airoh 0.3.2

A lightweight task library for reproducible workflows

Because reproducible science takes clean tasks. And why don't you have a cup of relaxing jasmin tea?

airoh is a lightweight Python task library built with invoke, designed for reproducible research workflows. It provides pre-written, modular task definitions that can be easily reused in your own tasks.py file — no boilerplate, just useful automation. Access the documentation of the library on the airoh docs website for a list of available airoh tasks.

Installation

Installation through PIP:

pip install airoh

For local deployment:

git clone https://github.com/airoh-pipeline/airoh.git
cd airoh
pip install -e .

Usage

You can use airoh in your project simply by importing tasks in your tasks.py file.

Minimal Example

# tasks.py
from airoh.utils import run_notebooks, setup_env_python

Now you can call:

invoke run-notebooks
invoke setup-env-python

Requirements

• Python ≥ 3.8
• invoke ≥ 2.0
• Docker (for container tasks)
• Apptainer (optional, for .sif support)
• jupyter (if using run-notebooks)

Note that a few more requirements are required for development, in particular pdoc which is used to generate the documentation website.

Philosophy

Inspired by Uncle Iroh from Avatar: The Last Airbender, airoh aims to bring simplicity, reusability, and clarity to research infrastructure — one well-structured task at a time. It is meant to support a concrete implementation of the YODA principles.

License

MIT © airoh contributors