纵向代码覆盖率与缺陷引入关联数据集
收藏资源简介:
该数据集由帕绍大学团队构建,聚焦JavaScript和TypeScript开源项目中代码覆盖率与缺陷引入的因果关系研究。核心数据涵盖纵向代码覆盖率指标、源代码度量、问题跟踪与代码评审内容,以及缺陷引入变更记录,通过WayPack Machine工具实现npm/yarn依赖环境的时序重建。数据采集严格限定于具备稳定测试套件的成熟项目,要求90%以上的提交可计算行级覆盖率,并整合GitHub Actions的CI日志。其创新性在于首次将因果推理框架应用于软件工程领域,旨在量化代码覆盖率对缺陷预防的实际效用,为开发实践中的测试策略优化提供实证依据。
This dataset was constructed by a research team from the University of Passau, focusing on causal research into the relationship between code coverage and defect introduction in open-source JavaScript and TypeScript projects. The core dataset covers longitudinal code coverage metrics, source code metrics, issue tracking and code review content, as well as change records tied to defect introductions. The temporal reconstruction of npm/yarn dependency environments is achieved via the WayPack Machine tool. Data collection is strictly constrained to mature projects with stable test suites, mandating that line-level coverage can be calculated for over 90% of commits, and integrates CI logs from GitHub Actions. Its innovative contribution lies in the first application of causal inference frameworks to the field of software engineering, aiming to quantify the actual efficacy of code coverage in defect prevention and provide empirical evidence for optimizing testing strategies in development practices.

- 1Causal Inference for the Effect of Code Coverage on Bug Introduction帕绍大学 · 2026年



