Update docs theme, use sphinx-tabs for CPU/GPU examples

[charlesbluca]

This PR bumps the dask-sphinx-theme to be more in line with Dask / Distributed's docs, and adds the sphinx-tabs extension so that code-blocks can be tabbed to show their GPU equivalent (when possible)

[codecov-commenter]

Codecov Report

Merging #394 (9d804fb) into main (c3ad6a9) will increase coverage by 0.15%.
The diff coverage is n/a.

@@            Coverage Diff             @@
##             main     #394      +/-   ##
==========================================
+ Coverage   89.02%   89.17%   +0.15%     
==========================================
  Files          69       69              
  Lines        3326     3326              
  Branches      653      653              
==========================================
+ Hits         2961     2966       +5     
+ Misses        296      287       -9     
- Partials       69       73       +4     

Impacted Files	Coverage Δ
dask_sql/_version.py	34.00% <0.00%> (+1.44%)	⬆️

📣 Codecov can now indicate which changes are the most critical in Pull Requests. Learn more

[charlesbluca]

@VibhuJawa if you get a chance could you check out machine_learning.rst and let me know if there's any other blocks in there that can be run on GPU? Based my examples off the GPU model tests but not sure if I missed anything 🙂

[VibhuJawa]

@VibhuJawa if you get a chance could you check out machine_learning.rst and let me know if there's any other blocks in there that can be run on GPU? Based my examples off the GPU model tests but not sure if I missed anything 🙂

Looking at it right now. We need to add to do two things:

1. Add explanation around wrap_fit and wrap_predict .
2. Use the same cuML and sklearn model to avoid confusion

Will make a PR with the suggested changes to this PR in a bit.

[VibhuJawa]

@charlesbluca , Raised a PR herre charlesbluca#2 .

Also, looking at the other sections (see below) of the machine-learning docs, they should probably work with GPU's . We don't have have tests for those now though. I think it might be worth-while adding tests for them and adding the GPU sections for those too. (Obviously in a follow up PR )

1. section 5
2. section-6
3. example

What do you think ?

[VibhuJawa]

Raised #397 to track the update to other ML sections.

[charlesbluca]

That sounds good to me - I can make TODOs for those examples so that we can remember to add GPU sections there in a follow up.

It's not immediately obvious what code changes would be made to example 6 to account for GPUs - or is the idea that a model created with a cuML class should have no issues being exported?

[VibhuJawa]

It's not immediately obvious what code changes would be made to example 6 to account for GPUs - or is the idea that a model created with a cuML class should have no issues being exported?

Depends on the model, i think most models work but some models like FIL have issues . We would have to test it to know. Anyways, having this under CI will give us confidence after putting stuff under CI.

That sounds good to me - I can make TODOs for those examples so that we can remember to add GPU sections there in a follow up.

Yup. Raised issue to track this, #397 . Lets add stuff there. :-)

[charlesbluca]

Going to go ahead and merge this in, as the gpuCI failures are unrelated