Python Hierarchy

.gitignore

@@ -130,3 +130,6 @@ dmypy.json
 
 # Pyre type checker
 .pyre/
+
+# PyCharm
+.idea/

.gitmodules

@@ -1,9 +1,14 @@
 [submodule "pybmix/src/bayesmix"]
 	path = pybmix/core/pybmixcpp/bayesmix
-	url = https://github.com/bayesmix-dev/bayesmix.git
+	url = https://github.com/taguhiM/bayesmix.git
 	depth = 1
 	update = merge
 [submodule "lib/pybind11"]
 	path = lib/pybind11
 	depth = 1
 	url = https://github.com/pybind/pybind11.git
+[submodule "lib/math"]
+	path = lib/math
+    depth = 1
+	url = https://github.com/bayesmix-dev/math.git
+    ignore = dirty

CMakeLists.txt

@@ -52,4 +52,5 @@ target_compile_definitions(pybmixcpp PRIVATE VERSION_INFO=${EXAMPLE_VERSION_INFO
 target_link_libraries(pybmixcpp PUBLIC bayesmixlib ${BAYESMIX_LINK_LIBRARIES})
 target_compile_options(pybmixcpp PUBLIC ${BAYESMIX_COMPILE_OPTIONS})
 
-add_custom_target(genterate_protos ALL DEPENDS ${PROTO_PYS})
+add_custom_target(generate_protos ALL DEPENDS ${PROTO_PYS})
+add_custom_target(two_to_three ALL COMMAND ${CMAKE_CURRENT_LIST_DIR}/convert_proto.sh DEPENDS generate_protos)

convert_proto.sh

@@ -0,0 +1,6 @@
+SCRIPT_DIR=$( cd -- "$( dirname -- "${BASH_SOURCE[0]}" )" &> /dev/null && pwd )/..
+
+# NEEDS TO BE SET MANUALLY
+ENV_DIR=/home/taguhi/miniconda3/envs/pybmix3.9
+
+$ENV_DIR/bin/2to3 --output-dir=$SCRIPT_DIR/pybmix/proto -W -n $SCRIPT_DIR/pybmix/proto

docs/examples/test_run.py

@@ -0,0 +1,48 @@
+import os
+import sys
+HERE = os.path.dirname(os.path.realpath(__file__))
+BUILD_DIR = os.path.join(HERE, "../../")
+sys.path.insert(0, os.path.realpath(BUILD_DIR))
+
+import numpy as np
+import matplotlib.pyplot as plt
+from pybmix.core.mixing import DirichletProcessMixing
+from pybmix.core.hierarchy import PythonHierarchy
+from pybmix.core.mixture_model import MixtureModel
+np.random.seed(2021)
+
+
+def sample_from_mixture(weights, means, sds, n_data):
+    n_comp = len(weights)
+    clus_alloc = np.random.choice(np.arange(n_comp), p=[0.5, 0.5], size=n_data)
+    return np.random.normal(loc=means[clus_alloc], scale=sds[clus_alloc])
+
+
+y = sample_from_mixture(np.array([0.5, 0.5]), np.array([-3, 3]), np.array([1, 1]), 200)
+
+
+plt.hist(y)
+plt.show()
+
+mixing = DirichletProcessMixing(total_mass=5)
+hierarchy = PythonHierarchy()
+hierarchy.make_default_fixed_params(y, 2)
+mixture = MixtureModel(mixing, hierarchy)
+
+mixture.run_mcmc(y, algorithm="Neal2", niter=2000, nburn=1000)
+
+from pybmix.estimators.density_estimator import DensityEstimator
+
+grid = np.linspace(-6, 6, 500)
+dens_est = DensityEstimator(mixture)
+densities = dens_est.estimate_density(grid)
+
+plt.hist(y, density=True)
+plt.plot(grid, np.mean(densities, axis=0), lw=3, label="predictive density")
+idxs = [5, 100, 300]
+
+for idx in idxs:
+    plt.plot(grid, densities[idx, :], "--", label="iteration: {0}".format(idx))
+    plt.legend()
+    plt.show()
+

make_.sh

@@ -0,0 +1,7 @@
+#!/usr/bin/env bash
+
+cd build/;
+cmake ..;
+make pybmixcpp;
+make generate_protos;
+make two_to_three;

