Online Adaptable Bug Localization for Rapidly Evolving Software

This repository contains the code and data associated with the following research paper:

A. Ciborowska, M.J. Decker, K. Damevski, "Online Adaptable Bug Localization for Rapidly Evolving Software"

Requirements

• Python3.6
• git
• Packages listed in requirements-3.txt file: pip install -r requirements.txt

Running the project

To run the project with default options simply use the following command:

python3.6 main.py [options]

Options:

• --datasets - path to datasets: datasets/corley, datasets/bench4bl or datasets/hcc. Default: datasets/bench4bl.
• --name - name of project to run the experiment on. If not provided, run experiments for all projects in the dataset.
• --model-type - type of BL model to use. Use joined to run JINGO, changesets to run Corley et al.[1]. Default: joined.
• --topics - number of topics for the bug report model and the changeset model. Format: [bug_topics, changeset_topics]. Default: [50, 100].
• --decays - decays for for the bug report model and the changeset model. Format [bug_decay, changeset_decay]. Default: [0.75, 1.0].
• --gamma - boosting parameter for JINGO prediction. Default: 1.0.
• --omega - boosting for the number of fixed issues to observe to train T matrix. Default: 1.0.
• --save-model - save model after training (flag). Default: off.
• -v - enable verbose output (flag). Default: off.

All options are defined at the top of main.py.

Example:

Run JINGO with the following configuration:

• dataset - datasets/corley
• project - BookKeeper
• the bug report model - 25 topics, 0.5 decay
• the changeset model - 200 topics, 0.85 decay
• omega = 1.0
• gamma = 5.0

python3.6 main.py --datasets datasets/corley --name bookkeeper --num-topics 25 200 --decays 0.5 0.85 --omega 1.0 --gamma 5.0

Results

Results for each project are stored in results/project_name folder.