Migrate as much as possible to pyproject.toml

.pre-commit-config.yaml

@@ -12,7 +12,7 @@ repos:
       rev: 5.0.4
       hooks:
           - id: flake8
-            args: [--config=python/setup.cfg]
+            args: [--config=python/.flake8]
             files: python/.*$
             types: [file]
             types_or: [python, cython]

ci/release/apply_wheel_modifications.sh

@@ -9,15 +9,12 @@ CUDA_SUFFIX=${2}
 # __init__.py versions
 sed -i "s/__version__ = .*/__version__ = \"${VERSION}\"/g" python/cuml/__init__.py
 
-# setup.py versions
-sed -i "s/version=.*,/version=\"${VERSION}\",/g" python/setup.py
+# pyproject.toml versions
+sed -i "s/^version = .*/version = \"${VERSION}\"/g" python/pyproject.toml
 
-# setup.py cuda suffixes
-sed -i "s/name=\"cuml\"/name=\"cuml${CUDA_SUFFIX}\"/g" python/setup.py
-sed -i "s/cudf/cudf${CUDA_SUFFIX}/g" python/setup.py
-sed -i "s/pylibraft/pylibraft${CUDA_SUFFIX}/g" python/setup.py
-sed -i "s/raft-dask/raft-dask${CUDA_SUFFIX}/g" python/setup.py
-
-# cudf pyproject.toml cuda suffixes
-sed -i "s/rmm/rmm${CUDA_SUFFIX}/g" python/pyproject.toml
+# pyproject.toml cuda suffixes
+sed -i "s/^name = \"cuml\"/name = \"cuml${CUDA_SUFFIX}\"/g" python/pyproject.toml
+sed -i "s/cudf/cudf${CUDA_SUFFIX}/g" python/pyproject.toml
 sed -i "s/pylibraft/pylibraft${CUDA_SUFFIX}/g" python/pyproject.toml
+sed -i "s/raft-dask/raft-dask${CUDA_SUFFIX}/g" python/pyproject.toml
+sed -i "s/rmm/rmm${CUDA_SUFFIX}/g" python/pyproject.toml

ci/release/update-version.sh

@@ -35,18 +35,13 @@ function sed_runner() {
 }
 
 
-# __init__.py and setup.py versions
+# __init__.py and pyproject.toml versions
 sed_runner "s/__version__ = .*/__version__ = \"${NEXT_FULL_TAG}\"/g" python/cuml/__init__.py
-sed_runner "s/version=.*,/version=\"${NEXT_FULL_TAG}\",/g" python/setup.py
-
-# Dependency versions in pyproject.toml
+sed_runner "s/^version = .*/version = \"${NEXT_FULL_TAG}\"/g" python/pyproject.toml
 sed_runner "s/rmm==.*\",/rmm==${NEXT_SHORT_TAG_PEP440}.*\",/g" python/pyproject.toml
+sed_runner "s/cudf==.*\",/cudf==${NEXT_SHORT_TAG_PEP440}.*\",/g" python/pyproject.toml
 sed_runner "s/pylibraft==.*\",/pylibraft==${NEXT_SHORT_TAG_PEP440}.*\",/g" python/pyproject.toml
-
-# Dependency versions in setup.py
-sed_runner "s/cudf==.*\",/cudf==${NEXT_SHORT_TAG_PEP440}.*\",/g" python/setup.py
-sed_runner "s/pylibraft==.*\",/pylibraft==${NEXT_SHORT_TAG_PEP440}.*\",/g" python/setup.py
-sed_runner "s/raft-dask==.*\",/raft-dask==${NEXT_SHORT_TAG_PEP440}.*\",/g" python/setup.py
+sed_runner "s/raft-dask==.*\",/raft-dask==${NEXT_SHORT_TAG_PEP440}.*\",/g" python/pyproject.toml
 
 
 # CMakeLists

python/setup.cfg → python/.flake8

@@ -40,19 +40,3 @@ per-file-ignores =
     # Cython Exclusions
     *.pyx: E999, E225, E226, E227, W503, W504
     *.pxd: E999, E225, E226, E227, W503, W504
-
-[pytest]
-markers =
-  unit: Quickest tests focused on accuracy and correctness
-  quality: More intense tests than unit with increased runtimes
-  stress: Longest running tests focused on stressing hardware compute resources
-  mg: Multi-GPU tests
-  memleak: Test that checks for memory leaks
-  no_bad_cuml_array_check: Test that should not check for bad CumlArray uses
-
-testpaths = cuml/tests
-
-filterwarnings =
-  error::FutureWarning:cuml[.*] # Catch uses of deprecated positional args in testing
-  ignore:[^.]*ABCs[^.]*:DeprecationWarning:patsy[.*]
-  ignore:(.*)alias(.*):DeprecationWarning:hdbscan[.*]

python/MANIFEST.in

@@ -0,0 +1,7 @@
+# Cython files
+recursive-include cuml *.pxd
+recursive-include cuml *.pyx
+
+# Build files. Don't use a recursive include on '.' in case the repo is dirty
+include . CMakeLists.txt
+recursive-include cuml CMakeLists.txt

python/pyproject.toml

@@ -13,7 +13,6 @@
 # limitations under the License.
 
