Add CuPy support to ArrayStore via Array API

[btjanaka]

Description

This PR supports CuPy via the Array API. At first I thought this might be challenging since CuPy has a different model of devices, i.e., it depends on context managers rather than passing in devices. However, array-api-compat seems to have solved these issues and made CuPy compatible in terms of devices. Here are some references that I looked at:

• [DNM] ENH: CuPy multi-device support data-apis/array-api-compat#293
• https://github.com/data-apis/array-api-compat/blob/main/array_api_compat/cupy/_aliases.py
• https://docs.cupy.dev/en/stable/reference/array_api_functions.html#cupy.array_api.ones

TODO

• Add cupy backend to conftest
• Add instructions for installing cupy
• Test ArrayStore with CuPy on GPU
• Update ArrayStore to consider CuPy in pandas return type
• Add cupy tests to CI -- I installed CuPy but the CuPy tests won't actually run because we don't have a GPU runner set up -- following scipy (https://github.com/scipy/scipy/blob/main/.github/workflows/gpu-ci.yml), I also set the CI to only run on ubuntu, since installing cupy on mac and windows seems to be problematic

Status

• I have read the guidelines in
  CONTRIBUTING.md
• I have formatted my code using yapf
• I have tested my code by running pytest
• I have linted my code with pylint
• I have added a one-line description of my change to the changelog in
  HISTORY.md
• This PR is ready to go