ontop/ontop

What it is.

Ontop converts the data stored in ordinary relational databases into knowledge graphs — a way of organizing data as a web of connected facts, rather than tables. The key twist: the graphs are virtual. The data never leaves the original database; Ontop just makes it look like a graph to anyone or anything querying it.

How it works.

• You query a knowledge graph in SPARQL, a query language for graph data.
• Ontop translates that SPARQL query into SQL, the language relational databases understand.
• The translated SQL is executed directly on your existing database, and the results come back to you.
• To know how the relational rows should appear as graph facts, Ontop uses R2RML mappings — a standard way to describe that correspondence.
• Optionally, it can also use lightweight ontologies (formal descriptions of concepts and relationships) to improve the mapping.

The bottom line. You get the benefits of a graph view over your data without building or maintaining a separate graph database — the source tables are the only copy.

Project details.

• Build system: Built with Maven, requiring Maven 3.6+ and Java 11.
• Licenses: The software is Apache License 2.0; the documentation is Creative Commons (attribution).
• Status indicators: The README includes badges for Maven Central distribution, Docker images, SourceForge downloads, and a Twitter account.
• Supporters: The project is backed by the Free University of Bozen-Bolzano and the company Ontopic s.r.l., with regular contributions from Birkbeck, University of London and Southeast University.
• Find it online: Official site, SourceForge, Docker Hub, GitHub (issues and code), plus community channels (Google Group, Facebook, Twitter, Mastodon).

In brief: Ontop is a translator that lets you query your existing relational databases as if they were graph databases, without moving or duplicating any data.