Cleanup dask dependency

[ericpre]

Checklist

• Used a personal fork of the feedstock to propose changes
• Bumped the build number (if the version is unchanged)
• [n/a] Reset the build number to 0 (if the version changed)
• [n/a] Re-rendered with the latest conda-smithy (Use the phrase @conda-forge-admin, please rerender in a comment in this PR for automated rerendering)
• [n/a] Ensured the license file is being packaged.

[conda-forge-webservices]

Hi! This is the friendly automated conda-forge-linting service.

I just wanted to let you know that I linted all conda-recipes in your PR (recipe) and found it was in an excellent condition.

[jakirkham]

Thanks Eric! 🙏

Think this would be best handled upstream first and then this could be updated accordingly in a subsequent release

[ericpre]

Thank you @jakirkham for the review. What does it need to handle upstream? As I understand the conda-forge dask-core package is the equivalent of installing dask without any extras:

dask: https://github.com/dask/dask/blob/461355e117958354ed30e29893afecd5e9258ff0/pyproject.toml#L29-L45
dask-core-feedstock: https://github.com/conda-forge/dask-core-feedstock/blob/e3cfc3964322f9421c7eed04af30d8e8924edddb/recipe/meta.yaml#L26-L35

The equivalent of the conda-forge dask package is dask['complete'], the conda-forge dask package has more dependencies (cytoolz and bokeh):

https://github.com/dask/dask/blob/461355e117958354ed30e29893afecd5e9258ff0/pyproject.toml#L65-L68
https://github.com/conda-forge/dask-feedstock/blob/c4a13519927e95614a24285ebb06916dd9b369fc/recipe/meta.yaml#L22-L31

[jakirkham]

Referring to dask-cuda's source as upstream (as indicated by the link above)

[ericpre]

Sorry, I don't understdand what needs to be change upstream: dask-cuda have dask (not dask['complete']) as dependency and this would translate to use dask-core in conda-forge.
For example, when installing dask-cuda from pypi, pyarrow and bokeh will not be installed, while it will be installed when installing from conda-forge. This is why dask-core need to be used instead of dask in conda-forge.

[jakirkham]

Could you please start by raising this as an issue on the dask-cuda repo? Think this will be easier to discuss there

[ericpre]

Done!