# SeC-GAN

**Repository Path**: passer_x/SeC-GAN

## Basic Information

- **Project Name**: SeC-GAN
- **Description**: No description available
- **Primary Language**: Unknown
- **License**: Not specified
- **Default Branch**: main
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 0
- **Created**: 2022-01-21
- **Last Updated**: 2022-01-21

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

# SeC-GAN - Generative Adversarial Network for Just-in-Time Defect Prediction with Semantic Changes Sensitivity

## I'm grateful that you're interested in my work! :)

I hope the following guide will make it easier for you to understand what we did.

## Requirements

Install all requirements in the "requirements.txt" file


## Step 1 - Extract Data:

1. Download the folder named bic and javadiff.
2. Requirements: 
    * Python 3.9 - then run: 
    ```
      python -m pip install --upgrade pip
      pip install pytest 
      pip install gitpython
      pip install jira
      pip install termcolor 
      pip install openpyxl  
      pip install javalang
      pip install pathlib
      pip install junitparser
      pip install pandas
      pip install numpy
      pip install pydriller
      pip install dataclasses
      pip install jsons
      python ./javadiff/setup.py install
     ```
    * java version 11 and 8
3. Checkout to directory name "./local_repo" the repository. For example: apache/deltaspike.
4. Execute: 
```
python Main.py [0]
```
  * Note: [0] - indicate  the number of commit extract in windown size of 50.  For large project set [0, 1, 2.....200].
 5. After the run the file save in "./results/" folder. 
 6. To merge all file Run:
  ```
    import pandas as pd
    import glob
    all_files = glob.glob("./results/*.csv")
    li = []
    for filename in all_files:
        df = pd.read_csv(filename, index_col=None, header=0)
        li.append(df)
    pd.concat(li, axis=0, ignore_index=True).to_csv('./results/all.csv', index=False
   ```
 7. All data save in './results/all.csv'.
    

## Step 2 - Preprocessing and SZZ algorithm

1. Open Data folder named  "Data".
2. In Data folder open for each project directory with the name NAME_PROJECT and put the file "all.csv" in the directory. For example  (Data -> mahout -> all.csv). 
3. Additionally, you can put the extracted  data from this repository in directory Data. 
4. In "variable.py" add the NAME_PROJECT and the key_issue (according to JIRA) to function get_key_issue().
5. Updare NAME_PROJECT in line 83 and Run :
   ```
   python main_create_data.py
   ```
7. After the run train and test file created in the path (Data -> NAME_PROJECT -> train_test). 

## Step 3 -  [Train CTGAN](https://github.com/sdv-dev/CTGAN)

1. Train CTGAN algorithm in folder CTGAN -> "run.py".
   ```
   python run.py
   ```
3. You can change the parameters() function with another parameters.
4. After this process update in (Data -> NAME_PROJECT -> CTGAN -> FILE_CREATE) the FILE_CREATE from this runing.


## Step 4 - Run Algoritham 
1. In folder Algorithm ave 3 file - "RF.py", "LR.py" and "Sec_GAN.py".
2. You can run each algorithm with '__main__' - you can change  the name project. 
3. Additionally,  you can change the parameters of "Sec_GAN.py" in the parameters dict. 


 
## Good luck !!!