 [build-system]
-
 requires = [
     "wheel",
     "setuptools",
@@ -28,6 +27,78 @@ requires = [
 ]
 build-backend = "setuptools.build_meta"
 
+[tool.pytest.ini_options]
+markers = [
+  "unit: Quickest tests focused on accuracy and correctness",
+  "quality: More intense tests than unit with increased runtimes",
+  "stress: Longest running tests focused on stressing hardware compute resources",
+  "mg: Multi-GPU tests",
+  "memleak: Test that checks for memory leaks",
+  "no_bad_cuml_array_check: Test that should not check for bad CumlArray uses",
+]
+
+testpaths = "cuml/tests"
+
+filterwarnings = [
+  "error::FutureWarning:cuml[.*]", # Catch uses of deprecated positional args in testing
+  "ignore:[^.]*ABCs[^.]*:DeprecationWarning:patsy[.*]",
+  "ignore:(.*)alias(.*):DeprecationWarning:hdbscan[.*]",
+]
+
+[project]
+name = "cuml"
+version = "23.04.00"
+description = "cuML - RAPIDS ML Algorithms"
+readme = { file = "README.md", content-type = "text/markdown" }
+authors = [
+    { name = "NVIDIA Corporation" },
+]
+license = { text = "Apache 2.0" }
+requires-python = ">=3.8"
+dependencies = [
+    "numba",
+    "scipy",
+    "seaborn",
+    "treelite==3.1.0",
+    "treelite_runtime==3.1.0",
+    "cudf==23.4.*",
+    "dask-cudf==23.4.*",
+    "pylibraft==23.4.*",
+    "raft-dask==23.4.*",
+]
+classifiers = [
+    "Intended Audience :: Developers",
+    "Programming Language :: Python",
+    "Programming Language :: Python :: 3.8",
+    "Programming Language :: Python :: 3.9",
+    "Programming Language :: Python :: 3.10",
+]
+
+[project.optional-dependencies]
+test = [
+    "pytest",
+    "hypothesis",
+    "pytest-xdist",
+    "pytest-benchmark",
+    "pytest-cases",
+    "nltk",
+    "dask-ml",
+    "numpydoc",
+    "umap-learn",
+    "statsmodels",
+    "scikit-learn==1.2",
+    "hdbscan @ git+https://github.com/scikit-learn-contrib/hdbscan.git@master",  # noqa:E501
+    "dask-glm @ git+https://github.com/dask/dask-glm@main",
+    "dask-cuda",
+]
+
+[project.urls]
+Homepage = "https://github.com/rapidsai/cuml"
+Documentation = "https://docs.rapids.ai/api/cuml/stable/"
+
+[tool.setuptools]
+license-files = ["LICENSE"]
+
 [tool.black]
 line-length = 79
 target-version = ["py38"]

python/setup.py

@@ -107,52 +107,7 @@ def clean_folder(path):
 # - Python package generation ------------------------------------------------
 
 setup(
-    name="cuml",
-    version="23.04.00",
-    description="cuML - RAPIDS ML Algorithms",
-    url="https://github.com/rapidsai/cuml",
-    author="NVIDIA Corporation",
-    license="Apache 2.0",
-    classifiers=[
-        "Intended Audience :: Developers",
-        "Programming Language :: Python",
-        "Programming Language :: Python :: 3.8",
-        "Programming Language :: Python :: 3.9",
-        "Programming Language :: Python :: 3.10",
-    ],
     include_package_data=True,
     packages=find_packages(include=["cuml", "cuml.*"]),
-    package_data={
-        key: ["*.pxd"] for key in find_packages(include=["cuml", "cuml.*"])
-    },
-    install_requires=[
-        "numba",
-        "scipy",
-        "seaborn",
-        "treelite==3.1.0",
-        "treelite_runtime==3.1.0",
-        "cudf==23.4.*",
-        "dask-cudf==23.4.*",
-        "pylibraft==23.4.*",
-        "raft-dask==23.4.*",
-    ],
-    extras_require={
-        "test": [
-            "pytest",
-            "hypothesis",
-            "pytest-xdist",
-            "pytest-benchmark",
-            "pytest-cases",
-            "nltk",
-            "dask-ml",
-            "numpydoc",
-            "umap-learn",
-            "statsmodels",
-            "scikit-learn==1.2",
-            "hdbscan @ git+https://github.com/scikit-learn-contrib/hdbscan.git@master",  # noqa:E501
-            "dask-glm @ git+https://github.com/dask/dask-glm@main",
-            "dask-cuda",
-        ]
-    },
     zip_safe=False,
 )