DCI

DCI (Detecting behavioral changes in CI) is a dataset comprising 60 commits selected from 6 notable open source Java projects. This dataset is curated to benchmark an approach that detects behavioral changes in commits, specifically focusing on changes that affect the behavior of unit tests. The dataset is used to evaluate the effectiveness of DCI in generating test methods that capture the behavioral differences between pre-commit and post-commit versions of the program. The dataset ensures that each commit introduces a behavioral change and includes a human-written test case that detects this change, serving as a ground truth for analysis. The application of DCI aims to enhance the quality assurance process in continuous integration by automatically generating relevant test cases for each commit.

DCI（Detecting behavioral changes in CI，持续集成行为变化检测）是一个包含60次代码提交的数据集，样本选自6个知名开源Java项目。该数据集专为基准测试一种可检测代码提交中行为变化的方法而精心构建，且重点聚焦于影响单元测试行为的变更。该数据集被用于评估DCI生成可捕捉程序提交前与提交后版本行为差异的测试方法的有效性。该数据集确保每一次代码提交均引入了明确的行为变化，同时包含人工编写的、可检测该变化的测试用例，以此作为分析所用的基准真值（ground truth）。DCI的应用旨在通过为每一次代码提交自动生成相关测试用例，优化持续集成流程中的质量保障工作。