An unbiased evaluation of Python packaging tools - Talk Python to Me Ep.436

How well do you know your Python packaging tools? These are things like pip which install your project's dependencies and their dependencies and so on. In this mix, we have more modern tools such as Poetry, Flit, Hatch and others. And even tools outside of Python itself which may attempt to manage Python itself in addition to the libraries. To make sense of all of this, we welcome back Anna-Lena Popkes for an unbiased evaluation of environment and packaging tools.
▬▬▬▬ About the podcast ▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬
This video is the uncut, live recording of the Talk Python To Me podcast ( talkpython.fm ). We cover Python-focused topics every week and publish the edited and polished version in audio form. Subscribe in your podcast player of choice (100% free) at talkpython.fm/subscribe.
▬▬▬▬ Guests ▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬
Anna-Lena's website: alpopkes.com
Anna-Lena on GitHub: github.com/zotroneneis
Accompanying Blog Post: alpopkes.com/posts/python/pac...
Talk from PyCon DE: • Anna-Lena Popkes: An u...
Talk from EuroPython: • An unbiased evaluation...
▬▬▬▬ Links and resources from the show ▬▬▬▬▬▬▬▬▬▬▬▬
Rye: github.com/mitsuhiko/rye
Poetry: python-poetry.org/
Material for MkDocs: squidfunk.github.io/mkdocs-ma...
100 Days of Python in a Magical Universe Episode: talkpython.fm/episodes/show/1...
pip-tools: pip-tools.readthedocs.io/en/l...
Hatch: hatch.pypa.io/latest/
PDM: pdm.fming.dev/latest/
Flit: flit.pypa.io/en/stable/
Conda: docs.conda.io/en/latest/
Pipenv: pipenv.pypa.io/en/latest/
Listen this episode on Talk Python: talkpython.fm/episodes/show/4...
Episode transcripts: talkpython.fm/episodes/transc...
▬▬▬▬ Dive deeper ▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬
Listen to the Talk Python To Me podcast at talkpython.fm Over 250 hours of Python courses at training.talkpython.fm/courses Follow us on on Mastodon. Michael: fosstodon.org/@mkennedy & Talk Python fosstodon.org/@talkpython

[@NostraDavid2]

Rye uses uv from Astral, instead of classic pip-tools nowadays. Heck, Rye has been taken over by Astral (with the blessings of its creator). Besides uv it also uses Ruff for formatting and linting.

[@LewisCowles]

pip can support package hashes, so it can be used for locking, if you so desire. Because of site-packages, I don't feel like lock-files are as useful

[@sillybuttons925]

I feel like there was not enough discussion on virtualenv and setuptools.

[@not_a_human_being]

this all should be prefixed by "if you're beginner - stick with conda". I know you guys don't like it, but it's good for most beginner/lower-intermediate user-groups. Otherwise "pipenv". And if you know really what you're doing you should go further. As for python version - that shouldn't concern you, UNLESS you're writing a package that you intend to publish. Stop mixing it all together, stop scaring beginners - python is an amazing language that will continue to dominate, but these "tables" they really missing the point. It's really question of how your code will be used: - locally - you'll create a report and that's that - api - you'll be running it yourself on a server - package - you'll be shipping it via PyPI or something for other people to use The further down this list you go the better you need to understand your package management.

[@antlister]

This is good advice.

[@NostraDavid2]

I think most beginner cases would work fine with python, pip and a requirements file. It's simple, literally anyone can use it, and there's enough information out there. If you want a modern Project Manager that doesn't break things every other version (looking at you, Poetry) then Rye is the way to go. It's all in one, and handles Python versions for you (which is such a breather). Oh, and it's fast (written in Rust, BTW 🦀), uses uv instead of pip, and Ruff instead of black, flake8, pylint, isort, and a bunch others.

[@DreamsAPI]

Hi, links to the article?