please also add test action run, example commit change refs/links to desc for review, thanks

cc @leewyang check if the workflow and actual doc content meet requirements, thanks

https://github.com/YanxuanLiu/spark-rapids-ml/actions/runs/4632181048
Success action run with tagging in fork repo

The commit made by the action:
YanxuanLiu@51182ee

LGTM for CI part. Please help verify if any permission issue for automation account in this repo after merge (retag v22.02.0), thanks!

Waiting for developer to help check the content, thanks!

I think we still need the cuml/rapids dependencies to be available/installed, e.g. like the CI/CD dockerfile. Otherwise, the API doc generation doesn't quite work, e.g.

• current docs (manually generated): https://nvidia.github.io/spark-rapids-ml/api/python/spark_rapids_ml.html#module-spark_rapids_ml.knn
• this PR: https://yanxuanliu.github.io/spark-rapids-ml/api/python/spark_rapids_ml.html#module-spark_rapids_ml.knn

Otherwise, the CI/automation part looks good to me.

I think we still need the cuml/rapids dependencies to be available/installed, e.g. like the CI/CD dockerfile. Otherwise, the API doc generation doesn't quite work, e.g.

• current docs (manually generated): https://nvidia.github.io/spark-rapids-ml/api/python/spark_rapids_ml.html#module-spark_rapids_ml.knn
• this PR: https://yanxuanliu.github.io/spark-rapids-ml/api/python/spark_rapids_ml.html#module-spark_rapids_ml.knn

Otherwise, the CI/automation part looks good to me.

@YanxuanLiu then I think we could build and run https://github.com/NVIDIA/spark-rapids-ml/blob/branch-23.04/ci/Dockerfile in head of docs steps, could try follow https://stackoverflow.com/a/61155718.
Or we can try find some minimal requirements, and add the installation steps in https://github.com/NVIDIA/spark-rapids-ml/blob/branch-23.04/ci/docs.sh, the goal is to make sure the output doc is like expected one, thanks~

FWIW, I think that the docs need to be built on a GPU host, because sphinx needs to be able to import the various packages to inspect them.

For example, I can successfully import cuml when using the CI docker image w/ --gpus 1, but if I run the same image without a GPU, the import fails with:

>>> import cuml
Traceback (most recent call last):
  File "/root/miniconda3/lib/python3.8/site-packages/cupy/__init__.py", line 18, in <module>
    from cupy import _core  # NOQA
  File "/root/miniconda3/lib/python3.8/site-packages/cupy/_core/__init__.py", line 3, in <module>
    from cupy._core import core  # NOQA
  File "cupy/_core/core.pyx", line 1, in init cupy._core.core
  File "/root/miniconda3/lib/python3.8/site-packages/cupy/cuda/__init__.py", line 8, in <module>
    from cupy.cuda import compiler  # NOQA
  File "/root/miniconda3/lib/python3.8/site-packages/cupy/cuda/compiler.py", line 14, in <module>
    from cupy.cuda import function
  File "cupy/cuda/function.pyx", line 1, in init cupy.cuda.function
  File "cupy/_core/_carray.pyx", line 1, in init cupy._core._carray
  File "cupy/_core/internal.pyx", line 1, in init cupy._core.internal
  File "cupy/cuda/memory.pyx", line 1, in init cupy.cuda.memory
ImportError: libcuda.so.1: cannot open shared object file: No such file or directory

So wondering if we can update gh-pages after a successful "build wheel and deploy" stage in Jenkins (on the same GPU build host)?

Yes, we can run the auto-updating steps on a blossom host, so we can add the steps to release pipeline.