merge

rapidsai · Nov 6, 2024 · ac596a2 · ac596a2
2 parents 51a382e + ac5b3ed
commit ac596a2
Show file tree

Hide file tree

Showing 588 changed files with 9,582 additions and 5,137 deletions.
diff --git a/.github/workflows/auto-assign.yml b/.github/workflows/auto-assign.yml
@@ -0,0 +1,17 @@
+name: "Auto Assign PR"
+
+on:
+  pull_request_target:
+    types:
+      - opened
+      - reopened
+      - synchronize
+
+jobs:
+  add_assignees:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions-ecosystem/action-add-assignees@v1
+        with:
+          github_token: "${{ secrets.GITHUB_TOKEN }}"
+          assignees: ${{ github.actor }}
diff --git a/.github/workflows/labeler.yml b/.github/workflows/labeler.yml
@@ -1,4 +1,5 @@
 name: "Pull Request Labeler"
+
 on:
 - pull_request_target
 

diff --git a/.pre-commit-config.yaml b/.pre-commit-config.yaml
@@ -16,17 +16,6 @@ repos:
             ^cpp/cmake/thirdparty/patches/.*|
             ^python/cudf/cudf/tests/data/subword_tokenizer_data/.*
           )
-  - repo: https://github.com/PyCQA/isort
-    rev: 5.13.2
-    hooks:
-      - id: isort
-        # Use the config file specific to each subproject so that each
-        # project can specify its own first/third-party packages.
-        args: ["--config-root=python/", "--resolve-all-configs"]
-        files: python/.*
-        exclude: |
-          (?x)^(^python/cudf_polars/.*)
-        types_or: [python, cython, pyi]
   - repo: https://github.com/MarcoGorelli/cython-lint
     rev: v0.16.2
     hooks:
@@ -150,6 +139,7 @@ repos:
     rev: v0.4.8
     hooks:
       - id: ruff
+        args: ["--fix"]
         files: python/.*$
       - id: ruff-format
         files: python/.*$
@@ -165,7 +155,7 @@ repos:
           )
       - id: verify-alpha-spec
   - repo: https://github.com/rapidsai/dependency-file-generator
-    rev: v1.13.11
+    rev: v1.16.0
     hooks:
       - id: rapids-dependency-file-generator
         args: ["--clean"]

diff --git a/CONTRIBUTING.md b/CONTRIBUTING.md
@@ -38,6 +38,7 @@ conduct. More information can be found at:
 8. Verify that CI passes all [status checks](https://docs.github.com/en/pull-requests/collaborating-with-pull-requests/collaborating-on-repositories-with-code-quality-features/about-status-checks).
    Fix if needed.
 9. Wait for other developers to review your code and update code as needed.
+   Changes to any C++ files require at least 2 approvals from the cudf-cpp-codeowners before merging.
 10. Once reviewed and approved, a RAPIDS developer will merge your pull request.
 
 If you are unsure about anything, don't hesitate to comment on issues and ask for clarification!
@@ -293,8 +294,8 @@ In order to run doxygen as a linter on C++/CUDA code, run
 ./ci/checks/doxygen.sh
 ```
 
-Python code runs several linters including [Black](https://black.readthedocs.io/en/stable/),
-[isort](https://pycqa.github.io/isort/), and [flake8](https://flake8.pycqa.org/en/latest/).
+Python code runs several linters including [Ruff](https://docs.astral.sh/ruff/)
+with its various rules  for Black-like formatting or Isort.
