pykiln

Version française

pykiln is the Python engine of Kiln: it helps configure new or existing projects — configuration files, CI/CD, documentation, tests, IDE run configurations — so they reach a high level of quality from the start.

pykiln is written in Python, but the projects it generates do not have to be: Python and Rust come first, other languages follow as language packs.

Status: early development. Version 0.0.1 on PyPI only reserves the name and provides no functionality. The command line interface is not settled; this README will document it once it is.

What it does

pykiln runs in four phases:

1. Inventory and questions — reads the existing project and asks what it cannot detect, then shows a plan. Nothing is written.
2. Build or transform — applies the approved plan.
3. Verify — checks that the environment syncs and that the quality toolchain and tests run. The previous setup is only removed once this passes.
4. Remote repository — configures the remote repository (labels, milestones, issue templates, mirror).

For a Python project, pykiln detects an existing virtual environment (venv, virtualenv, Poetry, Pipenv) and offers to move it to uv. It is an offer, not a forced migration.

IDE support targets PyCharm and RustRover first, through shared run configurations in .run/ and .editorconfig. pykiln does not write into .idea/.

Installation

Once a functional version is released, the recommended way to install pykiln will be:

uv tool install pykiln

uv installs and manages its own Python, so no system Python is needed. pip install pykiln will also work on a system with Python 3.10 or later.

Development

Requirements: uv.

git clone https://gitlab.com/open-works/kiln/pykiln.git
cd pykiln
uv sync            # creates .venv with the dev dependency group
uv run tox         # quality checks and test suite

Always run tox through uv run tox: uvx tox starts tox in an isolated environment without the tox-uv plugin.

Contributing

Bugs, features and tasks are tracked in this project's issues. Design decisions that apply to all of Kiln live in the governance project; decisions that only concern pykiln's own Python stack are filed here, with the same DT-NNN numbering.