Skip to content

Latest commit

 

History

History

shape-validation

Folders and files

NameName
Last commit message
Last commit date

parent directory

..
 
 
 
 
 
 

RDF ShapeValidation with Apache Jena

Here I show how easly simple RDFData validations can be done with SHACL

Project Setup

We need two dependencies to archive this, the Apache Jena Core and the SHACL extension

<dependency>
    <groupId>org.apache.jena</groupId>
    <artifactId>jena-core</artifactId>
</dependency>
<dependency>
    <groupId>org.apache.jena</groupId>
    <artifactId>jena-shacl</artifactId>
</dependency>

Test Data Instances

We have Bob who is perfectly fine

ex:Bob
    rdf:type     ex:Person ;
    ex:firstName "Bob" ;
    ex:lastName  "Bobson" ;
    ex:nickName  "the guy" ;
    ex:age       "100"^^xsd:integer .

And then we have Sue who has some violations

ex:Sue
    rdf:type     ex:Person ;
    ex:firstName "Sue" ;
    ex:firstName "Süe" ;
    ex:nickName  "first nicke" ;
    ex:nickName  "second nick" ;
    ex:age       "99" ;
    ex:undifined "undefined" .

The data can be seen here and the RDFSchema is given here

Test Data Shape

We define a simple Person dataset with the very restrictive shape:

  • A Person must have one firstName and lastName property, with type String.

  • A Person can have 0..n nickName properties of type String.

  • A Person must have am age of type Integer

  • A Person should not have any other property.

To archive that we define the SHACL property sh:minCount and sh:maxCount together with sh:datatype. The shapes can be seen here

Validation with Apache Jena

Sorurces see ValidateRDFWithSHACL.java.

First of all we need to load the data.ttl and shapes.ttl into a Graph, as seen in lines 21 - 27

Then we parse the shapes graph into a Shapes object instance, see line 29.

We validate the the whole graph by passing the shapes and data graph to a ShaclValidator instance, see line 40.

We can even validate single nodes as done in line 46.

The generated report is again RDF data, it is an empty graph if the data is conform or contains a validations report. See method in line 49ff