Automated Generation of REST API Specification from Plain HTML Documentation

REST is nowadays highly popular and widely adopted by Web services providers. However, most of the Web services providers only provide the documentation of their REST API in plain HTML pages, even if many specification formats exist such as WADL or OpenAPI for example. This prevents the Web Services users to benefit from all the advantages of having a machine-readable specification, such as generating client or server code, generating web services composition, checking formal properties, testing, etc. To face this issue, we provide a fully automated approach that builds a REST API specification from its corresponding plain HTML documentation. By given the root URL of the plain HTML API documentation, our approach automatically extracts the four mandatory parts that compose a specification: the base URL, the path templates, the HTTP verbs and the associated formal parameters. Our approach has been validated with topmost commercial REST based Web Services, and the validation shows that our approach achieves good precision and recall for popular Web Services.