LLM for Test Script Generation and Migration: Challenges, Capabilities, and Opportunities
Shengcheng Yu,Chunrong Fang,Yuchen Ling,Chentian Wu,Zhenyu Chen
2023-09-24
Abstract:This paper investigates the application of large language models (LLM) in the domain of mobile application test script generation. Test script generation is a vital component of software testing, enabling efficient and reliable automation of repetitive test tasks. However, existing generation approaches often encounter limitations, such as difficulties in accurately capturing and reproducing test scripts across diverse devices, platforms, and applications. These challenges arise due to differences in screen sizes, input modalities, platform behaviors, API inconsistencies, and application architectures. Overcoming these limitations is crucial for achieving robust and comprehensive test automation.
By leveraging the capabilities of LLMs, we aim to address these challenges and explore its potential as a versatile tool for test automation. We investigate how well LLMs can adapt to diverse devices and systems while accurately capturing and generating test scripts. Additionally, we evaluate its cross-platform generation capabilities by assessing its ability to handle operating system variations and platform-specific behaviors. Furthermore, we explore the application of LLMs in cross-app migration, where it generates test scripts across different applications and software environments based on existing scripts.
Throughout the investigation, we analyze its adaptability to various user interfaces, app architectures, and interaction patterns, ensuring accurate script generation and compatibility. The findings of this research contribute to the understanding of LLMs' capabilities in test automation. Ultimately, this research aims to enhance software testing practices, empowering app developers to achieve higher levels of software quality and development efficiency.
Software Engineering