pybmix/core/CMakeLists.txt

@@ -0,0 +1,220 @@
+cmake_minimum_required(VERSION 3.14.0)
+list(APPEND CMAKE_MODULE_PATH "${CMAKE_CURRENT_LIST_DIR}")
+message("CMAKE_MODULE_PATH ${CMAKE_MODULE_PATH}")
+
+find_program(CCACHE_PROGRAM ccache)
+if (CCACHE_PROGRAM)
+    set_property(GLOBAL PROPERTY RULE_LAUNCH_COMPILE "${CCACHE_PROGRAM}")
+endif()
+
+project(bayesmix)
+
+set(CMAKE_CXX_STANDARD 17)
+set(CMAKE_BUILD_TYPE Release)
+
+set(CMAKE_MODULE_PATH "${PROJECT_SOURCE_DIR}/cmake" ${CMAKE_MODULE_PATH})
+set(CMAKE_CXX_FLAGS_RELEASE "-O3 -funroll-loops -ftree-vectorize")
+set(CMAKE_FIND_PACKAGE_PREFER_CONFIG TRUE)
+
+# Require PkgConfig
+find_package(PkgConfig REQUIRED)
+find_package(OpenMP REQUIRED)
+find_package(pybind11 REQUIRED)
+
+# TBB CMake integration
+message(STATUS "Using math TBB")
+# Define TBB_ROOT Folder
+set(TBB_ROOT ${CMAKE_CURRENT_LIST_DIR}/pybmixcpp/bayesmix/lib/math/lib/tbb)
+file(COPY ${CMAKE_CURRENT_LIST_DIR}/pybmixcpp/bayesmix/lib/math/lib/tbb_2019_U8/ DESTINATION ${TBB_ROOT})
+# Build TBB Library with CMake Integration
+include(${TBB_ROOT}/cmake/TBBBuild.cmake)
+list(APPEND MAKE_ARGS "tbb_build_dir=${TBB_ROOT}")
+list(APPEND MAKE_ARGS "tbb_build_prefix=tbb")
+tbb_build(TBB_ROOT ${TBB_ROOT} CONFIG_DIR TBB_DIR MAKE_ARGS ${MAKE_ARGS})
+# Require TBB library (for compile / link options)
+find_package(TBB)
+
+# Check if Protobuf is present in system
+find_package(Protobuf)
+if (NOT Protobuf_FOUND AND NOT PROTOBUF_FOUND AND NOT TARGET protobuf::libprotobuf)
+    if (${CMAKE_VERSION} VERSION_LESS "3.20.0")
+        message(FATAL_ERROR
+                "Your cmake version is too old: either install a newer version (>=3.20)"
+                " or install google's protocol buffer (protobuf) library by hand.")
+    endif()
+
+    include(FetchContent)
+    set(FETCHCONTENT_QUIET OFF)
+    set(FETCHCONTENT_UPDATES_DISCONNECTED ON)
+    set(BUILD_SHARED_LIBS OFF)
+    set(CMAKE_POSITION_INDEPENDENT_CODE ON)
+    set(BUILD_TESTING OFF)
+
+    message(CHECK_START "Fetching Protobuf")
+    list(APPEND CMAKE_MESSAGE_INDENT "  ")
+
+    option(protobuf_BUILD_TESTS "" OFF)
+    set(protobuf_BUILD_EXPORT OFF)
+    set(protobuf_MSVC_STATIC_RUNTIME OFF)
+    FetchContent_Declare(
+            protobuf
+            GIT_REPOSITORY "https://github.com/protocolbuffers/protobuf.git"
+            GIT_TAG "v3.14.0"
+            GIT_SUBMODULES ""
+            SOURCE_SUBDIR cmake)
+    FetchContent_MakeAvailable(protobuf)
+
+    list(POP_BACK CMAKE_MESSAGE_INDENT)
+    message(CHECK_PASS "fetched")
+    message("Protobuf_FOUND ${Protobuf_FOUND}")
+    message("   --> PROTOBUF LIB: ${PROTOBUF_LIBRARIES}")
+    message("   --> PROTOBUF INCLUDE: ${PROTOBUF_INCLUDE_DIRS}")
+    message("   --> PROTOBUF VERSION: ${Protobuf_VERSION}")
+    message("   --> PROTOBUF Found: ${Protobuf_FOUND}")
+
+endif()
+
+option(DISABLE_TESTS
+        "If tests should be compiled or no" OFF)
+option(DISABLE_BENCHMARKS
+        "If benchmarks should be compiled or no" OFF)
+option(DISABLE_DOCS
+        "If docs should be generated or no" OFF)
