Installing from conda-envs results in upsupported python (3.13)

[williambdean]

Installing from yaml will result in dependency resolution error

conda env create -f conda-envs/environement-test.yml

Notes: I am using micromamba

Pinning will likely cause issues with the CI/CD:

https://github.com/wd60622/pymc-experimental/blob/5055262868473b61f487f42d82542f9d0fc0e273/.github/workflows/test.yml#L35-L42

Maybe some note about dev installation would be useful and sufficient

[ricardoV94]

Pinning will likely cause issues with the CI/CD:

Why?

CC @maresb

[williambdean]

Why?

Not sure if the duplicate key for python will be an issue or if the CI/CD lines that I linked will cause an issue if there is already a python key in env file

[ricardoV94]

Can we just do upper/lower pin in the yml?

[maresb]

I don't understand the formulation of this issue. Like what is even the expected vs actual behavior? What is the "dependency resolution error"?

[williambdean]

Putting lower and upper bounds in environment.yml will work.

I don't understand the formulation of this issue. Like what is even the expected vs actual behavior? What is the "dependency resolution error"?

I tried to install from yaml but python 3.13 was used didn't work with pymc versions

[maresb]

I think #403 should fix it since this should never happen on conda-forge

[williambdean]

I think #403 should fix it since this should never happen on conda-forge

Alright. I will give it a try after merge

[williambdean]

Thanks for making the adjustment.

Can confirm that

gh repo fork pymc-devs/pymc-extras
cd pymc-extras
conda env create -f conda-envs/environment-test.yml

Builds an environment with compatible python version (3.12)

[maresb]

Excellent, thanks @wd60622 !