Mono2Micro CodeQL Collector

Miguel do Carmo Ferreira

Key information

Authors:

Miguel do Carmo Ferreira (Miguel Do Carmo Ferreira)

Supervisors:

António Manuel Ferreira Rito da Silva (António Manuel Ferreira Rito da Silva)

Published in

October 31, 2025

Abstract

The Mono2Micro project enables the decomposition of monolithic applications into microservice architectures, relying on representation files to perform this process. These files are generated by code collectors that can only analyze Java applications built with Spring Data JPA or the Fénix Framework, since they rely on the Spoon library, which exclusively processes Java source code. This approach limits the applicability and versatility of Mono2Micro. This thesis proposes the development of a new code collection tool to address the limitations of the current solution. By integrating an existing code analysis tool, the new code collector aims to support codebases built with frameworks across different programming languages, provide easy extensibility, and produce accurate and reliable representation files. Existing code analysis tools face two main limitations: many support only a few programming languages, reinforcing the problem of limited language coverage. And even tools that claim multi-language or language-agnostic support often lack the granularity needed to capture framework-specific code elements. The new collector is implemented in Java, integrates CodeQL, and adopts a modular architecture for easy extension to new programming languages and frameworks. The tool will be evaluated using a variety of open-source projects spanning different programming languages. Evaluation metrics will include language and framework coverage, ease of extension and performance. The new collector tool is expected to expand the scope of the Mono2Micro project, offering support to a broader range of programming languages while also sustaining an extensible and modular design.