May 10, 2023
Tooling to debug Python code in VS Code
Debugging Python code, or any code, is often a challenging and time-consuming task, especially when dealing with large and complex code bases. Traditional Python debugging and Python code review tools are able to detect simple problems, usually related to style, syntax, as well as commonly listed security vulnerabilities. These tools can often be integrated into an IDE and certainly provide value in the development process. At the same time, they lack the ability to detect complex problems that require a contextual understanding of the code base.
Metabob is an AI-powered tool used for refactoring and debugging Python code. Metabob integrates a graph-attention-based problem detection model with generative AI models that the tool uses to create descriptions and code recommendations to fix the problems it detects in Python code. Unlike traditional, rules-based tools used for Python code reviews, Metabob is a powerful tool for debugging Python code due to its capability to detect complex problems that require a contextual understanding of the code base. Specifically, the tool looks to accelerate the process of debugging Python code by identifying problems such as memory leaks, race conditions, unhandled edge cases, and problems from hundreds of other categories. The Python debugging tool is currently available as a free VScode extension.
By utilizing the conversational capabilities of the integrated generative AI models, Metabob offers its users features that haven’t been introduced in previous Python code review extensions. These features include the capability to ask questions about problem descriptions and to pass more code context for problem descriptions. At the same time, Metabob also uses generative AI models to create context-sensitive code recommendations to rapidly fix the detected problems. Users can also pass in more context to the integrated generative AI models to further improve the results. In addition, users have the option to integrate different generative AI models such as GPT.
Overall, debugging Python code will become easier and more efficient through the latest advancements in machine learning and artificial intelligence.