Automated Test Data Generation for Typescript Web Applications

Abstract

This research proposes a Pattern-based Unit Testing method, namely PUT, to generate test data automatically for Typescript web applications. The main idea is to analyze the internal structure of source code to detect patterns of attribute usages. These patterns are then used to generate test data directly. An implemented tool and experimental results are presented to demonstrate the effectiveness of PUT in comparison with the automated random method. It shows that PUT could increase function coverage by 3.57 %-23.81 %, statement coverage by 0.5%-27.41 %, and branch coverage by 3.3%-39.24%. Therefore, PUT could provide potential usefulness for automated unit testing of Typescript web applications in practice.