OWLAPI

mOWL interfaces the OWLAPI by using JPype to connect to the Java Virtual Machine. JPype allows to access Java objects and methods from a Python script. For example, we can use the OWLAPI from Python:

import mowl
mowl.init_jvm("10g")
from org.semanticweb.owlapi.apibinding import OWLManager

manager = OWLManager.createOWLOntologyManager()
ontology = manager.createOntology()

ontology.getClassesInSignature()

Note

Notice that accessing to Java objects is only possible after starting the JVM

Shortcuts to the OWLAPI

In order to ease the ontology management, mOWL provides an adapter class that wraps some objects found in the OWLAPI such as the OWLManager and the DataFactory:

from mowl.owlapi import OWLAPIAdapter

adapter = OWLAPIAdapter()
owl_manager = adapter.owl_manager # Instance of org.semanticweb.owlapi.apibinding.OWLManager
data_factory = adapter.data_factory # Equivalent to owl_manager.getOWLDataFactory()

Shortcuts to OWL Reasoners

mOWL provides some shortcuts for performing reasoning over ontologies. This works by wrapping OWL Reasoners that are instances of org.semanticweb.owlapi.reasoner.OWLReasoner. The following example shows how to obtain inferences from an ontology.

from mowl.datasets.builtin import FamilyDataset
from mowl.reasoning.base import MOWLReasoner
from org.semanticweb.HermiT import Reasoner

dataset = FamilyDataset()

reasoner = Reasoner.ReasonerFactory().createReasoner(dataset.ontology)
reasoner.precomputeInferences()

mowl_reasoner = MOWLReasoner(reasoner)
classes_to_infer_over = list(dataset.ontology.getClassesInSignature())

subclass_axioms = mowl_reasoner.infer_subclass_axioms(classes_to_infer_over)
equivalence_axioms = mowl_reasoner.infer_equivalent_class_axioms(classes_to_infer_over)
disjointness_axioms = mowl_reasoner.infer_disjoint_class_axioms(classes_to_infer_over)

The previous example uses the HermiT reasoner. We can also use ELK:

from org.semanticweb.elk.owlapi import ElkReasonerFactory
reasoner_factory = ElkReasonerFactory()
reasoner = reasoner_factory.createReasoner(dataset.ontology)

mowl_reasoner = MOWLReasoner(reasoner)