Implementing GraphQL as a Query Language for Deductive Databases in SWI-Prolog Using DCGs, Quasi Quotations, and Dicts

TL;DR

This paper introduces a Prolog library that implements GraphQL as a query language for deductive databases, enabling flexible data access and validation using SWI-Prolog's language features.

Contribution

It presents the first implementation of GraphQL in SWI-Prolog for deductive databases, combining DCGs, quasi quotations, and dicts for type-safe querying.

Findings

Successfully implemented GraphQL type system in SWI-Prolog

Provides a unified interface for data access and validation

Enables flexible querying over deductive databases

Abstract

The methods to access large relational databases in a distributed system are well established: the relational query language SQL often serves as a language for data access and manipulation, and in addition public interfaces are exposed using communication protocols like REST. Similarly to REST, GraphQL is the query protocol of an application layer developed by Facebook. It provides a unified interface between the client and the server for data fetching and manipulation. Using GraphQL's type system, it is possible to specify data handling of various sources and to combine, e.g., relational with NoSQL databases. In contrast to REST, GraphQL provides a single API endpoint and supports flexible queries over linked data. GraphQL can also be used as an interface for deductive databases. In this paper, we give an introduction of GraphQL and a comparison to REST. Using language features recently added to SWI-Prolog 7, we have developed the Prolog library GraphQL.pl, which implements the GraphQL type system and query syntax as a domain-specific language with the help of definite clause grammars (DCG), quasi quotations, and dicts. Using our library, the type system created for a deductive database can be validated, while the query system provides a unified interface for data access and introspection.