araras 2.3.0

A collection of tools for Machine Learning and Data Science

This is a python module that provides a set of tools for working with machine learning models. It includes utilities for neural architecture search using optuna, builders and helpers for keras/tensorflow, a monitoring system for the kernel, and several other features. The module is designed to be easy to use and flexible, allowing users to customize their machine learning workflows.

Table of Contents

• Table of Contents
• 📚 API Documentation
• ⚙️ Installation Instructions
• 🐧 Linux GPU Venv Setup (TensorFlow + Torch)
• 🚀 Release Flow
• 🔖 Versioning Policy
• 🤝 Contributing
• 📜 License
• 🤝 Collaborators

📚 API Documentation

For comprehensive documentation, examples, and detailed usage guides, please visit our Documentation Wiki.

⚙️ Installation Instructions

Install only the feature set you need:

pip install araras
pip install araras[tensorflow]
pip install araras[torch]
pip install araras[viz]
pip install araras[notebook]
pip install araras[gnn]
pip install araras[all]

Notes:

• The base install is lightweight and excludes heavyweight ML backends.
• TensorFlow support is enabled via the tensorflow extra.
• PyTorch support is enabled via the torch extra.
• Visualization and notebook extras are optional and independent.

🐧 Linux GPU Venv Setup (TensorFlow + Torch)

For Linux users with an NVIDIA GPU, this repository includes an installer script at venvs/tf-gpu.sh that creates and configures a virtual environment with:

• TensorFlow (tensorflow[and-cuda])
• PyTorch (torch, torchinfo, torchviz)
• Optuna and common ML/data-science utilities
• araras[all]

The script also verifies CPU and GPU TensorFlow availability at the end.

Run it from the repository root:

chmod +x venvs/tf-gpu.sh
./venvs/tf-gpu.sh

Important notes:

• Linux only (script uses apt and nvidia-smi).
• You will be prompted for the virtual environment name and location.
• A supported NVIDIA driver is required (the script checks this automatically).
• The script uses python3.12 by default.

🚀 Release Flow

Maintainer quick path:

python -m build
twine check dist/*
git tag v1.0.0
git push origin v1.0.0

Tag pushes matching v* trigger the publish workflow.

🔖 Versioning Policy

• The value in pyproject.toml project.version must match the Git tag version.
• Release order: bump version, merge to main, tag as v, push tag.
• PyPI versions are immutable and cannot be re-used.

🤝 Contributing

Contributions are what make the open-source community amazing. To contribute:

1. Fork the project.
2. Create a feature branch (git checkout -b feature/new-feature).
3. Commit your changes (git commit -m 'Add some feature').
4. Push to the branch (git push origin feature/new-feature).
5. Open a Pull Request.

📜 License

This project is licensed under the General Public License.

🤝 Collaborators

We thank the following people who contributed to this project:

Matheus Ferreira