OntoMatch

Introduction

This work won the award for the best project at LDAC2019 Hackathon . We extended the work by Faria et al. and modified their tool AgreementMakerLight for this project.

Our implementation can be used to exploring the capabilities of AgreementMakerLight tool for ontology matching. We used Python scripts to automate creating different outputs by trying out different settings of AgreementMakerLight to determine the best property combination for matching your domain ontology.

Requirements and Steps

Python libraries required

1. Python 3.7
2. RDFlib
3. Numpy
4. Itertools
5. re
6. os

Precompiled tools required

AgreementMakerLight

Steps

1. Clone this repo.
2. Download and unzip AgreementMakerLight inside the repo such that AgreementMakerLight.jar and OntoMatch.py and Folder 'store' are in the root folder.
3. Place the ontologies to be compared in your root folder.
4. Open OntoMatch.py using VS code or similar.
5. Edit the combination of properties in the main() in the file OntoMatch.py.
6. Run the code.
7. You would see results in count_result and result file.
8. The respective config.ini files containing combinations of AML methods and the resulting 'alignment.rdf' are stored in ./store/config_files/ and ./store/alignment_files/ respectively.

Using Knowledge Bases to improve matching

The AgreementMakerLight has options to different matching steps, the default one being "Background Knowledge Matcher". The .py file in this repository runs using all the available owl files in the ./store/knowledge/. folder by default.

1. An owl file of any ontology has to be copied in the following root folder of the tool: ./store/knowledge/.
2. In the tool, go to Match>Custom Match>Match Options>Settings>Matching Settings, and select your loaded knowledge base prior to initiating the matching.