DBOWLizer

About

DBOWLizer is a novel framework for the automatic creation/extraction of expressive ontologies from (normalized) relational databases. This approach extracts and accurately represents the information found in a database by examining its composition. We anticipate that this work will help to expose the rich content currently stored in thousands of databases in a way that users, in particular scientists, can be more effective in retrieving information that supports their scientific endeavours.

Semantics Extraction

Extracts the semantics from a relational database and creates an expressive OWL ontology.

Mappings

Generates R2RML mappings between a relational database and the extracted ontology.

Rule-Based

Customizable heuristics to extract an ontology from a relational database.

Ontologies

The relational database schema composition is described in the relational-model ontology; the rules to analyze the schema composition and map it into an ontology term are defined as necessary sufficient conditions of classes and SWRL rules in the relational-to-ontology-mapping ontology. All DBOWLizer ontologies can also be downloaded at the cybershare ontology site.

Relational Model Ontologies

Ontologies that represent the relational model.

Ontology	Documentation	Source
Relational Model Primitive
Relational Model Complex
Relational Model Individuals

Mapping Ontologies

Ontologies that represent mappings between the relational model and ontologies.

Ontology	Documentation	Source
Relational Ontology Mapping Primitive
Relational Ontology Mapping Complex

DBOWLizer default mapping heuristics

Ontology that contains the default heuristics to map relational model components to OWL components used in the DBOWLizer framework.

Ontology	Documentation	Source
DBOWLizer Default Heuristics

System Architecture

DBOWLizer extends a set of external modules such as: SchemaCrawler for accessing and parsing the relational database schema, OWLAPI and R2RML-API to generate mappings and create the extracted ontology, and Jersey for the REST Web Service interface.

Uses View of External Elements

Testing

The DBOWLizer application was compiled using Java JDK 1.8 and tested against a sample database under MySQL 5.6. The application code is available at the Github repository for this project, the sample test can also be executed directly from the DBOWLizer endpoint service deployed into Tomcat 8.

Inputs

>> Sample Database: Employees MySQL Script
>> Sample Database: Schema Diagram

Outputs

Ontology	Documentation	Source
Relational Model Complex Output
After Reasoner Populated Ontology	Not Available
Employees Mapped by DBOWLizer Primitive
Employees Mapped by DBOWLizer Complex

>> R2ML Mappings: Employees R2ML Mappings

DBOWLizer Endpoint Service

The DBOWLizer application is exposed as a Rest Webservice. It receives a JSON specification of the database connection settings, and outputs a JSON consisting of download links to the generated ontologies and R2ML mapping files.

Endpoint Service Source Code

Resources

Source Code

>> DBOWLizer Releases: DBOWLizer Release v1.0 (Github Repository)

Linked Open Vocabularies (LOV)

Relational to Ontology Mapping Primitive

Our Team Members

Automated discovery, integration, analysis and interpretation of scientific data by providing context, or meaning to information using Semantic Web technologies.

Natalia Villanueva

Principal Investigator

Luis works as a Research Associate with full involvement during the software engineering lifecycle of scientific applications.

Luis Garnica Chavira

Implementation Lead

Diego is a PhD student whose research interest lie in the areas of Machine Learning, Recommender Systems, Natural Language Processing, and Semantic Web.

Diego Aguirre

Lead Developer

Erick is an undergraduate student that aims to graduate in summer 2016, interested in Semantic Web technologies and development of mobile apps.