The OpenAPI specification provides a language-agnostic way to describe REST APIs, which serves as documentation and enables automated generation of client libraries and server stubs.
Large language models can extend OpenAPI's code generation capabilities beyond simple stubs to create functional backend code, particularly for database-centric operations.
The code generation process targets a specific Spring framework architecture, breaking the problem down into generating controllers, repositories, entities, and schema classes.
A structured four-part prompt including task, rules, input, and context is used to reliably generate schema classes and database entities from the OpenAPI specification.
By providing the LLM with the operation specification and previously generated classes as context, it can generate complete controller endpoints and database repositories.
While the generated code is reliable for database-centric tasks, limitations include placing logic in controllers, lacking authorization, and the inherent incompleteness of the OpenAPI spec.
31:12Daniel Töws
32:28Hugo Guerrero
32:26Krzysztof Cieślak
31:39Rainer Stropek
32:14Timo Salm & Sandra Ahlgrimm
31:53Timo Salm
28:22Patrick Schnell
26:30Simon A.T. Jiménez
Jobs that call for the skills explored in this talk.
CONTIAMO GMBH
Berlin, Germany
webLyzard
Vienna, Austria
Amdocs
Kontich, Belgium