KBS_Just-in-time defect prediction based on AST change embedding At present, the source code files uploaded by this repository are not sorted out, and will be further supplemented later.
-
"corpus" directory: The corpus of the project is stored in the "corpus" directory.
-
"data" directory: Relevant data is stored in the "data" directory, and "tradition_data" stores traditional features, "raw_ data" stores the extracted ace change node, "ACE_ data" stores the digital vector of the pre processed AST change node and the related submission information.
-
"jars" directory: The jar files related to gumtree are stored in the "jars" directory
-
"model" directory: The training model is stored in the "model" directory
-
"project" directory: The project history data is stored in the "project" directory. The "ACE_pygit2_test_multi.py" file is used to extract the AST change token from the project.
6.preprocess.py file for processing raw data in "data/raw_data" directory.
7.train_ test_ sets_ generation.py file is used to convert the original data in the "data/raw_data" directory into training sets and test sets that can be learned directly, and the generated files are stored in "data/ACE_data" directory
This repository still includes the following issues:
-
if the file reading path in the source code is not modified, an error will be reported if the file is run directly.
-
there is a lack of necessary instructions for running steps. If you do not understand our paper, you may not know the running order of each Python file.
-
some intermediate files are still missing, including ast change node sequence file under "data/raw_ data" directory, word vector model file generated by GloVe, etc.
We will revise and supplement these in the coming period.
中文说明：
1.corpus目录下存放的是项目的语料库。
2.data目录下存放的是相关数据，tradition_data存放的是传统特征，raw_data存放的是提取出来的ACE变更节点，ACE_data存放的是预处理好的AST变更节点的数字向量，以及相关的提交信息。
3.jars目录下存放的是gumtree的相关jar文件
4.model目录下存放的是训练模型
5.project目录中存放的是项目历史数据，“ACE_pygit2_test_multi.py”文件用于从项目中提取AST更改令牌。
6.preprocess.py文件用于处理raw_data中的原始数据
7.train_test_sets_generation.py文件用于将data/raw_data中的原始数据转换成可直接学习的训练集和测试集，生成的文件存放在data/ACE_data目录下