🍪🔮 Template to create Translator Reasoner APIs with the TRAPI Predict Kit

A cookiecutter template to bootstrap machine learning prediction experiments, and integrate the predictions generated by this experiment to a Translator Reasoner API.

🛠️ Getting Started

1. Install cookiecutter with:

pip install cookiecutter

2. Run cookiecutter with:

cookiecutter https://github.com/MaastrichtU-IDS/cookiecutter-trapi-predict-kit

3. Enter the requested information, then win! Remember, package names should only have letters, numbers, and underscores.

4. If you're working under version control, copy the repository into your folder tracked under git, commit the files, and push to your remote.

💪 Features

Your new python package will have the following:

• Standard src/ layout
• Declarative setup with pyproject.toml
• Reproducible tests with pytest
• Data management with dvc
• Automated running of tests on each push with GitHub Actions

🏗️ Development

One liner to quickly reset a create default repo and run training:

cd .. && rm -rf my-model && cookiecutter ./cookiecutter-trapi-predict-kit && cd my-model && hatch run train