一句话：提出端到端可扩展框架 PySTAAR，集成测试生成、故障定位、补丁合成与验证，借助大模型自动修复 Python 类型错误。
本文介绍 PySTAAR，一个面向 Python 程序的端到端、可扩展的类型错误自动检测与修复框架。现有 Python 类型错误修复工作通常只覆盖流水线中的单个环节，难以在真实应用上全自动完成从发现问题到确认补丁的完整流程。PySTAAR 将测试生成、故障定位、补丁合成与补丁验证整合为一条全自动流水线，并结合大语言模型与当前最先进的修复技术。框架采用模块化、可扩展的架构，并提供可视化整个工作流的易用 Web 界面，使不同经验水平的开发者都能使用。作者在真实 Python 应用上进行了工具演示，表明 PySTAAR 无需人工干预即可自动检测并修复类型错误，并取得较高的修复率。
We present PySTAAR, an end-to-end, extensible framework for automatically detecting and repairing type errors in Python programs. PySTAAR integrates test generation, fault localization, patch synthesis, and patch validation into a fully automated pipeline, leveraging large language models and state-of-the-art repair techniques. The framework features a modular, extensible architecture and a user-friendly web interface that visualizes the workflow, making it accessible to developers of all experience levels. We demonstrate that PySTAAR can automatically detect and repair type errors in real-world Python applications, achieving high repair rate without manual intervention. Our tool demonstration is available at https://youtu.be/VizRQsrtsdk.
@inproceedings{paperbot11542,
title = {PySTAAR: An End-to-End, Extensible Framework for Automated Python Type Error Repair},
author = {Wonseok Oh and Hyobin Park and Miryeong Kang and Seungbin Choi and Yunja Choi and Hakjoo Oh},
booktitle = {ICSE},
year = {2026},
doi = {10.1145/3774748.3787613}
}