Learning-based Testing of aLarge Scale Django Web Application : An Exploratory Case Study Using LBTest

Abstract: Learning-based testing is an emerging software testing paradigm. With it, automatic test case generation is achieved by combining model-based testing with machine learning. Recent research has led to the development of LBTest, a tool that is getting ready for use in the software industry. The tool has mostly been used in small academic case studies and only two previous industrial case studies. This thesis investigated the use of LBTest for testing a large scale web application from the financial IT sector, built using the popular Django framework. When errors were injected into the system under test, LBTest successfully generated test cases showing how they were violating given software requirements. Efficiency problems were discovered and solved in a general way, repeatable for testing any Django system with LBTest.