Integration of GraphQL data sources into an RDF federation engine

Sammanfattning

As data interoperability has become increasingly more important as of late, so has the topic of database query language translations. Being able to translate between different query languages, both old and new, could potentially improve a lot of different systems. One of these systems is HeFQUIN, a federated query engine that primarily focuses on querying many different RDF data sources by sending out SPARQL sub-queries. Exploring ways to extend this engine to include non-RDF data sources, like GraphQL for example, would naturally increase the uses of such an engine. This thesis presents, and implements, a generalized approach for translating SPARQL queries, primarily Basic Graph Patterns (BGPs), into GraphQL queries, after which the resulting JSON response received by executing the query is translated back to solution mappings. The performance of the translator implementation is then evaluated by using a set of pre-determined, specially designed, test queries. These queries are structured to encompass many different scenarios the proposed translator could face. The results of the evaluation tests reveal that the translator implementation with its steps only constitute a comparatively small fraction of the full request to the GraphQL endpoints. Finally, it is concluded that due in part to the generalized nature of the approach, a wide range of different structured basic graph patterns are able to be translated, as long as they follow the set up requirements imposed on the BGP and the GraphQL endpoint.