Design of Cross-Repository API Documentation Automation for Projects Governed by Version Control

Abstract

Proper API documentation is vitally important for developing with the API. It should be provided in a clear and understandable format, with heightened accessibility and accuracy. This thesis is an attempt to ease the production of API documentation, by automating the process of its creation in a multi- repository project structure. The thesis will also survey the current state of generative AI usage within this space, and acquire knowledge on how following improvements for the implementation of this design can utilize AI. This, and finding a way to organize this automation system, such that it could scale for the inclusion of other projects, are the research objectives of this thesis. Creating the design and its implementation follow Design Science Research, while the literature review follows an exclusion method to find relevant articles for research within context. The design can be presumed to scale well with the anticipation of new projects, although a definitive and objective answer was not possible to be deduced from the limited data gathered. There were a few possibilities on how to include AI within this design, but none are mature enough yet to be used without human intervention. More research on the AI side is still necessary. The automation system is in production, and more data will be gathered from it in-house.