PR_CodeSmells

Replication package for the paper detection of code smells in pull requests on GitHub.

How-to

1. Clone the repository

$ git clone https://github.com/Tools-Demo/PR_CodeSmells.git

2. Place the dataset file in PR_CodeSmells-main/Dataset (Download from 25_projects_PRs.csv)

3. Change directory

$ cd PR_CodeSmells-main

4. Execute python classify_quality_PRs_1.py

5. Output for each executed method will be generated as a CSV file under "Results" folder (Existing results are placed in "Results_old" folder)

6. Methods that can be executed in the main function:

$ train_models_using_10FoldCV(df) [Default]  -- Train all classifiers with 10-fold cross-validation
$ Baseline_10FoldCV(df) -- Train baseline classifier with 10-fold cross-validation
$ baseline_classifer(df)  -- Train baseline classifier
$ hyperparamGridSearch(df)  -- Execute grid search for HPO