SUMA is an efficient and scalable query answering system. It develops a partial materialization-based approach to extending RDF data. SUMA supports the OWL 2 DL ontology.

This documentation is for SUMA 1.0.0.

• Produces efficient materialization by low complexity materialization algorithm
• Builds three types of indexes for data and rules
• Provides reasoning API
• Can integrate off-the-shelf efficient SPARQL query engines
• Supports OWL 2 DL ontologies
• Supports SPARQL queries

• There are two ways to use SUMA: first, run the SUMA source code, and second, run SUMA.jar by command line.

• For instance, in IntelliJ IDEA, you can load in the following way: Project Structure -> Libraries -> new project library -> Java;

• pathTBox: the ontology path (*.owl)

• pathABox: the RDF data path (*.nt/ttl)

• n: the step of materialization

• pathExtendedABox: the path of extended RDF data (*.nt)

• isQueryByJena: whether to run the Jena query, which defaults to true

• initIsRoleWriting(true): whether to apply role rewriting algorithm, which defaults to true

• queryPath: the SPARQL query path (.sparql)

• answerPath: the answer path

• After you complete the first three steps, you can run the main class testSUMARunTest.java directly. The default path in the main class is our running example. It points to src/main/java/com/tju/suma/data, where includes standard.sparql (query file), univ-bench-dl.owl (ontology file), and uobm1.nt (RDF file).

• There are two ways to run SUMA.jar: first, directly run SUMA.jar by command line, and second, regenerate the jar from the SUMA source.

• java -jar SUMA.jar ONTOLOGY_PATH DATA_PATH NEW_DATA_PATH
• For example: java -jar SUMA.jar uobm.owl uobm1.nt uobm1_new.nt

• LUBM data generator;
• UOBM data generator;
• Ontologies used at the experiment of SUMA at DASFAA'20;
• Queries used at the experiment of SUMA at DASFAA'20;
• DBPedia+ ontology, data, and query.

• PAGOdA: PAGOdA exploits a hybrid approach to answering conjunctive queries over OWL 2 ontologies that combines a datalog reasoner with a fully-fledged OWL 2 reasoner in order to provide scalable "pay as you go" performance.
• RDF3x: RDF3x is an efficient SPARQL query engine.