+option(BUILD_RUN "" ON)
+
+# set(CMAKE_BUILD_WITH_INSTALL_RPATH ON)
+set(CMAKE_INSTALL_RPATH_USE_LINK_PATH TRUE)
+set(BASEPATH "${CMAKE_CURRENT_LIST_DIR}")
+set(INCLUDE_PATHS
+        ${BASEPATH}
+        ${CMAKE_CURRENT_LIST_DIR}/pybmixcpp/bayesmix/lib/math
+        ${CMAKE_CURRENT_LIST_DIR}/pybmixcpp/bayesmix/lib/math/lib/boost_1.72.0
+        ${CMAKE_CURRENT_LIST_DIR}/pybmixcpp/bayesmix/lib/math/lib/eigen_3.3.9
+        # TBB already included
+        ${CMAKE_CURRENT_BINARY_DIR}
+        ${protobuf_SOURCE_DIR}/pybmixcpp/bayesmix/src
+        ${protobuf_SOURCE_DIR}/src
+        )
+
+
+set(LINK_LIBRARIES
+        pthread
+        protobuf::libprotobuf
+        TBB::tbb
+        OpenMP::OpenMP_CXX
+        pybind11::embed
+        pybind11::pybind11
+        )
+
+set(COMPILE_OPTIONS -D_REENTRANT -fPIC)
+
+file(GLOB ProtoFiles "${BASEPATH}/pybmixcpp/bayesmix/src/proto/*.proto")
+set(PROTO_DIR proto)
+
+foreach(PROTO_FILE IN LISTS ProtoFiles)
+    message(STATUS "protoc proto(cc): ${PROTO_FILE}")
+    get_filename_component(PROTO_DIR ${PROTO_FILE} DIRECTORY)
+    get_filename_component(PROTO_NAME ${PROTO_FILE} NAME_WE)
+    set(PROTO_HDR ${CMAKE_CURRENT_BINARY_DIR}/${PROTO_NAME}.pb.h)
+    set(PROTO_SRC ${CMAKE_CURRENT_BINARY_DIR}/${PROTO_NAME}.pb.cc)
+    message(STATUS "protoc hdr: ${PROTO_HDR}")
+    message(STATUS "protoc src: ${PROTO_SRC}")
+    add_custom_command(
+            OUTPUT ${PROTO_SRC} ${PROTO_HDR}
+            COMMAND protobuf::protoc "--proto_path=${BASEPATH}/pybmixcpp/bayesmix/src/proto"
+            ${PROTO_DIRS} "--cpp_out=${PROJECT_BINARY_DIR}" ${PROTO_FILE}
+            DEPENDS ${PROTO_FILE} protobuf::protoc
+            COMMENT "Generate C++ protocol buffer for ${PROTO_FILE}"
+            VERBATIM)
+    list(APPEND PROTO_HDRS ${PROTO_HDR})
+    list(APPEND PROTO_SRCS ${PROTO_SRC})
+endforeach()
+
+SET_SOURCE_FILES_PROPERTIES(${PROTO_SRCS} ${PROTO_HDRS} PROPERTIES GENERATED TRUE)
+
+get_directory_property(HAS_PARENT PARENT_DIRECTORY)
+if(HAS_PARENT)
+    set(BAYESMIX_INCLUDE_PATHS ${INCLUDE_PATHS} PARENT_SCOPE)
+    set(BAYESMIX_LINK_LIBRARIES ${LINK_LIBRARIES} PARENT_SCOPE)
+    set(BAYESMIX_COMPILE_OPTIONS ${COMPILE_OPTIONS} PARENT_SCOPE)
+    set(PROTO_HEADERS ${PROTO_HDRS} PARENT_SCOPE)
+    set(PROTO_SOURCES ${PROTO_SRCS} PARENT_SCOPE)
+    set(ProtoFiles ${ProtoFiles} PARENT_SCOPE)
+endif()
+
+# Build library object
+add_library(bayesmix OBJECT)
+target_sources(bayesmix PUBLIC ${PROTO_SRCS} ${PROTO_HDRS})
+add_subdirectory(pybmixcpp/bayesmix/src)
+target_include_directories(bayesmix PUBLIC ${INCLUDE_PATHS})
+target_include_directories(bayesmix PUBLIC pybmixcpp/bayesmix)
+target_link_libraries(bayesmix PUBLIC ${LINK_LIBRARIES})
+target_compile_options(bayesmix PUBLIC ${COMPILE_OPTIONS})
+
+# Build static library
+add_library(bayesmixlib $<TARGET_OBJECTS:bayesmix>)
+
+if (BUILD_RUN)
+    # Build run executable
+    add_executable(run_mcmc $<TARGET_OBJECTS:bayesmix> pybmixcpp/run_mcmc.cc)
+    target_include_directories(run_mcmc PUBLIC ${INCLUDE_PATHS})
