FoX: a Fo rmal eX plainer for JIT Defect Prediction

Experiment Replication

To replicate our experiment in the paper, please go to this repository.

FoX

@underreview{
}

FoX Example

git clone https://github.com/trustablefox/quickstart_example.git
cd quickstart_example

Then, open the DEMO.ipynb file for a concrete example.

Quick Start FoX Tool

pip3 install foxplainer

from foxplainer.explainer import FoX
fx = FoX(global_model_name="LR", 
           xnum='all', 
           global_model_path="./global_model/openstack_LR_global_model.pkl", 
           proj_name="openstack", 
           data_path="./dataset/",
           inst_id=5)
fx.explain(in_jupyter=False)

A user needs to provide the following parameters:

1. global_model_name: should be either "RF" (Random Forest) or "LR" (Logistic Regression)

2. xtype (optional): Only needed when xnum=1. Support two types of explanations, specify "abd" for Abductive Explanation or "con" for Contrastive Explanation

3. xnum: should be either 1 or "all" (return all explanations)

4. global_model_path: the path to your trained global model (model should be trained using sci-kit learn library)

5. proj_name: the project name of your dataset

6. data_path: the path to the required data, this path should contain the following 2 files:

   6.a {proj_name}.csv - the complete file consisting of all training + testing data with features and label

   6.b {proj_name}_X_test.csv - the testing data that only contains feature columns without label

7. inst_id: the row number of the instance to be explained in your testing data

Table of Contents

• Installation

  • Install via pip
  • Install via conda
  • Local install via poetry

• Contributions

• Documentation

• License

Install FoX Tool

pip3 install -r requirements.txt

Install via pip

pip3 install foxplainer

Install via conda

conda install -c conda-forge foxplainer

Local install via poetry

git clone https://github.com/trustablefox/foxplainer.git
poetry install

If you do not have poetry installed yet, simply run the command below to install it, for more information, click here.

(Invoke-WebRequest -Uri https://raw.githubusercontent.com/python-poetry/poetry/master/get-poetry.py -UseBasicParsing).Content | python -

Contributions

We welcome and recognize all contributions. You can see a list of current contributors in the contributors tab.

Documentation

Access our official documentation here.

License

MIT License