rohhaiil/SysMLv2_Repair_with_SLMs

该数据集用于SysML v2代码的修复任务，包含两种配置：default和full。default配置包含经过筛选的训练/验证/测试集，而full配置包含完整的数据集。数据集中的每个实例都有唯一的标识符、源代码标识符、突变类别、突变类型、错误代码、正确代码、差异补丁等字段。数据集的创建过程包括从公共和作者生成的SysML v2代码中创建种子集，引入合成错误，生成上下文信息，并最终生成目标输出。数据集发布在CC BY 4.0许可证下，要求引用原始作者。

This dataset is used for the task of repairing SysML v2 code and includes two configurations: default and full. The default configuration contains filtered train/validation/test sets, while the full configuration contains the complete dataset. Each instance in the dataset has unique identifiers, source code identifiers, mutation categories, mutation types, bad code, good code, diff patches, and other fields. The dataset creation process involves creating a seed set from public and author-generated SysML v2 code, introducing synthetic errors, generating contextual information, and finally producing target outputs. The dataset is released under the CC BY 4.0 license, requiring attribution to the original authors.