Merge pull request #235226 from natsukium/catboost/update

python3Packages.catboost: 1.0.5 -> 1.2.2; build with cmake

pkgs/development/libraries/catboost/default.nix

@@ -0,0 +1,113 @@
+{ lib
+, config
+, stdenv
+, fetchFromGitHub
+, cmake
+, libiconv
+, llvmPackages
+, ninja
+, openssl
+, python3Packages
+, ragel
+, yasm
+, zlib
+, cudaSupport ? config.cudaSupport
+, cudaPackages ? {}
+, pythonSupport ? false
+}:
+
+stdenv.mkDerivation (finalAttrs: {
+  pname = "catboost";
+  version = "1.2.2";
+
+  src = fetchFromGitHub {
+    owner = "catboost";
+    repo = "catboost";
+    rev = "refs/tags/v${finalAttrs.version}";
+    hash = "sha256-A1zCIqPOW21dHKBQHRtS+/sstZ2o6F8k71lmJFGn0+g=";
+  };
+
+  patches = [
+    ./remove-conan.patch
+  ];
+
+  postPatch = ''
+    substituteInPlace cmake/common.cmake \
+      --replace  "\''${RAGEL_BIN}" "${ragel}/bin/ragel" \
+      --replace "\''${YASM_BIN}" "${yasm}/bin/yasm"
+
+    shopt -s globstar
+    for cmakelists in **/CMakeLists.*; do
+      sed -i "s/OpenSSL::OpenSSL/OpenSSL::SSL/g" $cmakelists
+      ${lib.optionalString (lib.versionOlder cudaPackages.cudaVersion "11.8") ''
+        sed -i 's/-gencode=arch=compute_89,code=sm_89//g' $cmakelists
+        sed -i 's/-gencode=arch=compute_90,code=sm_90//g' $cmakelists
+      ''}
+    done
+  '';
+
+  outputs = [ "out" "dev" ];
+
+  nativeBuildInputs = [
+    cmake
+    llvmPackages.bintools
+    ninja
+    (python3Packages.python.withPackages (ps: with ps; [ six ]))
+    ragel
+    yasm
+  ] ++ lib.optionals cudaSupport (with cudaPackages; [
+    cuda_nvcc
+  ]);
+
+  buildInputs = [
+    openssl
+    zlib
+  ] ++ lib.optionals stdenv.isDarwin [
+    libiconv
+  ] ++ lib.optionals cudaSupport (with cudaPackages; [
+    cuda_cudart
+    cuda_cccl
+    libcublas
+  ]);
+
+  env = {
+    CUDAHOSTCXX = lib.optionalString cudaSupport "${stdenv.cc}/bin/cc";
+    NIX_CFLAGS_LINK = lib.optionalString stdenv.isLinux "-fuse-ld=lld";
+    NIX_LDFLAGS = "-lc -lm";
+  };
+
+  cmakeFlags = [
+    "-DCMAKE_BINARY_DIR=$out"
+    "-DCMAKE_POSITION_INDEPENDENT_CODE=on"
+    "-DCATBOOST_COMPONENTS=app;libs${lib.optionalString pythonSupport ";python-package"}"
+  ] ++ lib.optionals cudaSupport [
+    "-DHAVE_CUDA=on"
+  ];
+
+  installPhase = ''
+    runHook preInstall
+
+    mkdir $dev
+    cp -r catboost $dev
+    install -Dm555 catboost/app/catboost -t $out/bin
+    install -Dm444 catboost/libs/model_interface/static/lib/libmodel_interface-static-lib.a -t $out/lib
+    install -Dm444 catboost/libs/model_interface/libcatboostmodel${stdenv.hostPlatform.extensions.sharedLibrary} -t $out/lib
+    install -Dm444 catboost/libs/train_interface/libcatboost${stdenv.hostPlatform.extensions.sharedLibrary} -t $out/lib
+
+    runHook postInstall
+  '';
+
+  meta = with lib; {
+    description = "High-performance library for gradient boosting on decision trees";
+    longDescription = ''
+      A fast, scalable, high performance Gradient Boosting on Decision Trees
+      library, used for ranking, classification, regression and other machine
+      learning tasks for Python, R, Java, C++. Supports computation on CPU and GPU.
+    '';
+    license = licenses.asl20;
+    platforms = platforms.unix;
+    homepage = "https://catboost.ai";
+    maintainers = with maintainers; [ PlushBeaver natsukium ];
+    mainProgram = "catboost";
+  };
+})

