ROS Bug Dataset Collection

Tools for collecting a bug dataset focused on behavioral architecture miscomposition bugs in ROS (Robot Operating System).

Prerequisites

• Docker installed on your machine
• Python 3.8.10
• Pipenv
• GPU support for running machine learning models (optional, but recommended)

Setup

1. Clone the Repository

git clone <your-repository-url>
cd <repository-name>

2. Install Dependencies

pipenv install

3. Set Up GitHub Token

1. Create a GitHub Personal Access Token:

   • Visit GitHub Token Settings
   • Generate a token with appropriate repository access permissions

2. Run the Setup Script:

   bash setup.sh

   • The script will prompt you to enter your GitHub Personal Access Token
   • It creates a .env file with secure, restricted permissions

4. Set Up Llama Model

1. Start Ollama Docker Container:

   docker compose up -d

   Note: If your machine does not have a GPU, you need to adjust the compose.yml file by uncommenting the following lines to remove GPU requirements:

   # deploy:
   #   resources:
   #     reservations:
   #       devices:
   #         - driver: nvidia
   #           count: 1
   #           capabilities: [gpu]

2. Initialize Llama Model:

   docker exec -it ollama_service ollama run llama3

Usage

Running Commit Difference Analysis

To analyze commit differences for a specific repository:

pipenv run python commit_diff.py <Owner/repo>

Example

pipenv run python commit_diff.py autowarefoundation/autoware

Output

After running the script, a CSV file will be generated with the following format:

Filename: autoware_verified_commits.csv

Columns:

• link: Direct link to the GitHub commit
• answer: Binary classification (True/False) indicating whether the commit represents a Behavioral Architecture Miscomposition Bug (BABC)
• reason: Detailed explanation of why the commit was classified as a BABC or not

Example CSV Contents:

link,answer,reason
https://github.com/autowarefoundation/autoware/commit/abc123,True,"Significant changes in node communication pattern indicating potential behavioral architecture miscomposition"
https://github.com/autowarefoundation/autoware/commit/def456,False,"Minor refactoring without substantial behavioral changes"

Configuration

Arguments

• repo_name: GitHub repository name in the format 'Owner/repo'
  • Example: autowarefoundation/autoware

Troubleshooting

• Ensure Docker is running before executing scripts
• Verify your GitHub token has the necessary permissions
• Check that Pipenv environment is activated