v23.04.00

🚨 Breaking Changes

• Pin dask and distributed for release (#5333) @galipremsagar

🐛 Bug Fixes

• Skip pickle notebook during nbsphinx (#5342) @dantegd
• Avoid race condition in FIL predict_per_tree (#5334) @wphicks
• Ensure experimental FIL shmem usage is below device limits (#5326) @wphicks
• Update cuda architectures for threads per sm restriction (#5323) @wphicks
• Run experimental FIL tests in CI (#5316) @wphicks
• Run memory leak pytests without parallelism to avoid sporadic test failures (#5313) @dantegd
• Update cupy version for pip wheels (#5311) @dantegd
• Fix for raising attributeerors erroneously for ipython methods (#5299) @dantegd
• Fix cuml local cpp docs build (#5297) @galipremsagar
• Don't run dask tests twice when testing wheels (#5279) @benfred
• Remove MANIFEST.in use auto-generated one for sdists and package_data for wheels (#5278) @vyasr
• Removing remaining include of raft/distance/distance_type.hpp (#5264) @cjnolet
• Enable hypothesis testing for nightly test runs. (#5244) @csadorf
• Support numeric, boolean, and string keyword arguments to class methods during CPU dispatching (#5236) @beckernick
• Allowing large data in kmeans (#5228) @cjnolet

📖 Documentation

• Fix docs build to be pydata-sphinx-theme=0.13.0 compatible (#5259) @galipremsagar
• Add supported CPU/GPU operators to API docs and update docstrings (#5239) @beckernick
• Fix documentation author (#5126) @bdice

🚀 New Features

• Modify default batch size in HDBSCAN soft clustering (#5335) @tarang-jain
• reduce memory pressure in membership vector computation (#5268) @tarang-jain
• membership_vector for HDBSCAN (#5247) @tarang-jain
• Provide FIL implementation for both CPU and GPU (#4890) @wphicks

🛠️ Improvements

• Remove deprecated Treelite CI API from FIL (#5348) @hcho3
• Updated forest inference to new dask worker api for 23.04 (#5347) @taureandyernv
• Pin dask and distributed for release (#5333) @galipremsagar
• Pin cupy in wheel tests to supported versions (#5312) @vyasr
• Drop pickle5 (#5310) @jakirkham
• Remove CUDA_CHECK macro (#5308) @hcho3
• Revert faiss removal pinned tag (#5306) @cjnolet
• Upgrade to Treelite 3.2.0 (#5304) @hcho3
• Implement predict_per_tree() in FIL (#5303) @hcho3
• remove faiss from cuml (#5293) @benfred
• Stop setting package version attribute in wheels (#5285) @vyasr
• Add libfaiss runtime dependency to libcuml. (#5284) @bdice
• Move faiss_mr from raft (#5281) @benfred
• Generate pyproject dependencies with dfg (#5275) @vyasr
• Updating cuML to use consolidated RAFT libs (#5272) @cjnolet
• Add codespell as a linter (#5265) @benfred
• Pass AWS_SESSION_TOKEN and SCCACHE_S3_USE_SSL vars to conda build (#5263) @ajschmidt8
• Update to GCC 11 (#5258) @bdice
• Drop Python 3.7 handling for pickle protocol 4 (#5256) @jakirkham
• Migrate as much as possible to pyproject.toml (#5251) @vyasr
• Adapt to rapidsai/rmm#1221 which moves allocator callbacks (#5249) @wence-
• Add dfg as a pre-commit hook. (#5246) @vyasr
• Stop using versioneer to manage versions (#5245) @vyasr
• Enhance cuML benchmark utility and refactor hdbscan import utilities (#5242) @beckernick
• Fix GHA build workflow (#5241) @AjayThorve
• Support innerproduct distance in the pairwise_distance API (#5230) @benfred
• Enable hypothesis for 23.04 (#5221) @csadorf
• Reduce error handling verbosity in CI tests scripts (#5219) @AjayThorve
• Bump pinned pip wheel deps to 23.4 (#5217) @sevagh
• Update shared workflow branches (#5215) @ajschmidt8
• Unpin dask and distributed for development (#5209) @galipremsagar
• Remove gpuCI scripts. (#5208) @bdice
• Move date to build string in conda recipe (#5190) @ajschmidt8
• Kernel shap improvements (#5187) @vinaydes
• test out the raft bfknn replacement (#5186) @benfred
• Forward merge 23.02 into 23.04 (#5182) @vyasr
• Add detail namespace for linear models (#5107) @lowener
• Add pre-commit configuration (#4983) @csadorf