pkgs/development/libraries/catboost/remove-conan.patch

@@ -0,0 +1,34 @@
+diff --git a/CMakeLists.txt b/CMakeLists.txt
+index becd2ad03c..7e3c8c99b1 100644
+--- a/CMakeLists.txt
++++ b/CMakeLists.txt
+@@ -27,7 +27,6 @@ cmake_policy(SET CMP0104 OLD)
+
+ include(cmake/archive.cmake)
+ include(cmake/common.cmake)
+-include(cmake/conan.cmake)
+ include(cmake/cuda.cmake)
+ include(cmake/cython.cmake)
+ include(cmake/fbs.cmake)
+@@ -37,21 +36,6 @@ include(cmake/recursive_library.cmake)
+ include(cmake/swig.cmake)
+ include(cmake/global_vars.cmake)
+
+-if (CMAKE_CROSSCOMPILING)
+-  include(${CMAKE_BINARY_DIR}/conan_paths.cmake)
+-else()
+-  conan_cmake_autodetect(settings)
+-  conan_cmake_install(
+-    PATH_OR_REFERENCE ${CMAKE_SOURCE_DIR}
+-    INSTALL_FOLDER ${CMAKE_BINARY_DIR}
+-    BUILD missing
+-    REMOTE conancenter
+-    SETTINGS ${settings}
+-      ENV "CONAN_CMAKE_GENERATOR=${CMAKE_GENERATOR}"
+-      CONF "tools.cmake.cmaketoolchain:generator=${CMAKE_GENERATOR}"
+-  )
+-endif()
+-
+ if (CMAKE_SYSTEM_NAME STREQUAL "Linux" AND CMAKE_SYSTEM_PROCESSOR STREQUAL "x86_64" AND NOT HAVE_CUDA)
+   include(CMakeLists.linux-x86_64.txt)
+ elseif (CMAKE_SYSTEM_NAME STREQUAL "Linux" AND CMAKE_SYSTEM_PROCESSOR STREQUAL "x86_64" AND HAVE_CUDA)

pkgs/development/python-modules/catboost/default.nix

@@ -1,64 +1,50 @@
-{ buildPythonPackage, fetchFromGitHub, lib, pythonOlder
-, clang_12, python
-, graphviz, matplotlib, numpy, pandas, plotly, scipy, six
-, withCuda ? false, cudatoolkit }:
-
-buildPythonPackage rec {
-  pname = "catboost";
-  # nixpkgs-update: no auto update
-  version = "1.0.5";
-
-  disabled = pythonOlder "3.4";
+{ lib
+, buildPythonPackage
+, catboost
+, python
+, graphviz
+, matplotlib
+, numpy
+, pandas
+, plotly
+, scipy
+, setuptools
+, six
+, wheel
+}:
+
+buildPythonPackage {
+  inherit (catboost) pname version src meta;
+  format = "pyproject";
+
+  sourceRoot = "source/catboost/python-package";
+
+  nativeBuildInputs = [
+    setuptools
+    wheel
+  ];
 
-  src = fetchFromGitHub {
-    owner = "catboost";
-    repo = "catboost";
-    rev = "refs/tags/v${version}";
-    hash = "sha256-ILemeZUBI9jPb9G6F7QX/T1HaVhQ+g6y7YmsT6DFCJk";
-  };
+  propagatedBuildInputs = [
+    graphviz
+    matplotlib
+    numpy
+    pandas
+    plotly
+    scipy
+    six
+  ];
 
-  nativeBuildInputs = [ clang_12 ];
+  buildPhase = ''
+    runHook preBuild
 
-  propagatedBuildInputs = [ graphviz matplotlib numpy pandas scipy plotly six ]
-    ++ lib.optionals withCuda [ cudatoolkit ];
+    # these arguments must set after bdist_wheel
+    ${python.pythonForBuild.interpreter} setup.py bdist_wheel --no-widget --prebuilt-extensions-build-root-dir=${lib.getDev catboost}
 
-  patches = [
-    ./nix-support.patch
-  ];
-
-  postPatch = ''
-    # substituteInPlace is too slow for these large files, and the target has lots of numbers in it that change often.
-    sed -e 's|\$(YMAKE_PYTHON3-.*)/python3|${python.interpreter}|' -i make/*.makefile
+    runHook postBuild
   '';
 
-  preBuild = ''
-    cd catboost/python-package
-  '';
-  setupPyBuildFlags = [ "--with-ymake=no" ];
-  CUDA_ROOT = lib.optional withCuda cudatoolkit;
-  enableParallelBuilding = true;
+  # setup a test is difficult
+  doCheck = false;
 
-  # Tests use custom "ya" tool, not yet supported.
-  dontUseSetuptoolsCheck = true;
   pythonImportsCheck = [ "catboost" ];
-
-  passthru = {
-    # Do not update to catboost 1.1.x because the patch doesn't apply cleanly
-    skipBulkUpdate = true;
-  };
-
-  meta = with lib; {
-    description = "High-performance library for gradient boosting on decision trees.";
-    longDescription = ''
-      A fast, scalable, high performance Gradient Boosting on Decision Trees
-      library, used for ranking, classification, regression and other machine
-      learning tasks for Python, R, Java, C++. Supports computation on CPU and GPU.
-    '';
-    license = licenses.asl20;
-    platforms = [ "x86_64-linux" ];
-    homepage = "https://catboost.ai";
-    maintainers = with maintainers; [ PlushBeaver ];
-    # _catboost.pyx.cpp:226822:19: error: use of undeclared identifier '_PyGen_Send'
-    broken = withCuda;
-  };
 }