GITS-Eval
收藏资源简介:
GITS-Eval数据集由谷歌研究团队创建,旨在评估基于代理的自动程序修复系统在企业环境中的表现。该数据集包含178个bug,其中78个由人工报告,100个由机器报告,涵盖了多种编程语言和项目。数据集的筛选过程经过多阶段过滤,确保每个bug都有可执行的测试和明确的修复补丁。该数据集主要用于研究自动程序修复系统在复杂企业环境中的性能,旨在解决企业代码库中的bug修复问题。
The GITS-Eval dataset was developed by the Google Research team to evaluate the performance of agent-based automated program repair systems in enterprise environments. This dataset comprises 178 bugs, 78 of which are manually reported and 100 of which are machine-reported, covering diverse programming languages and projects. The dataset's screening process employs multi-stage filtering to ensure that every bug has an executable test case and a definitive repair patch. This dataset is primarily utilized to research the performance of automated program repair systems in complex enterprise environments, with the goal of addressing bug repair challenges within enterprise codebases.




