Skip to main content

nox-poetry

Project description

nox-poetry

PyPI Python Version License

Read the documentation at https://nox-poetry.readthedocs.io/ Tests Codecov

Use Poetry inside Nox sessions

This package provides a drop-in replacement for the nox.session decorator, and for the nox.Session object passed to user-defined session functions. This enables session.install to install packages at the versions specified in the Poetry lock file.

from nox_poetry import session

@session(python=["3.10", "3.9"])
def tests(session):
    session.install("pytest", ".")
    session.run("pytest")

Disclaimer: This project is not affiliated with Nox, and not an official Nox plugin.

Installation

Install nox-poetry from the Python Package Index:

$ pip install nox-poetry

Important: This package must be installed into the same environment that Nox is run from. If you installed Nox using pipx, use the following command to install this package into the same environment:

$ pipx inject nox nox-poetry

Requirements

  • Python 3.7+
  • Poetry >= 1.0.0

You need to have a Poetry installation on your system. nox-poetry uses Poetry via its command-line interface.

Usage

Import the @session decorator from nox_poetry instead of nox. There is nothing else you need to do. The session.install method automatically honors the Poetry lock file when installing dependencies. This allows you to manage packages used in Nox sessions as development dependencies in Poetry.

This works because session functions are passed instances of nox_poetry.Session, a proxy for nox.Session adding Poetry-related functionality. Behind the scenes, nox-poetry uses Poetry to export a constraints file and build the package.

For more fine-grained control, additional utilities are available under the session.poetry attribute:

  • session.poetry.installroot(distribution_format=["wheel"|"sdist"])
  • session.poetry.build_package(distribution_format=["wheel"|"sdist"])
  • session.poetry.export_requirements()

Note that distribution_format is a keyword-only parameter.

Here is a comparison of the different installation methods:

  • Use session.install(...) to install specific development dependencies, e.g. session.install("pytest").
  • Use session.install(".") (or session.poetry.installroot()) to install your own package.
  • Use session.run_always("poetry", "install", external=True) to install your package with all development dependencies.

Please read the next section for the tradeoffs of each method.

Why?

Let's look at an example:

from nox_poetry import session

@session(python=["3.10", "3.9"])
def tests(session):
    session.install("pytest", ".")
    session.run("pytest")

This session performs the following steps:

  • Build a wheel from the local package.
  • Install the wheel as well as the pytest package.
  • Invoke pytest to run the test suite against the installation.

Consider what would happen in this session if we had imported @session from nox instead of nox_poetry:

  • Package dependencies would only be constrained by the wheel metadata, not by the lock file. In other words, their versions would not be pinned.
  • The pytest dependency would not be constrained at all.
  • Poetry would be installed as a build backend every time.

Unpinned dependencies mean that your checks are not reproducible and deterministic, which can lead to surprises in Continuous Integration and when collaborating with others. You can solve these issues by declaring pytest as a development dependency, and installing your package and its dependencies using poetry install:

@nox.session
def tests(session: Session) -> None:
    """Run the test suite."""
    session.run_always("poetry", "install", external=True)
    session.run("pytest")

Unfortunately, this approach comes with its own set of problems:

  • Checks run against an editable installation of your package, i.e. your local copy of the code, instead of the installed wheel your users see. In the best case, any mistakes will still be caught during Continuous Integration. In the worst case, you publish a buggy release because you forgot to commit some changes.
  • The package is installed, as well as all of its core and development dependencies, no matter which tools a session actually runs. Code formatters or linters, for example, don't need your package installed at all. Besides being wasteful, it goes against the idea of running checks in isolated environments.

nox-poetry uses a third approach:

  • Installations are performed by pip, via the session.install method.
  • When installing your own package, Poetry is used to build a wheel, which is passed to pip.
  • When installing third-party packages, Poetry is used to export a constraints file, which is passed to pip along with the packages. The constraints file ensures that package versions are pinned by the lock file, without forcing an installation of every listed dependency and sub-dependency.

In summary, this approach brings the following advantages:

  • You can manage tools like pytest as development dependencies in Poetry.
  • Dependencies are pinned by Poetry's lock file, making checks predictable and deterministic.
  • You can run checks against an installed wheel, instead of your local copy of the code.
  • Every tool can run in an isolated environment with minimal dependencies.
  • No need to install your package with all its dependencies if all you need is some linter.

Contributing

Contributions are very welcome. To learn more, see the Contributor Guide.

License

Distributed under the terms of the MIT license, nox-poetry is free and open source software.

Issues

If you encounter any problems, please file an issue along with a detailed description.

Credits

This project was generated from @cjolowicz's Hypermodern Python Cookiecutter template.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

nox-poetry-1.0.2.tar.gz (12.9 kB view hashes)

Uploaded Source

Built Distribution

nox_poetry-1.0.2-py3-none-any.whl (11.1 kB view hashes)

Uploaded Python 3

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page