On the Semantic Web, vocabularies and ontologies play a fundamental role to express the terminology and rules of certain domains. New technologies like SHACL provide the possibility to express data schemata specific to certain data sets, applications, and domains. However, the domain modeling process is collaborative and when using RDF, it requires technical knowledge. In this paper, we present a tool to support a two-step-process to model a terminology and a schema with a combined graphical RDF Schema editor and visual SHACL editor. This tool allows domain experts to create a terminology and schema without the need for a deep understanding of RDF Schema or SHACL.
experts into the process. Several systems to deal with SHACL shapes have been proposed
(cf. section 2). So far, some of these tools provide a visualization of SHACL shapes and only
one system has a prototypical interactive visual editor for SHACL shapes. Consequently, our
far goal is to establish a joined distributed collaborative domain modeling process, based on the
Quit Store [
This paper is structured as followsS.ection 2 provides an overview on the literaturSee;ction 3shows some implementation details and demonstrates the tool; finallSye,ction 4presents the conclusions and a prospect to future work.
There are many tools that contribute to creating a visual vocabulary and ontology editors. For the sake of brevity of this demo paper we have created a comparison in the Open Research Knowledge Graph (ORKG) 3[] that represents the related work2][. It is available athttps://doi .org/10.48366/R113089.
The OntoPad is implemented as a web application using JavaScript and the VueJS framework. The source code is available as FLOSS licensed under the terms of the GPL-3.0 on GitHub (https://github.com/AKSW/OntoPad). This allows to use state-of-the-art technology for the user interfaces and provide a responsive interaction independent of the users client platform. Figure 1shows an overview of the domain modeling interface of the OntoPad.
To store the data, we send SPARQL Update requests to a SPARQL endpoint. In order to support the highly collaborative process of domain modeling, we use the Quit Sto1]reas[ SPARQL endpoint. The Quit Store tracks all changes performed with SPARQL Update requests in a Git repository and allows to synchronize the repository with remote collaborators. This allows each collaborator to setup their own independent workspace while still participating in a common distributed collaborative workspace.
In the first step, the terminology of a domain is defined. For this purpose the OntoPad provides a terminology component as can be seen in the left column ofigf. 1 . In the top it lists all explicit instances of rdfs:Class, as well as inferred instances orfdfs:Classby considering objects ofrdf :type-triples and instances ofowl:Class, that are found in the current graph. By clicking on the plus-button a new class term can be added. The interface to add a new class is shownfigi.n2 . The user can specify the IRI as well as a label using the properrtydfs:labeland a comment using rdfs:comment. To edit a class, the user can select the respective term and switch to the edit tab. The edit tab provides a triple view on the class that allows to change the existing statements and also to add new statements. Additionally a source-tab is available to edit the resource using the Turtle syntax. In the lower part of the terminology component, properties can be defined and edited as instances of rdf:Propertyin the same way as for class terms. The term creation process is kept at a bare minimum to provide a very simple user interface that does not distract the domain experts from the discursive process.
In the right part of the domain modeling interface, as depicted figin. 1 , the visual diagram component is provided. The graphical notation of the diagram is inspired by the notation of UML class diagrams as used in software engineering. The modeling tool allows the user to create new shapes by dragging anrdfs:Class from the left hand side and dropping it in the right hand shape composer. This action automatically creates a new instancesohf:NodeShape with the class specified as sh:targetClass. In the same way the user can drag anrdf:Property from the left hand list of terms and drop it inside a node shape. This automatically creates a new instance of sh:PropertyShapewith the property specified as sh:path. The sh:PropertyShapeis attached to the sh:NodeShape with the sh:propertyproperty. To specify a shape as constraint for the value nodes of a property the user can click the square behind the property and draw a line to theIRI input port of the respective node shape. Also thseh:PropertyShapeand sh:Node Shape resources can be viewed and edited with the triple editor or with the source-tab. By clicking the save button in the top right corner, the created node shapes and property shapes are stored to the underlying Quit Store. When opening the view, the diagram is initialized with already existing instances ofsh:NodeShape and their linkedsh:PropertyShapeinstances.
We presented to our knowledge the first system that allows the user to create and edit an RDF terminology in conjunction with a visual diagram based editor for SHACL shapes. The OntoPad supports the two step vocabulary creation process: it allows to create and edit an RDF terminology in a graphical user interface and it provides an editor based on a visual graph diagram of SHACL Shapes. Our system is complemented by an editor for the RDF serialization and a graphical user interface to edit triples. OntoPad is built on top of Quit Store which provides a versioned RDF triple store, allowing the user to keep track of the creation process, provides the possibility for distributed collaboration, and tracks the relevant provenance information.
To actually make the vision of an interlinked vocabulary space in the Semantic Web true, we
want to extend the system to also import existing vocabularies from vocabulary repositories
like Linked Open Vocabularies9[] or DBpedia Archivo5[]. This should allow to foster the reuse
of existing terminology and provide a mix-and-match environment to construct SHACL shapes
for specific use cases and applications. As future work, we will further implement more of the
SHACL constraints, like cardinality, datatypes, and string patterns, into the editor to support
use cases like the creation of input forms with RDFormht(tps://github.com/simeonackermann
/RDForm/). Even though the underlying Quit Store provides many unparalleled features, we
want to untangle the dependency on this specific triple store to be able to use the OntoPad in
combination with any endpoint that provides the standard SPARQL 1.1 Query and Update
interface. To support the full linked data life-cycle the system shall also be tested in combination
with and integrated with systems like ASTREA4[] in order to enable the automatic generation
of SHACL shapes and test an ontology using formally pre-defined guidelines or custom SHACL
tests with SHARK [
This work was partly supported by the Bundesanstalt für Materialforschung und -prüfung (BAM) and by a grant from the German Federal Ministry for Economic Afairs and Energy (BMWi) for the PlatonaM project (01MT19005A).