+    target_include_directories(run_mcmc PUBLIC pybmixcpp/bayesmix/)
+    target_include_directories(run_mcmc PUBLIC pybmixcpp/bayesmix/src)
+    target_include_directories(run_mcmc PUBLIC build)
+    target_link_libraries(run_mcmc PUBLIC ${LINK_LIBRARIES})
+    target_compile_options(run_mcmc PUBLIC ${COMPILE_OPTIONS})
+endif()
+
+add_subdirectory(pybmixcpp)
+
+if (NOT DISABLE_TESTS)
+    add_subdirectory(test)
+endif()
+
+if (NOT DISABLE_BENCHMARKS)
+    add_subdirectory(benchmarks)
+endif()
+
+if (NOT DISABLE_DOCS)
+    add_subdirectory(docs)
+endif()
+
+if (NOT DISABLE_PLOTS)
+    include(FetchContent)
+
+    set(FETCHCONTENT_QUIET OFF)
+    set(FETCHCONTENT_UPDATES_DISCONNECTED ON)
+    set(BUILD_SHARED_LIBS OFF)
+    set(CMAKE_POSITION_INDEPENDENT_CODE ON)
+    set(BUILD_TESTING OFF)
+
+    message(CHECK_START "Fetching Matplotplusplus")
+    list(APPEND CMAKE_MESSAGE_INDENT "  ")
+
+    FetchContent_Declare(matplotplusplus
+            GIT_REPOSITORY https://github.com/alandefreitas/matplotplusplus
+            GIT_TAG origin/master # or whatever tag you want
+            )
+
+    FetchContent_GetProperties(matplotplusplus)
+    if(NOT matplotplusplus_POPULATED)
+        FetchContent_Populate(matplotplusplus)
+        add_subdirectory(${matplotplusplus_SOURCE_DIR} ${matplotplusplus_BINARY_DIR}
+                EXCLUDE_FROM_ALL)
+    endif()
+    # find_package(Matplot++ REQUIRED)
+    message("matplot " ${matplotplusplus_SOURCE_DIR})
+
+    add_executable(plot_mcmc $<TARGET_OBJECTS:bayesmix> pybmixcpp/plot_mcmc.cc
+            pybmixcpp/bayesmix/src/plots/plot_utils.h
+            pybmixcpp/bayesmix/src/plots/plot_utils.cc)
+    target_include_directories(plot_mcmc PUBLIC
+            ${INCLUDE_PATHS} ${matplotplusplus_SOURCE_DIR}/source)
+    target_include_directories(plot_mcmc PUBLIC pybmixcpp/bayesmix)
+    target_link_libraries(plot_mcmc PUBLIC ${LINK_LIBRARIES} matplot)
+    target_compile_options(plot_mcmc PUBLIC ${COMPILE_OPTIONS})
+endif()
+
+if (NOT DISABLE_EXAMPLES)
+    add_subdirectory(pybmixcpp/bayesmix/examples)
+endif()

pybmix/core/algo_python.asciipb

@@ -0,0 +1,27 @@
+##### GENERIC SETTINGS FOR ALL ALGORITHMS #####
+# Algorithm ID string, e.g. "Neal2"
+algo_id: "Neal3"
+
+# RNG initial seed: any nonnegative integer
+rng_seed: 20201124
+
+# Number of iterations of the algorithm
+iterations: 1100
+
+# Number of initial iterations discarded by the algorithm
+burnin: 100
+
+# Number of clusters in which data will be first initialized
+# (NOTE: If you wish to initialize one datum per cluster, please write 0.)
+# (NOTE: This value is ONLY used for initialization, and it may be overwritten
+#  by certain mixing objects, such as LogSBMixing. Please check a mixing's
+#  initialize() function to know for sure whether or not it will override this
+#  value.)
+init_num_clusters: 3
+
+
+##### ALGORITHM-SPECIFIC SETTINGS #####
+# Neal8 number of auxiliary blocks
+# (NOTE: 3 is the recommended value in most cases, please change it only if you
+#  know what you're doing.)
+neal8_n_aux: 3