Integrating A Large Language Model Into Search-based Automated Program Repair: Experiment Results

The results of the experiments presented in the paper "Integrating A Large Language Model Into Search-based Automated Program Repair" at SANER 2026. geneseer-saner-2026.zip contains the implementation of the approaches, geneseer-1.0.0-SNAPSHOT-jar-with-dependencies.jar contains the built version of the tool (executable JAR). The three JSON files contain the raw experiment output. The format is a list of objects for each bug-repair attempt. Each such execution has the following fields: id executionIndex commandLine (list of the command-line to launch the repair tool) exitStatus (of the repair tool process) result (the result object, as documented in the README of the geneseer tool) logOutput started (timestamp) ended (timestamp) data.ods contains and aggregated overview of the repair results, as well as the diagrams presented in the paper.本数据集为发表于SANER 2026的论文《Integrating A Large Language Model Into Search-based Automated Program Repair》（《将大语言模型（Large Language Model）集成于基于搜索的自动化程序修复》）中的实验结果。 geneseer-saner-2026.zip 包含了对应方法的实现代码；geneseer-1.0.0-SNAPSHOT-jar-with-dependencies.jar 为该工具的可执行编译版本（含所有依赖的JAR包）。 三个JSON文件存储了原始实验输出，其格式为每一次程序修复尝试对应的对象列表，单次执行包含以下字段：id、executionIndex、commandLine（启动修复工具的命令行列表）、exitStatus（修复工具进程的退出状态）、result（结果对象，详见geneseer工具的README文档）、logOutput（日志输出）、started（启动时间戳）、ended（结束时间戳）。 data.ods 文件包含了修复结果的汇总概览，以及论文中展示的相关图表。