stefanocarrera/autophagycode_D_he_train-mercury_Qwen3-8B_strategy_trust_t1.25_g6_run0_metrics
收藏资源简介:
该数据集是一个用于编程任务代码质量评估的数据集,包含164个训练样本。每个样本对应一个编程任务,记录了任务ID、入口函数、可执行状态、正确性判断、测试通过与失败数量、测试运行时间、错误类型等基本信息,同时提供了丰富的代码质量度量指标,包括Halstead复杂度(词汇量、长度、体积、难度、工作量、时间)、圈复杂度、可维护性指数、代码行数(LOC和SLOC)、注释比例、词汇丰富度(TTR)、词汇表、香农熵、预测熵(均值和最大值)以及定义函数数量等。数据集适用于代码质量分析、自动化评估、编程任务复杂度研究等领域。
This dataset is designed for code quality assessment of programming tasks, containing 164 training examples. Each example corresponds to a programming task and includes basic information such as task ID, entry point function, executability status, correctness judgment, number of tests passed/failed, test run time, and error type. It also provides comprehensive code quality metrics, including Halstead complexity (vocabulary, length, volume, difficulty, effort, time), cyclomatic complexity, maintainability index, lines of code (LOC and SLOC), comment percentage, lexical richness (TTR), token dictionary, Shannon entropy, predictive entropy (mean and maximum), and number of defined functions. The dataset is suitable for research in code quality analysis, automated assessment, and programming task complexity studies.