 
 cuDF also uses [codespell](https://github.com/codespell-project/codespell) to find spelling
 mistakes, and this check is run as a pre-commit hook. To apply the suggested spelling fixes,

diff --git a/ci/build_cpp.sh b/ci/build_cpp.sh
@@ -15,8 +15,12 @@ rapids-print-env
 
 rapids-logger "Begin cpp build"
 
+sccache --zero-stats
+
 # With boa installed conda build forward to boa
 RAPIDS_PACKAGE_VERSION=$(rapids-generate-version) rapids-conda-retry mambabuild \
     conda/recipes/libcudf
 
+sccache --show-adv-stats
+
 rapids-upload-conda-to-s3 cpp
diff --git a/ci/build_python.sh b/ci/build_python.sh
@@ -19,6 +19,8 @@ rapids-logger "Begin py build"
 
 CPP_CHANNEL=$(rapids-download-conda-from-s3 cpp)
 
+sccache --zero-stats
+
 # TODO: Remove `--no-test` flag once importing on a CPU
 # node works correctly
 # With boa installed conda build forwards to the boa builder
@@ -28,12 +30,18 @@ RAPIDS_PACKAGE_VERSION=$(head -1 ./VERSION) rapids-conda-retry mambabuild \
   --channel "${CPP_CHANNEL}" \
   conda/recipes/pylibcudf
 
+sccache --show-adv-stats
+sccache --zero-stats
+
 RAPIDS_PACKAGE_VERSION=$(head -1 ./VERSION) rapids-conda-retry mambabuild \
   --no-test \
   --channel "${CPP_CHANNEL}" \
   --channel "${RAPIDS_CONDA_BLD_OUTPUT_DIR}" \
   conda/recipes/cudf
 
+sccache --show-adv-stats
+sccache --zero-stats
+
 RAPIDS_PACKAGE_VERSION=$(head -1 ./VERSION) rapids-conda-retry mambabuild \
   --no-test \
   --channel "${CPP_CHANNEL}" \
@@ -46,6 +54,8 @@ RAPIDS_PACKAGE_VERSION=$(head -1 ./VERSION) rapids-conda-retry mambabuild \
   --channel "${RAPIDS_CONDA_BLD_OUTPUT_DIR}" \
   conda/recipes/cudf_kafka
 
+sccache --show-adv-stats
+
 RAPIDS_PACKAGE_VERSION=$(head -1 ./VERSION) rapids-conda-retry mambabuild \
   --no-test \
   --channel "${CPP_CHANNEL}" \

diff --git a/ci/build_wheel.sh b/ci/build_wheel.sh
@@ -3,7 +3,8 @@
 
 set -euo pipefail
 
-package_dir=$1
+package_name=$1
+package_dir=$2
 
 source rapids-configure-sccache
 source rapids-date-string
@@ -12,4 +13,14 @@ rapids-generate-version > ./VERSION
 
 cd "${package_dir}"
 
-python -m pip wheel . -w dist -v --no-deps --disable-pip-version-check
+sccache --zero-stats
+
+rapids-logger "Building '${package_name}' wheel"
+python -m pip wheel \
+    -w dist \
+    -v \
+    --no-deps \
+    --disable-pip-version-check \
+    .
+
+sccache --show-adv-stats
diff --git a/ci/build_wheel_cudf.sh b/ci/build_wheel_cudf.sh
@@ -18,7 +18,7 @@ echo "libcudf-${RAPIDS_PY_CUDA_SUFFIX} @ file://$(echo /tmp/libcudf_dist/libcudf
 echo "pylibcudf-${RAPIDS_PY_CUDA_SUFFIX} @ file://$(echo /tmp/pylibcudf_dist/pylibcudf_*.whl)" >> /tmp/constraints.txt
 export PIP_CONSTRAINT="/tmp/constraints.txt"
 
-./ci/build_wheel.sh ${package_dir}
+./ci/build_wheel.sh cudf ${package_dir}
 
 python -m auditwheel repair \
     --exclude libcudf.so \

diff --git a/ci/build_wheel_cudf_polars.sh b/ci/build_wheel_cudf_polars.sh
@@ -5,7 +5,7 @@ set -euo pipefail
 
 package_dir="python/cudf_polars"
 
-./ci/build_wheel.sh ${package_dir}
+./ci/build_wheel.sh cudf-polars ${package_dir}
 
 RAPIDS_PY_CUDA_SUFFIX="$(rapids-wheel-ctk-name-gen ${RAPIDS_CUDA_VERSION})"
