Framework Based on Streams and Graphs for Model Extraction from Source Code

Abstract

Software visualization is an important issue in software maintenance domains, software architecture reconstruction and reverse engineering. In this context the use of some toolssupport is relevant for medium and large applications where the amount of source code to review and analyze is highest. This paper presents a framework based on streams and graphs for the extraction of models from code source. Some important features of this framework include stream parallelization, representation of elements based on graphs, definition of rules based on DSL and generation of UML diagrams in format XMI Java Web applications from GitHub repositories were used to evaluate the implementations of the framework; The results are shown.