Modeling and retrieving linked RESTful APIs: A graph database approach

Abstract

This paper describes an approach that combines Linked Data and Graph Database concepts for modeling and retrieving RESTFul Linked APIs (Web APIs). We propose a multi-levels graph structure where the Web APIs (vertices) are connected by means of different types of links: The graph is stored in a graph database to allow graph exploration, e.g., to achieve an API discovery task for instance. The exploration/ retrieval process is performed by means of graph queries: on the user (client) side, a keyword search interface is provided while, on the server side (graph database), a graph query is issued. The approach has been implemented on top of Neo4j and its cypher query language, and an experimentation has been conducted with real datasets.