-RAPIDS_PY_WHEEL_NAME="cudf_polars_${RAPIDS_PY_CUDA_SUFFIX}" RAPIDS_PY_WHEEL_PURE="1" rapids-upload-wheels-to-s3 ${package_dir}/dist
+RAPIDS_PY_WHEEL_NAME="cudf_polars_${RAPIDS_PY_CUDA_SUFFIX}" RAPIDS_PY_WHEEL_PURE="1" rapids-upload-wheels-to-s3 python ${package_dir}/dist
diff --git a/ci/build_wheel_dask_cudf.sh b/ci/build_wheel_dask_cudf.sh
@@ -5,7 +5,7 @@ set -euo pipefail
 
 package_dir="python/dask_cudf"
 
-./ci/build_wheel.sh ${package_dir}
+./ci/build_wheel.sh dask-cudf ${package_dir}
 
 RAPIDS_PY_CUDA_SUFFIX="$(rapids-wheel-ctk-name-gen ${RAPIDS_CUDA_VERSION})"
-RAPIDS_PY_WHEEL_NAME="dask_cudf_${RAPIDS_PY_CUDA_SUFFIX}" RAPIDS_PY_WHEEL_PURE="1" rapids-upload-wheels-to-s3 ${package_dir}/dist
+RAPIDS_PY_WHEEL_NAME="dask_cudf_${RAPIDS_PY_CUDA_SUFFIX}" RAPIDS_PY_WHEEL_PURE="1" rapids-upload-wheels-to-s3 python ${package_dir}/dist
diff --git a/ci/build_wheel_libcudf.sh b/ci/build_wheel_libcudf.sh
@@ -3,10 +3,30 @@
 
 set -euo pipefail
 
+package_name="libcudf"
 package_dir="python/libcudf"
 
+rapids-logger "Generating build requirements"
+
+rapids-dependency-file-generator \
+  --output requirements \
+  --file-key "py_build_${package_name}" \
+  --file-key "py_rapids_build_${package_name}" \
+  --matrix "cuda=${RAPIDS_CUDA_VERSION%.*};arch=$(arch);py=${RAPIDS_PY_VERSION};cuda_suffixed=true" \
+| tee /tmp/requirements-build.txt
+
+rapids-logger "Installing build requirements"
+python -m pip install \
+    -v \
+    --prefer-binary \
+    -r /tmp/requirements-build.txt
+
+# build with '--no-build-isolation', for better sccache hit rate
+# 0 really means "add --no-build-isolation" (ref: https://github.com/pypa/pip/issues/5735)
+export PIP_NO_BUILD_ISOLATION=0
+
 export SKBUILD_CMAKE_ARGS="-DUSE_NVCOMP_RUNTIME_WHEEL=ON"
-./ci/build_wheel.sh ${package_dir}
+./ci/build_wheel.sh "${package_name}" "${package_dir}"
 
 RAPIDS_PY_CUDA_SUFFIX="$(rapids-wheel-ctk-name-gen ${RAPIDS_CUDA_VERSION})"
 
@@ -16,4 +36,4 @@ python -m auditwheel repair \
     -w ${package_dir}/final_dist \
     ${package_dir}/dist/*
 
-RAPIDS_PY_WHEEL_NAME="libcudf_${RAPIDS_PY_CUDA_SUFFIX}" rapids-upload-wheels-to-s3 cpp ${package_dir}/final_dist
+RAPIDS_PY_WHEEL_NAME="${package_name}_${RAPIDS_PY_CUDA_SUFFIX}" rapids-upload-wheels-to-s3 cpp "${package_dir}/final_dist"
diff --git a/ci/build_wheel_pylibcudf.sh b/ci/build_wheel_pylibcudf.sh
@@ -16,12 +16,12 @@ RAPIDS_PY_WHEEL_NAME="libcudf_${RAPIDS_PY_CUDA_SUFFIX}" rapids-download-wheels-f
 echo "libcudf-${RAPIDS_PY_CUDA_SUFFIX} @ file://$(echo /tmp/libcudf_dist/libcudf_*.whl)" > /tmp/constraints.txt
 export PIP_CONSTRAINT="/tmp/constraints.txt"
 
-./ci/build_wheel.sh ${package_dir}
+./ci/build_wheel.sh pylibcudf ${package_dir}
 
 python -m auditwheel repair \
     --exclude libcudf.so \
     --exclude libnvcomp.so \
     -w ${package_dir}/final_dist \
     ${package_dir}/dist/*
 
-RAPIDS_PY_WHEEL_NAME="pylibcudf_${RAPIDS_PY_CUDA_SUFFIX}" rapids-upload-wheels-to-s3 ${package_dir}/final_dist
+RAPIDS_PY_WHEEL_NAME="pylibcudf_${RAPIDS_PY_CUDA_SUFFIX}" rapids-upload-wheels-to-s3 python ${package_dir}/final_dist
