A Dataflow Graph Transformation Language and Query Rewriting System for RDF Ontologies

Users interested in biological and biomedical information sets on the semantic web are frequently not computer scientists. These researchers often find it difficult to use declarative query and view definition languages to manipulate these RDF data sets. We define a language IML consisting of a small number of graph transformations that can be composed in a dataflow style to transform RDF ontologies. The language's operations closely map to the high-level manipulations users undertake when transforming ontologies using a visual editor. To reduce the potentially high cost of evaluating queries over these transformations on demand, we describe a query rewriting engine for evaluating queries on IML views. The rewriter leverages IML's dataflow style and optimizations to eliminate unnecessary transformations in answering a query over an IML view. We evaluate our rewriter's performance on queries over use case view definitions on one or more biomedical ontologies.