pyscaffold / pyscaffoldext-dsproject

💫 PyScaffold extension for data-science projects

Home Page:https://pyscaffold.org/projects/dsproject

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Test to ensure that `environment.yml` can install without error

sgbaird opened this issue · comments

Describe your use-case

I've often found that the environment.yml file can't be successfully resolved when users first try out the repository in development mode (rather than installing the Python package directly). See for example sparks-baird/self-driving-lab-demo#164. I tend to keep setup.cfg up-to-date, but I'm not regularly checking environment.yml.

Describe the solution you would like to see for your use-case

It would be nice to add a step to GitHub actions that ensures the environment.yml file can be successfully installed. This could also be a step that only runs when environment.yml gets changed to reduce the overhead.

Alternatively, there could be a note in the README installation section to try:

conda create -n <name-of-package> python==3.10.*
pip install -e .

in case conda create -f environment.yml produces an error, and manually install additional packages from environment.yml as needed (e.g., for running notebooks).

Hi @sgbaird, thanks for suggesting this improvement and sorry for my late reply.

In which cases does conda create -f environment.yml produce an error? Would this then rather be an error within the conda package repository itself?