lilypond 2.25.12

A redistribution of LilyPond to use it easily from Python code.

LilyPond wheels

What's this?

GNU LilyPond is a text-based music typesetter. Wheels are the standard binary distribution format for Python packages. This repository contains a script that is being used to package LilyPond in wheel format. The resulting wheels are uploaded on PyPI (see the lilypond project page). The goal is to make it easy to use LilyPond from Python code, installing LilyPond as if it were a normal Python package — which allows to include it in your dependencies, pin it, makes tools like tox able to install it, and generally makes all the tooling around standard Python packages just work.

Usage

If you want to have the LilyPond wrapper package available in your Python environment, you can install it using

pip install lilypond

As always, consider installing packages in a virtual environment instead of globally.

You can of course use it in all the other normal ways a package on PyPI can be used: pyproject.toml project.dependencies list, dependencies in tox.ini, requirements.txt files, etc.

pip install lilypond does NOT make lilypond available as a command — this is normal! This wrapper is specifically intended for using LilyPond from Python code, not for installing LilyPond in general. In order to use the bundled LilyPond, you should do:

import subprocess

import lilypond

subprocess.run([lilypond.executable(), "file.ly"])

That is, import the lilypond module, and use the executable() function to get the full path to the bundled lilypond executable, as a pathlib.Path object, then use the subprocess module to execute it.

Scripts shipped with LilyPond are also available; use the optional argument to executable(), e.g., executable("convert-ly"), to access them.

Maintenance

When a new LilyPond release is out, run

hatch run build <version> 0

This leaves wheels of LilyPond <version> in build/. If they look good (use a venv to check), upload them with

python -m twine upload build/*.whl

The 0 in the hatch command is the build number. In case something goes wrong in published wheels, fixed wheels should be uploaded with a build number of 1 (then 2, etc.).

For type checking:

hatch run mypy:check