diff --git a/ci/cudf_pandas_scripts/run_tests.sh b/ci/cudf_pandas_scripts/run_tests.sh
@@ -54,15 +54,8 @@ else
     RAPIDS_PY_WHEEL_NAME="libcudf_${RAPIDS_PY_CUDA_SUFFIX}" rapids-download-wheels-from-s3 cpp ./dist
     RAPIDS_PY_WHEEL_NAME="pylibcudf_${RAPIDS_PY_CUDA_SUFFIX}" rapids-download-wheels-from-s3 python ./dist
 
-    echo "" > ./constraints.txt
-    if [[ $RAPIDS_DEPENDENCIES == "oldest" ]]; then
-        # `test_python_cudf_pandas` constraints are for `[test]` not `[cudf-pandas-tests]`
-        rapids-dependency-file-generator \
-            --output requirements \
-            --file-key test_python_cudf_pandas \
-            --matrix "cuda=${RAPIDS_CUDA_VERSION%.*};arch=$(arch);py=${RAPIDS_PY_VERSION};dependencies=${RAPIDS_DEPENDENCIES}" \
-        | tee ./constraints.txt
-    fi
+    # generate constraints (possibly pinning to oldest support versions of dependencies)
+    rapids-generate-pip-constraints test_python_cudf_pandas ./constraints.txt
 
     python -m pip install \
         -v \

diff --git a/ci/test_wheel_cudf.sh b/ci/test_wheel_cudf.sh
@@ -12,15 +12,8 @@ RAPIDS_PY_WHEEL_NAME="pylibcudf_${RAPIDS_PY_CUDA_SUFFIX}" rapids-download-wheels
 
 rapids-logger "Install cudf, pylibcudf, and test requirements"
 
-# Constrain to minimum dependency versions if job is set up as "oldest"
-echo "" > ./constraints.txt
-if [[ $RAPIDS_DEPENDENCIES == "oldest" ]]; then
-    rapids-dependency-file-generator \
-        --output requirements \
-        --file-key py_test_cudf \
-        --matrix "cuda=${RAPIDS_CUDA_VERSION%.*};arch=$(arch);py=${RAPIDS_PY_VERSION};dependencies=${RAPIDS_DEPENDENCIES}" \
-      | tee ./constraints.txt
-fi
+# generate constraints (possibly pinning to oldest support versions of dependencies)
+rapids-generate-pip-constraints py_test_cudf ./constraints.txt
 
 # echo to expand wildcard before adding `[extra]` requires for pip
 python -m pip install \

diff --git a/ci/test_wheel_cudf_polars.sh b/ci/test_wheel_cudf_polars.sh
@@ -29,15 +29,9 @@ RAPIDS_PY_WHEEL_NAME="libcudf_${RAPIDS_PY_CUDA_SUFFIX}" rapids-download-wheels-f
 RAPIDS_PY_WHEEL_NAME="pylibcudf_${RAPIDS_PY_CUDA_SUFFIX}" rapids-download-wheels-from-s3 python ./dist
 
 rapids-logger "Installing cudf_polars and its dependencies"
-# Constraint to minimum dependency versions if job is set up as "oldest"
-echo "" > ./constraints.txt
-if [[ $RAPIDS_DEPENDENCIES == "oldest" ]]; then
-    rapids-dependency-file-generator \
-        --output requirements \
-        --file-key py_test_cudf_polars \
-        --matrix "cuda=${RAPIDS_CUDA_VERSION%.*};arch=$(arch);py=${RAPIDS_PY_VERSION};dependencies=${RAPIDS_DEPENDENCIES}" \
-      | tee ./constraints.txt
-fi
+
+# generate constraints (possibly pinning to oldest support versions of dependencies)
+rapids-generate-pip-constraints py_test_cudf_polars ./constraints.txt
 
 # echo to expand wildcard before adding `[test]` requires for pip
 python -m pip install \

diff --git a/ci/test_wheel_dask_cudf.sh b/ci/test_wheel_dask_cudf.sh
@@ -12,15 +12,9 @@ RAPIDS_PY_WHEEL_NAME="libcudf_${RAPIDS_PY_CUDA_SUFFIX}" rapids-download-wheels-f
