Wolpertinger

Building

In order to build Wolpertinger (Standalone and Protege-plugin), ensure that you have Git and Maven installed.

Clone a copy of the repo:

git clone https://github.com/wolpertinger-reasoner/Wolpertinger.git

Change to the directory:

cd Wolpertinger

Run maven to build the artifacts:

mvn clean package

Installing

For the plug-in, it remains to copy the JAR-file from the target directory into your plugins directory of the local Protege installation. You need to have Clingo (https://github.com/potassco/clingo) in your machine for doing reasoning tasks.

Running

3-Coloring Examples

We use an ontology that encodes the 3-Coloring problem to show how to retreive models. The ontology is available in the sub-repository Ontologies\3Coloring. The core axioms of the problem encoding can be found in 3Coloring.owl. Assume,we want to find cloring of the graph G = (V, E):

• introduce a new individual "v_i" for each vertex v in V as an instance of concept Node.
• for each edge "(v_i, v_j)" in E, insert an object property assertions "edge(v_i, v_j)"

We provide an instance example using a graph with 4 vertices and 4 edges in 3Coloring-4Nodes.owl. We can then ask Wolpertinger for 3 fixed-domain models of the ontology using the command:

java -jar [wolpertinger-path] --models=3 --abox=3coloring-models [ontology-path]

* the ontology-path is an URI. for example, in windows machines:

java -jar C:\wolpertinger.jar --models=3 --abox=3coloring-models "file:/C:/Ontologies/3Coloring/3Coloring-4Nodes.owl"

The ontologies that represent models will be written in the folder 3coloring-models. One can derive the solution from each model by looking at the instances of RedNode, BlueNode, and GreenNode. For example, one of the models is:

node(i_n4) node(i_n2) node(i_n3) node(i_n1) edge(i_n1,i_n2) edge(i_n1,i_n3) edge(i_n2,i_n3) edge(i_n2,i_n4) edge(i_n4,i_n2) edge(i_n3,i_n2) edge(i_n3,i_n1) edge(i_n2,i_n1) rednode(i_n2) greennode(i_n1) bluenode(i_n4) bluenode(i_n3)

Justifications

Wolpertinger provides a translation that allows to compute justifications for inconsistency using asprin. The parameter "--justification" therefore translates an input ontology into an answer-set program, such that its answer-sets correspond to justifications. The resulting program is dedicated to be processed with asprin (which in turn uses Clingo). In the current development state, asprin needs to be called seperately.