can one sparql constraint generate multiple violations in pySHACL?

[habdirad]

Hello all,
I am experimenting with pySHACL and comparing validation outputs against the topquadrant's validator.

Issue/Question:
I have some sparql constraints that check a parent node against the child nodes in a graph. I noticed that pySHACL returns only one of the offending childs in the violation report. Can one sparql constraint generate multiple violations in pySHACL?

Reproducible example:
the following example should generate 3 violations (one for each person) but only returns one (only for ex:John).
graph:

@prefix ex: <http://example.org/ns#> .
@prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#> .
@prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> .
@prefix schema: <http://schema.org/> .
@prefix xsd: <http://www.w3.org/2001/XMLSchema#> .

ex:Bob
    a schema:Person ;
    schema:givenName "Robert" ;
    schema:familyName "Junior" ;
	schema:family ex:JuniorFamily .

ex:Molly
    a schema:Person ;
    schema:givenName "Molly" ;
    schema:familyName "Junior2" ;
	schema:family ex:JuniorFamily .
	
ex:John
    a schema:Person ;
    schema:givenName "John" ;
    schema:familyName "Junior3" ;
	schema:family ex:JuniorFamily .

	
ex:JuniorFamily
	a schema:Family;
	schema:CorrectfamilyName "Junior" .

shapes:

@prefix dash: <http://datashapes.org/dash#> .
@prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#> .
@prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> .
@prefix schema: <http://schema.org/> .
@prefix sh: <http://www.w3.org/ns/shacl#> .
@prefix xsd: <http://www.w3.org/2001/XMLSchema#> .
@prefix ex: <http://example.org/ns#> .

schema:PersonShape
    a sh:NodeShape ;
    sh:targetClass schema:Family;
sh:property schema:family-familyname.

schema:family-familyname
        rdf:type     sh:PropertyShape ;
        sh:path      schema:CorrectfamilyName ;
        sh:name      "familyname" ;
        sh:sparql    schema:onefamilyname ;
        sh:nodeKind  sh:Literal .

schema:onefamilyname 
        rdfs:label  "onefamilyname " ;
        rdf:type    sh:SPARQLConstraint ;
        sh:message  "{?this} {?person}" ;
        sh:select   "prefix dash: <http://datashapes.org/dash#>\r\n prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#>\r\n prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#>\r\n prefix schema: <http://schema.org/>\r\n prefix sh: <http://www.w3.org/ns/shacl#>\r\n prefix xsd: <http://www.w3.org/2001/XMLSchema#>\r\n prefix ex: <http://example.org/ns#>\r\n\r\n\r\nSELECT ?this ?person\r\nWHERE {\r\n  ?this a schema:Family.\r\n ?person schema:family ?this.\r\n}" .

script:

conforms, results_graph, results_text = validate(
    graph,
    shacl_graph=shapes,

    data_graph_format="turtle",
    shacl_graph_format="turtle",
    inference=None,
    advanced = True,
    debug=True,
)

output showing 1 violation from pySHACL:
'Validation Report\nConforms: False\nResults (1):\nConstraint Violation in SPARQLConstraintComponent (http://www.w3.org/ns/shacl#SPARQLConstraintComponent):\n\tSeverity: sh:Violation\n\tSource Shape: schema1:family-familyname\n\tFocus Node: ex:JuniorFamily\n\tResult Path: schema1:CorrectfamilyName\n\tSource Constraint: schema1:onefamilyname\n\tMessage: http://example.org/ns#JuniorFamily http://example.org/ns#John\n'

output from Topquadrant's validator showing 3 violations:

[ashleysommer]

Hi @habdirad
Thanks for this report.
Yes, you're right, indeed a single sparql constraint can/should be able to return multiple violations.
I thought there were tests to ensure this behaviour in the SHACL test suite, but perhaps there are not.
I think this may be caused by a bug due to the use of SPARQLConstraint on a sh:PropertyShape (which is valid, but not well tested). I'll look into it.

[ashleysommer]

Hi @habdirad
I have finally got a chance to investigate your described issue.
I tested your reproduction shapes and datagraph as you've given it, and I am seeing the 3 errors reported same as you show from TopQuadrant validator.

Validation Report
Conforms: False
Results (3):
Constraint Violation in SPARQLConstraintComponent (http://www.w3.org/ns/shacl#SPARQLConstraintComponent):
	Severity: sh:Violation
	Source Shape: schema1:family-familyname
	Focus Node: ex:JuniorFamily
	Result Path: schema1:CorrectfamilyName
	Source Constraint: schema1:onefamilyname
	Message: http://example.org/ns#JuniorFamily http://example.org/ns#Bob
Constraint Violation in SPARQLConstraintComponent (http://www.w3.org/ns/shacl#SPARQLConstraintComponent):
	Severity: sh:Violation
	Source Shape: schema1:family-familyname
	Focus Node: ex:JuniorFamily
	Result Path: schema1:CorrectfamilyName
	Source Constraint: schema1:onefamilyname
	Message: http://example.org/ns#JuniorFamily http://example.org/ns#John
Constraint Violation in SPARQLConstraintComponent (http://www.w3.org/ns/shacl#SPARQLConstraintComponent):
	Severity: sh:Violation
	Source Shape: schema1:family-familyname
	Focus Node: ex:JuniorFamily
	Result Path: schema1:CorrectfamilyName
	Source Constraint: schema1:onefamilyname
	Message: http://example.org/ns#JuniorFamily http://example.org/ns#Molly

See my gist for exactly the PySHACL code I am running:
https://gist.github.com/ashleysommer/62704fd8cba732ffcabf5f07de8a74a9

[habdirad]

Thank you @ashleysommer. I was using version 0.23. 0 inadvertently. Updating to 0.26.0 fixed this issue. Apologies for the false alarm.