Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Use poetry #46

Open
peterdudfield opened this issue Feb 9, 2024 · 3 comments
Open

Use poetry #46

peterdudfield opened this issue Feb 9, 2024 · 3 comments

Comments

@peterdudfield
Copy link
Contributor

Detailed Description

Should we use poetry in this repo.

Context

  • tough upgrades sometimes, casuing odd error
  • perhaps less flexible for R&D work
  • but more robust for MLops (which this repo is mainly for)

Possible Implementation

Make this a peotry project

@peterdudfield peterdudfield added enhancement New feature or request discussion and removed enhancement New feature or request labels Feb 9, 2024
@jacobbieker
Copy link
Member

Potentially conda instead? I think it might be worth standardizing on one install method, rather than a mix of poetry and conda that we have in our repos. Conda should ensure everything installed works with each other too, and can be easier for some geospatial stuff, although admittedly isn't a huge problem with this repo I think.

@peterdudfield
Copy link
Contributor Author

sorry, perhaps I got something a bit confused here.

Poetry I think is mainly used to pin requirements. Then we can use pip or conda or whatever we want to do that.

I'm not a fan of using conda as a standard, as it comes with a lot of overhead sometimes for production code. I do agree it does take care of somethings under the hood, which can be tricky otherwise. I think the divide between R&D repos, can use conda and Production ML apps using poetry could be somewhat of a nice balance.

@jacobbieker
Copy link
Member

Ah okay, I guess, for pinning requirements, any of them can work well enough, they all pin requirements the same, so I don't think that would make poetry stand out, if anything, I think it should stick to a single package manager, mixing multiple is another potential source of failure. Conda does have overhead, but that can be minimized a bit with miniconda or something like that.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

No branches or pull requests

2 participants