Fix HDBSCAN python namespace

[cjnolet]

Addresses #4724.

[cjnolet]

Hey @vyasr, quick question here- I'm trying to embed the hdbscan cython in a directory within cuml/cluster but scikit-build doesn't seem to like it (and I'm sure I'm not doing it right). For example, CI is giving me this error:

15:45:48 CMake Error at /workspace/.conda-bld/_h_env_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placeh/lib/python3.8/site-packages/skbuild/resources/cmake/UseCython.cmake:213 (get_directory_property):
15:45:48   get_directory_property DIRECTORY argument provided but requested directory
15:45:48   not found.  This could be because the directory argument was invalid or, it
15:45:48   is valid but has not been processed yet.
15:45:48 Call Stack (most recent call first):
15:45:48   _skbuild/linux-aarch64-3.8/cmake-build/_deps/rapids-cmake-src/rapids-cmake/cython/create_modules.cmake:99 (add_cython_target)
15:45:48   cuml/cluster/CMakeLists.txt:30 (rapids_cython_create_modules)
15:45:48 
15:45:48 
15:45:48 _modinit_prefix:PyInit_
15:45:48 CMake Error at /workspace/.conda-bld/_h_env_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placeh/lib/python3.8/site-packages/skbuild/resources/cmake/UseCython.cmake:213 (get_directory_property):
15:45:48   get_directory_property DIRECTORY argument provided but requested directory
15:45:48   not found.  This could be because the directory argument was invalid or, it
15:45:48   is valid but has not been processed yet.
15:45:48 Call Stack (most recent call first):
15:45:48   _skbuild/linux-aarch64-3.8/cmake-build/_deps/rapids-cmake-src/rapids-cmake/cython/create_modules.cmake:99 (add_cython_target)
15:45:48   cuml/cluster/CMakeLists.txt:30 (rapids_cython_create_modules)

Here's the offending line: https://github.com/rapidsai/cuml/pull/4895/files#diff-dc75864989b6c5d0c5ab6c45578b8e678d325a9532758d60afc5651edfc87e45R18. Do you have any ideas on what I'm doing wrong?

[vyasr]

It's this scikit-build issue that prevents us from having nested directories. You'll need to create another CMakeLists.txt in the subdirectory and add it here.

[codecov-commenter]

Codecov Report

Base: 78.02% // Head: 79.65% // Increases project coverage by +1.62% 🎉

Coverage data is based on head (8eb8fd2) compared to base (7a0ab85).
Patch has no changes to coverable lines.

Additional details and impacted files

@@               Coverage Diff                @@
##           branch-22.10    #4895      +/-   ##
================================================
+ Coverage         78.02%   79.65%   +1.62%     
================================================
  Files               180      181       +1     
  Lines             11385    11489     +104     
================================================
+ Hits               8883     9151     +268     
+ Misses             2502     2338     -164     

Flag	Coverage Δ
dask	46.16% <ø> (-0.06%)	⬇️
non-dask	68.99% <ø> (+1.72%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
python/cuml/common/array.py	95.10% <0.00%> (-2.88%)	⬇️
python/cuml/cluster/__init__.py	100.00% <0.00%> (ø)
python/cuml/metrics/__init__.py	100.00% <0.00%> (ø)
python/cuml/preprocessing/__init__.py	100.00% <0.00%> (ø)
python/cuml/benchmark/automated/utils/utils.py	0.00% <0.00%> (ø)
python/cuml/dask/preprocessing/LabelEncoder.py	93.33% <0.00%> (ø)
...cuml/_thirdparty/sklearn/preprocessing/__init__.py	100.00% <0.00%> (ø)
python/cuml/cluster/hdbscan/__init__.py	100.00% <0.00%> (ø)
python/cuml/thirdparty_adapters/adapters.py	91.54% <0.00%> (+0.05%)	⬆️
...cuml/_thirdparty/sklearn/utils/skl_dependencies.py	57.64% <0.00%> (+0.50%)	⬆️
... and 4 more

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

☔ View full report at Codecov.
📢 Do you have feedback about the report comment? Let us know in this issue.

[beckernick]

I think this will address the namespace issue mentioned in #4724 but won't fully close the issue, as the issue is also for hdbscan.membership_vector

[dantegd]

@gpucibot merge

[ajschmidt8]

Approving ops-codeowner file changes