups.google.com/d/msgid/topbraid-users/CAGUep86_5U4TAKX-Jjb5erVdzZZZ9Xrfvv5%3DOzNJi-69gCjdzg%40mail.gmail.com?utm_medium=email&utm_source=footer> > . > > -- > You received this message because you are subscribed to the Google Groups > "TopBraid Suite Users" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to [email protected]. > To view this discussion on the web visit > https://groups.google.com/d/msgid/topbraid-users/f0606b84-d191-ad63-014b-a1ca870980ce%40topquadrant.com > <https://groups.google.com/d/msgid/topbraid-users/f0606b84-d191-ad63-014b-a1ca870980ce%40topquadrant.com?utm_medium=email&utm_source=footer> > . > -- You received this message because you are subscribed to the Google Groups "TopBraid Suite Users" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/topbraid-users/CAGUep87iYAM%3DzFHyQXKicx54ZBGRLWHsUzQdp_SsQ2hkxHE%2Bfw%40mail.gmail.com.
