Characteristics of AIOps Projects - Replication Package

This repository contains the replication package for the paper "Studying the Characteristics of AIOps Projects on GitHub".

Introduction

We organize the replication package into three file folders.

1. Experiments: this folder contains code for our main experiments (i.e., extracting GitHub metrics for projects (RQ1) and analyze projects in terms of code quality (RQ3));
2. Results: this folder contains the results CSV files for all of our research questions;
3. Results analysis: this folder contains code for analyzing the dataset and experiment results.

The needed libraries and versions can be found on requirements.txt file.

Install

git clone https://github.com/AIOpsstudy/understanding_aiops
pip install -r requirements.txt

Experiments

This part contains code for our main experiments for RQ1 and RQ3. All code could be found under the experiments folder. We have the following experiment code available:

• github_aiops.ipynb contains code for extracting GitHub metrics for our set of AIOps projects.
• github_ml.ipynb randomly selects repositories with the keywords machine-learning and deep-learning, and then extract their GitHub metrics.
• github_general.ipynb randomly selects repositories all over GitHub and then extract their GitHub metrics.
• sonarqube contains code for cloning projects, sending the codes to SonarQube server, analyzing the source codes, and extracting the results from SonarQube. To download and install a SonarQube instance, please refer to their documentations.

Results

This part contains the output data from our main experiments. All output CSV files could be found under the results folder.

Result Analysis

This part contains code for the analysis of our experiment results. All code could be found under the analysis folder.