google-colab package error when using new base environment

[proppy]

I get the following error when using condacolab master with manually built https://github.com/conda/constructor/ installers?

Collecting package metadata (current_repodata.json): ...working... done
Solving environment: ...working... failed with initial frozen solve. Retrying with flexible solve.
Collecting package metadata (repodata.json): ...working... done
Solving environment: ...working... failed with initial frozen solve. Retrying with flexible solve.

PackagesNotFoundError: The following packages are not available from current channels:

  - google-colab

Current channels:

  - https://repo.anaconda.com/pkgs/main/linux-64
  - https://repo.anaconda.com/pkgs/main/noarch
  - https://repo.anaconda.com/pkgs/r/linux-64
  - https://repo.anaconda.com/pkgs/r/noarch

To search for alternate channels that may provide the conda package you're
looking for, navigate to

    https://anaconda.org

and use the search bar at the top of the page.

Reverting to pip install https://github.com/conda-incubator/condacolab/archive/28521d7c5c494dd6377bb072d97592e30c44609c.tar.gz seems to solve the issue, so I suspect this was introduced with #31?

[jaimergp]

Thanks for the report! Yes, #31 is behind this. We might need to force the usage of conda-forge in the patching instructions.

For now, you might be able to work around it with CONDA_CHANNELS=conda-forge,defaults or, conversely, CONDA_CHANNELS=defaults,conda-forge.

[proppy]

We might need to force the usage of conda-forge in the patching instructions.

Curious if this will cause the conda-forge index to be fetched by default during the condacolab.install()?

[jaimergp]

Yes, it will incur some overhead.

The idea was to provide google-colab dependencies through conda and then override with the pip install --no-deps command.

[proppy]

The idea was to provide google-colab dependencies through conda

Oh does that mean that if we bundle google-colab in our constructor installer, we don't need to enable conda-forge?

[jaimergp]

Yes, that would be the best option! We still need to document this workflow with the new option, but first we need to figure it out 😬