NullGTN artifact

A tool that automatically reannotates Java projects using graph learning models.

Undo split:

cat nullgtn-update.tar.part-* > nullgtn-update.tar

To run:

sudo docker run -d -p 2222:22 --gpus all nullgtn-update:latest
OR
sudo docker start nullgtn-update

ssh root@127.0.0.1 -p 2222
Password: NullGTNv.4

Requirements:

Host OS: Tested on Ubuntu 24.04. Use other Linux at your own risk.
NVIDIA GPU required with at least 16 GB memory.

Make sure drivers are installed and update on both host and guest:

# 1. Update package lists
sudo apt-get update

# 2. List available NVIDIA drivers for your GPU
sudo ubuntu-drivers list --gpgpu

# 3. Automatically install the best server/compute driver
sudo ubuntu-drivers install --gpgpu

# 4. Reboot to load the new driver
sudo reboot

On host, make sure the --gpus flag is available for Docker:

# A. Add NVIDIA.s GPG key and repository
curl -fsSL https://nvidia.github.io/libnvidia-container/gpgkey \
  | sudo gpg --dearmor -o /usr/share/keyrings/nvidia-container-toolkit-keyring.gpg \
  && curl -s -L https://nvidia.github.io/libnvidia-container/stable/deb/nvidia-container-toolkit.list \
     | sed 's#deb https://#deb [signed-by=/usr/share/keyrings/nvidia-container-toolkit-keyring.gpg] https://#g' \
     | sudo tee /etc/apt/sources.list.d/nvidia-container-toolkit.list

# B. Install the toolkit
sudo apt-get update
sudo apt-get install -y nvidia-container-toolkit

# C. Tell Docker to use it and restart
sudo nvidia-ctk runtime configure --runtime=docker
sudo systemctl restart docker

Directories:

precision_recall: Show precision. Use the conda environment: conda activate naenv
warnings: Generate warning logs. Use the conda environment: conda activate nullgtn

See README.md in each subdirectory for directions.