New port: misc/xgboost: Scalable, Portable and Distributed Gradient Boosting (GBDT, GBRT, GBM)

svn path=/head/; revision=520144

misc/Makefile

@@ -535,6 +535,7 @@
     SUBDIR += xfce4-weather-plugin
     SUBDIR += xfce4-wm-themes
     SUBDIR += xgas
+    SUBDIR += xgboost
     SUBDIR += xless
     SUBDIR += xosd
     SUBDIR += xtail

misc/xgboost/Makefile

@@ -0,0 +1,35 @@
+# $FreeBSD$
+
+PORTNAME=	xgboost
+DISTVERSIONPREFIX=	v
+DISTVERSION=	0.90-235
+DISTVERSIONSUFFIX=	-gad4a1c73
+CATEGORIES=	misc # machine-learning
+
+MAINTAINER=	yuri@FreeBSD.org
+COMMENT=	Scalable, Portable and Distributed Gradient Boosting (GBDT, GBRT, GBM)
+
+LICENSE=	APACHE20
+LICENSE_FILE=	${WRKSRC}/LICENSE
+
+LIB_DEPENDS=	libdmlc.so:devel/dmlc-core
+
+USES=		cmake compiler:c++11-lang localbase
+USE_GITHUB=	yes
+GH_ACCOUNT=	dmlc
+GH_TUPLE=	dmlc:rabit:2f25347:rabit/rabit \
+		NVlabs:cub:b20808b:cub/cub
+USE_LDCONFIG=	yes
+
+LDFLAGS+=	${LOCALBASE}/lib/libdmlc.so -lexecinfo
+
+OPTIONS_DEFINE=		OPENMP
+
+OPENMP_CMAKE_BOOL=	USE_OPENMP
+OPENMP_LIB_DEPENDS=	libomp.so:devel/openmp
+OPENMP_BROKEN=		error: unsupported argument 'libomp  -I/usr/local/include -L/usr/local/lib' to option 'fopenmp='
+
+post-patch: # https://github.com/dmlc/rabit/issues/131
+	@${FIND} ${WRKSRC} -name "*.h" -o -name "*.cc" | ${XARGS} ${REINPLACE_CMD} -i '' -e 's| _assert(| _xassert(|; s| _assert)| _xassert)|'
+
+.include <bsd.port.mk>

misc/xgboost/distinfo

@@ -0,0 +1,7 @@
+TIMESTAMP = 1576356559
+SHA256 (dmlc-xgboost-v0.90-235-gad4a1c73_GH0.tar.gz) = c5fd4f9c152ba56961a43b677397251f20f78e534571926247f8d24275f9642b
+SIZE (dmlc-xgboost-v0.90-235-gad4a1c73_GH0.tar.gz) = 1131611
+SHA256 (dmlc-rabit-2f25347_GH0.tar.gz) = 9a9f9b754d1f8f8f0949faded75859e791cc56c75c38928467649fc227bee7e0
+SIZE (dmlc-rabit-2f25347_GH0.tar.gz) = 90555
+SHA256 (NVlabs-cub-b20808b_GH0.tar.gz) = 5d613bf9c4bd3f60803720b982291c585e5e9cce83a93b5ef576bd3ce6d141e4
+SIZE (NVlabs-cub-b20808b_GH0.tar.gz) = 390976

misc/xgboost/files/patch-CMakeLists.txt

@@ -0,0 +1,26 @@
+--- CMakeLists.txt.orig	2019-12-14 20:50:22 UTC
++++ CMakeLists.txt
+@@ -101,16 +101,16 @@ endif (USE_CUDA)
+ 
+ # dmlc-core
+ msvc_use_static_runtime()
+-add_subdirectory(${xgboost_SOURCE_DIR}/dmlc-core)
+-set_target_properties(dmlc PROPERTIES
+-  CXX_STANDARD 11
+-  CXX_STANDARD_REQUIRED ON
+-  POSITION_INDEPENDENT_CODE ON)
+-list(APPEND LINKED_LIBRARIES_PRIVATE dmlc)
++#add_subdirectory(${xgboost_SOURCE_DIR}/dmlc-core)
++#set_target_properties(dmlc PROPERTIES
++#  CXX_STANDARD 11
++#  CXX_STANDARD_REQUIRED ON
++#  POSITION_INDEPENDENT_CODE ON)
++#list(APPEND LINKED_LIBRARIES_PRIVATE dmlc)
+ 
+ # rabit
+ set(RABIT_BUILD_DMLC OFF)
+-set(DMLC_ROOT ${xgboost_SOURCE_DIR}/dmlc-core)
++set(DMLC_ROOT ${CMAKE_INSTALL_PREFIX})
+ set(RABIT_WITH_R_LIB ${R_LIB})
+ add_subdirectory(rabit)
+ 

