Implementing Graph Transformation Languages using RDF Storage and SPARQL Queries

The Resource Description Framework (RDF) format has seen signficant interest from industry as a flexible format for graph representation. However, RDF has so far not been widely used in the context of graph rewriting tools. In this paper, we take AToMPM, our research-oriented meta-modelling and graph rewriting tool and extend it by adding RDF-based graph transformation functionality. We address the graph matching problem, generating custom SPARQL queries for each rule pattern, and use that to efficiently perform matches on a source RDF graph. In this way we create a research system for RDF graph transformations that also gives us the ability to investigate different possible approaches to using RDF and evaluate the effects on the system. Our experience and performance data serves as a guideline for evaluating the trade-off between the inherent efficiency of a specialized representation and the flexibility of using a standardized representation.