 RAPIDS_PY_WHEEL_NAME="pylibcudf_${RAPIDS_PY_CUDA_SUFFIX}" rapids-download-wheels-from-s3 python ./dist
 
 rapids-logger "Install dask_cudf, cudf, pylibcudf, and test requirements"
-# Constraint to minimum dependency versions if job is set up as "oldest"
-echo "" > ./constraints.txt
-if [[ $RAPIDS_DEPENDENCIES == "oldest" ]]; then
-    rapids-dependency-file-generator \
-        --output requirements \
-        --file-key py_test_dask_cudf \
-        --matrix "cuda=${RAPIDS_CUDA_VERSION%.*};arch=$(arch);py=${RAPIDS_PY_VERSION};dependencies=${RAPIDS_DEPENDENCIES}" \
-      | tee ./constraints.txt
-fi
+
+# generate constraints (possibly pinning to oldest support versions of dependencies)
+rapids-generate-pip-constraints py_test_dask_cudf ./constraints.txt
 
 # echo to expand wildcard before adding `[extra]` requires for pip
 python -m pip install \

diff --git a/conda/environments/all_cuda-118_arch-x86_64.yaml b/conda/environments/all_cuda-118_arch-x86_64.yaml
@@ -46,6 +46,7 @@ dependencies:
 - librdkafka>=2.5.0,<2.6.0a0
 - librmm==24.12.*,>=0.0.0a0
 - make
+- mmh3
 - moto>=4.0.8
 - msgpack-python
 - myst-nb
@@ -58,14 +59,14 @@ dependencies:
 - numpy>=1.23,<3.0a0
 - numpydoc
 - nvcc_linux-64=11.8
-- nvcomp==4.0.1
+- nvcomp==4.1.0.6
 - nvtx>=0.2.1
 - openpyxl
 - packaging
 - pandas
 - pandas>=2.0,<2.2.4dev0
 - pandoc
-- polars>=1.11,<1.12
+- polars>=1.11,<1.13
 - pre-commit
 - ptxcompiler
 - pyarrow>=14.0.0,<18.0.0a0
@@ -76,6 +77,7 @@ dependencies:
 - pytest-xdist
 - pytest<8
 - python-confluent-kafka>=2.5.0,<2.6.0a0
+- python-xxhash
 - python>=3.10,<3.13
 - pytorch>=2.1.0
 - rapids-build-backend>=0.3.0,<0.4.0.dev0

diff --git a/conda/environments/all_cuda-125_arch-x86_64.yaml b/conda/environments/all_cuda-125_arch-x86_64.yaml
@@ -45,6 +45,7 @@ dependencies:
 - librdkafka>=2.5.0,<2.6.0a0
 - librmm==24.12.*,>=0.0.0a0
 - make
+- mmh3
 - moto>=4.0.8
 - msgpack-python
 - myst-nb
@@ -56,14 +57,14 @@ dependencies:
 - numba-cuda>=0.0.13
 - numpy>=1.23,<3.0a0
 - numpydoc
-- nvcomp==4.0.1
+- nvcomp==4.1.0.6
 - nvtx>=0.2.1
 - openpyxl
 - packaging
 - pandas
 - pandas>=2.0,<2.2.4dev0
 - pandoc
-- polars>=1.11,<1.12
+- polars>=1.11,<1.13
 - pre-commit
 - pyarrow>=14.0.0,<18.0.0a0
 - pydata-sphinx-theme!=0.14.2
@@ -74,6 +75,7 @@ dependencies:
 - pytest-xdist
 - pytest<8
 - python-confluent-kafka>=2.5.0,<2.6.0a0
+- python-xxhash
 - python>=3.10,<3.13
 - pytorch>=2.1.0
 - rapids-build-backend>=0.3.0,<0.4.0.dev0

diff --git a/conda/recipes/libcudf/conda_build_config.yaml b/conda/recipes/libcudf/conda_build_config.yaml
@@ -35,7 +35,7 @@ spdlog_version:
   - ">=1.14.1,<1.15"
 
 nvcomp_version:
-  - "=4.0.1"
+  - "=4.1.0.6"
 
 zlib_version:
   - ">=1.2.13"

diff --git a/cpp/CMakeLists.txt b/cpp/CMakeLists.txt
@@ -369,6 +369,9 @@ add_library(
   src/filling/sequence.cu
   src/groupby/groupby.cu
   src/groupby/hash/compute_groupby.cu
+  src/groupby/hash/compute_mapping_indices.cu
+  src/groupby/hash/compute_mapping_indices_null.cu
+  src/groupby/hash/compute_shared_memory_aggs.cu
   src/groupby/hash/compute_single_pass_aggs.cu
   src/groupby/hash/create_sparse_results_table.cu
   src/groupby/hash/flatten_single_pass_aggs.cpp