Evaluation of Static Analysis Methods of Python']Python Programs

Static analysis is a method for detecting code smells and possible software bugs by examining the source code without executing the program. While we have considerable experiences for programming languages with static type system, especially for C, C++, and Java, languages with dynamic behavior requires different approaches. Python is an important programming language with a dynamic type system, used in many emerging areas, including data science, machine learning, and web applications. In this work we overview static analysis methods currently applied for Python, investigate their advantages and shortages, and highlight the restrictions of current tools and suggest further research directions to tackle these problems. We report our experiences applying static analysis methods on an open source Python software system where we found numerous issues confirmed by the developers. Based on these findings, we suggest refined configuration settings on static analysis tools.