misc/xgboost/files/patch-rabit_CMakeLists.txt

@@ -0,0 +1,34 @@
+--- rabit/CMakeLists.txt.orig	2019-12-14 20:43:47 UTC
++++ rabit/CMakeLists.txt
+@@ -13,7 +13,7 @@ set(DMLC_ROOT ${CMAKE_CURRENT_LIST_DIR}/../dmlc-core)
+ 
+ # moved from xgboost build
+ if(R_LIB OR MINGW OR WIN32)
+-  add_library(rabit src/engine_empty.cc src/c_api.cc)
++  add_library(rabit SHARED src/engine_empty.cc src/c_api.cc)
+   set(rabit_libs rabit)
+   set_target_properties(rabit
+           PROPERTIES CXX_STANDARD 11
+@@ -21,10 +21,10 @@ if(R_LIB OR MINGW OR WIN32)
+           POSITION_INDEPENDENT_CODE ON)
+ else()
+   find_package(Threads REQUIRED)
+-  add_library(rabit_empty src/engine_empty.cc src/c_api.cc)
+-  add_library(rabit_base src/allreduce_base.cc src/engine_base.cc src/c_api.cc)
++  add_library(rabit_empty SHARED src/engine_empty.cc src/c_api.cc)
++  add_library(rabit_base SHARED src/allreduce_base.cc src/engine_base.cc src/c_api.cc)
+ 
+-  add_library(rabit src/allreduce_base.cc src/allreduce_robust.cc src/engine.cc src/c_api.cc)
++  add_library(rabit SHARED src/allreduce_base.cc src/allreduce_robust.cc src/engine.cc src/c_api.cc)
+   add_library(rabit_mock_static src/allreduce_base.cc src/allreduce_robust.cc src/engine_mock.cc src/c_api.cc)
+   add_library(rabit_mock SHARED src/allreduce_base.cc src/allreduce_robust.cc src/engine_mock.cc src/c_api.cc)
+   target_link_libraries(rabit Threads::Threads)
+@@ -43,7 +43,7 @@ if(RABIT_BUILD_MPI)
+   if (NOT MPI_CXX_FOUND)
+     message(FATAL_ERROR "CXX Interface for MPI is required for building MPI backend.")
+   endif (NOT MPI_CXX_FOUND)
+-  add_library(rabit_mpi src/engine_mpi.cc ${MPI_INCLUDE_PATH})
++  add_library(rabit_mpi SHARED src/engine_mpi.cc ${MPI_INCLUDE_PATH})
+   target_link_libraries(rabit_mpi ${MPI_CXX_LIBRARIES})
+   list(APPEND rabit_libs rabit_mpi)
+ endif()

misc/xgboost/files/patch-src_common_json.cc

@@ -0,0 +1,11 @@
+--- src/common/json.cc.orig	2019-12-14 20:32:28 UTC
++++ src/common/json.cc
+@@ -630,7 +630,7 @@ Json JsonReader::ParseNumber() {
+     // multiply zero by inf which gives nan.
+     if (f != 0.0) {
+       // Only use exp10 from libc on gcc+linux
+-#if !defined(__GNUC__) || defined(_WIN32) || defined(__APPLE__)
++#if !defined(__GNUC__) || defined(_WIN32) || defined(__APPLE__) || defined(__FreeBSD__)
+ #define exp10(val) std::pow(10, (val))
+ #endif  // !defined(__GNUC__) || defined(_WIN32) || defined(__APPLE__)
+       f *= exp10(exponent);

misc/xgboost/files/patch-src_common_timer.cc

@@ -0,0 +1,14 @@
+--- src/common/timer.cc.orig	2019-12-14 20:35:26 UTC
++++ src/common/timer.cc
+@@ -44,8 +44,9 @@ std::vector<Monitor::StatMap> Monitor::CollectFromOthe
+     statistic[kv.first] = Object();
+     auto& j_pair = statistic[kv.first];
+     j_pair["count"] = Integer(kv.second.count);
+-    j_pair["elapsed"] = Integer(std::chrono::duration_cast<std::chrono::microseconds>(
+-        kv.second.timer.elapsed).count());
++    // https://github.com/dmlc/xgboost/issues/5117
++    j_pair["elapsed"] = 1; //Integer(std::chrono::duration_cast<std::chrono::microseconds>(
++        //kv.second.timer.elapsed).count());
+   }
+ 
+   std::stringstream ss;

misc/xgboost/pkg-descr

@@ -0,0 +1,5 @@
+Scalable, Portable and Distributed Gradient Boosting (GBDT, GBRT or GBM)
+Library, for Python, R, Java, Scala, C++ and more. Runs on single machine,
+Hadoop, Spark, Flink and DataFlow.
+
+WWW: https://xgboost.ai/

misc/xgboost/pkg-plist

@@ -0,0 +1,46 @@
+bin/xgboost
+include/rabit/c_api.h
+include/rabit/internal/engine.h
+include/rabit/internal/io.h
+include/rabit/internal/rabit-inl.h
+include/rabit/internal/socket.h
+include/rabit/internal/thread_local.h
+include/rabit/internal/timer.h
+include/rabit/internal/utils.h
+include/rabit/rabit.h
+include/rabit/serializable.h
+include/xgboost/base.h
+include/xgboost/c_api.h
+include/xgboost/data.h
+include/xgboost/feature_map.h
+include/xgboost/gbm.h
+include/xgboost/generic_parameters.h
+include/xgboost/host_device_vector.h
+include/xgboost/json.h
+include/xgboost/json_io.h
+include/xgboost/learner.h
+include/xgboost/linear_updater.h
+include/xgboost/logging.h
+include/xgboost/metric.h
+include/xgboost/model.h
+include/xgboost/objective.h
+include/xgboost/parameter.h
+include/xgboost/predictor.h
+include/xgboost/span.h
+include/xgboost/tree_model.h
+include/xgboost/tree_updater.h
+include/xgboost/version_config.h
+lib/cmake/rabit/rabitConfig.cmake
+lib/cmake/rabit/rabitConfigVersion.cmake
+lib/cmake/rabit/rabitTargets-%%CMAKE_BUILD_TYPE%%.cmake
+lib/cmake/rabit/rabitTargets.cmake
+lib/cmake/xgboost/XGBoostTargets-%%CMAKE_BUILD_TYPE%%.cmake
+lib/cmake/xgboost/XGBoostTargets.cmake
+lib/cmake/xgboost/xgboost-config-version.cmake
+lib/cmake/xgboost/xgboost-config.cmake
+lib/librabit.so
+lib/librabit_base.so
+lib/librabit_empty.so
+lib/librabit_mock.so
+lib/librabit_mock_static.a
+lib/libxgboost.so