Open dgunning opened 5 years ago
Since we might want to use pyviz it might cause an issue since pyviz works with conda and not that well with pip.
Make it an optional dependency?
Test that we can run the project on checkout using these instructions
https://github.com/ocan-data/OpenCanadaData/blob/master/conda-environments.md
The canadadataml.yml and canadadata.yml work and I was able to both create environments and Jupyter Kernels.
I had issues with canadadataviz.yml
tomi@tomi-Fenix ~/S/OpenCanadaData> conda env create -f canadadataviz.yml Collecting package metadata: done Solving environment: failed
UnsatisfiableError: The following specifications were found to be in conflict:
Ok .. PyViz requires Python 3.6 https://pyviz.org/installation.html. I'll change this .. since you aren't really going hard into visualization just now - I think you were going to do ML.
Can you also update? conda env update -f ocandata.yml
One a side note I've found out that if we add a dependency to the yml file, and update the conda environment using conda env update -f ocandata.yml we should also refresh the Jupyter kernel so the new library shows up in Jupyter python -m ipykernel install --user --name ocandata --display-name ocandata
I will blog about this and maybe create test cases. My future workflow will be just
We want to make sure that the project is runnable when checked out. So add or change the required python files to allow a new user to checkout the project.
To run the project will need a test script or notebook although the preference is to make notebooks primary