Enhancement(Infra): Move away from Poetry

[JacobCoffee]

Summary

This issue serves as a launchpad to potentially swap from Poetry to another solution. We don't have to change or go anywhere, but this is just to get an idea for options, thoughts, and concerns.

Poetry causes many issues, its dependency resolution sometimes takes multiple minutes, they have more than once made questionable changes (python-poetry/poetry#6297), dependencies leak outside of their intended target (https://discord.com/channels/919193495116337154/919193495690936353/1131650830060245094, python-poetry/poetry#4401).

Basic Example

No response

Drawbacks and Impact

• New tool = new learning curve for everyone
• CI Changes

Unresolved questions

Where to?

• PDM
• Hatch
• PDM + Hatch's build system, Hatchling
• Rye (experimental)
• Huak (experimental)

Other

This doesn't really solve the "The Python packaging ecosystem is terrible overall" issue...

NOTE: PDM does have a potential issue around PEP582 being rejected.

Funding

• If you would like to see an issue prioritized, make a pledge towards it!
• We receive the pledge once the issue is completed & verified

[peterschutt]

I noticed recently that hatch has been adopted by attrs (they are a project I've always viewed as setting a good standard as far as project/build/testing configuration goes).

One thing hatch doesn't have is lock files: ref, ref. However, that's not really an end-user issue, and seems we could do something to work around that by using pip-tools to compile a hashed set of requirements for development environments.

Given that hatch is a PyPA project - it probably will be a bit more "boring" than other package managers out there - but sometimes boring is a good thing.

Just my 2c.

[cofin]

I recently tested PDM + hatchling and found that it was a really solid poetry replacement. It had nearly drop-in feature parity with most of the poetry commands I use.