Writing Tests for Legacy Code is Slow – AI Can Help You Do It Faster

A new developer is tasked with enhancing a Flask-based application that serves ski resorts globally by calculating ski lift pass prices, a process complicated by an incomplete codebase left by a previous developer. The app currently handles single pass pricing via a REST API but needs updates to support multiple passes and specific discounts for young night skiers. However, the absence of automated tests complicates the task, as the developer is unsure of the existing scenarios and risks introducing bugs under tight deadlines. The text introduces qodo (formerly Codium), an AI-assisted tool integrating with IDEs like VS Code and JetBrains, which aids developers by quickly generating relevant tests and code analyses, thus facilitating a better understanding of unfamiliar code and mitigating the risk of regressions. Despite initial limitations such as improperly resolved mock paths, qodo significantly reduces the time needed to establish a safety net of tests, enabling developers to confidently implement new features without breaking existing functionalities.