Skip to main content
MLPP uses CMake 3.20+ as its build system, providing flexible configuration options and modern package management integration. The library can be used as a standalone project or integrated into larger CMake-based builds.

​
CMake configuration options

MLPP exposes several options to control build behavior and optional dependencies: 
option(MLPP_BUILD_TESTS      "Build tests"                     ON)
option(MLPP_BUILD_EXAMPLES   "Build examples"                  ON)
option(MLPP_WITH_PYPLOT      "Enable matplotlibcpp (Python+NumPy)" ON)
option(MLPP_WITH_OPENCV      "Enable OpenCV support"           ON)
option(MLPP_USE_EIGEN        "Use Eigen when available"        ON)

​
Core options

​
MLPP_BUILD_TESTS
boolean
default:"ON"
Build the test suite. Useful during development but can be disabled for production integration.
​
MLPP_BUILD_EXAMPLES
boolean
default:"ON"
Compile example programs demonstrating library usage. Disable to speed up builds when examples aren’t needed.
​
MLPP_USE_EIGEN
boolean
default:"ON"
Enable Eigen 3.4+ integration for linear algebra operations. Highly recommended for performance.
​
MLPP_WITH_PYPLOT
boolean
default:"ON"
Enable matplotlib visualization support via matplotlibcpp. Requires Python3 and NumPy.
​
MLPP_WITH_OPENCV
boolean
default:"ON"
Enable OpenCV integration for computer vision tasks. Optional but useful for image-based ML.

​
Dependency management

MLPP uses find_package() to locate optional dependencies and automatically configures compile definitions: 
if(MLPP_USE_EIGEN)
    find_package(Eigen3 3.4 QUIET NO_MODULE)
    if(Eigen3_FOUND)
        target_link_libraries(mlpp INTERFACE Eigen3::Eigen)
        target_compile_definitions(mlpp INTERFACE MLPP_HAS_EIGEN)
    endif()
endif()

if(MLPP_WITH_PYPLOT)
    find_package(Python3 REQUIRED COMPONENTS Interpreter Development NumPy)
    target_link_libraries(mlpp INTERFACE Python3::Python Python3::NumPy)
    target_compile_definitions(mlpp INTERFACE MLPP_WITH_PYPLOT)
endif()

if(MLPP_WITH_OPENCV)
    find_package(OpenCV QUIET COMPONENTS core imgproc highgui)
    if(OpenCV_FOUND)
        target_link_libraries(mlpp INTERFACE opencv_core opencv_imgproc opencv_highgui)
        target_compile_definitions(mlpp INTERFACE MLPP_HAS_OPENCV)
    endif()
endif()
Dependencies are linked with INTERFACE visibility, meaning they propagate to consuming targets automatically.

​
Integration patterns

MLPP supports multiple integration workflows depending on your project structure. Use CMake’s FetchContent module to download and integrate MLPP directly: 
cmake_minimum_required(VERSION 3.20)
project(my_ml_project)

include(FetchContent)

FetchContent_Declare(
    mlpp
    GIT_REPOSITORY https://github.com/your-org/mlpp.git
    GIT_TAG        v0.3.0
)

# Optionally configure MLPP options before making it available
set(MLPP_BUILD_TESTS OFF)
set(MLPP_BUILD_EXAMPLES OFF)

FetchContent_MakeAvailable(mlpp)

add_executable(my_app main.cpp)
target_link_libraries(my_app PRIVATE mlpp::mlpp)

​
Method 2: find_package (installed library)

If MLPP is installed system-wide, use find_package() with the generated config file: 
cmake_minimum_required(VERSION 3.20)
project(my_ml_project)

find_package(mlpp 0.3 REQUIRED)

add_executable(my_app main.cpp)
target_link_libraries(my_app PRIVATE mlpp::mlpp)

​
Method 3: add_subdirectory (local copy)

For projects with MLPP as a subdirectory: 
cmake_minimum_required(VERSION 3.20)
project(my_ml_project)

add_subdirectory(external/mlpp)

add_executable(my_app main.cpp)
target_link_libraries(my_app PRIVATE mlpp::mlpp)

​
Package installation

MLPP includes complete installation support with CMake package config files: 
include(GNUInstallDirs)

install(DIRECTORY include/mlpp DESTINATION ${CMAKE_INSTALL_INCLUDEDIR})

install(TARGETS mlpp EXPORT mlppTargets
    INCLUDES DESTINATION ${CMAKE_INSTALL_INCLUDEDIR}
)

install(EXPORT mlppTargets
    FILE mlppTargets.cmake
    NAMESPACE mlpp::
    DESTINATION ${CMAKE_INSTALL_LIBDIR}/cmake/mlpp
)

​
Installation workflow

1

Configure

Configure the build with desired options:
cmake -B build -DCMAKE_BUILD_TYPE=Release \
               -DMLPP_BUILD_TESTS=OFF \
               -DMLPP_BUILD_EXAMPLES=OFF
2

Install

Install headers and package configuration:
cmake --install build --prefix /usr/local
3

Use in projects

Projects can now find MLPP via find_package(mlpp).

​
Package configuration

The installed package config automatically finds required dependencies: 
@PACKAGE_INIT@

include(CMakeFindDependencyMacro)

if(@MLPP_USE_EIGEN@)
    find_dependency(Eigen3 3.4)
endif()

if(@MLPP_WITH_PYPLOT@)
    find_dependency(Python3 COMPONENTS NumPy)
endif()

if(@MLPP_WITH_OPENCV@)
    find_dependency(OpenCV COMPONENTS core imgproc highgui)
endif()

include("${CMAKE_CURRENT_LIST_DIR}/mlppTargets.cmake")
check_required_components(mlpp)
This ensures consuming projects automatically get all required transitive dependencies.

​
Compiler requirements

MLPP requires C++20 support: 
target_compile_features(mlpp INTERFACE cxx_std_20)

​
Tested compilers

  • GCC 10+
  • Clang 12+
  • MSVC 2019+
  • AppleClang 13+

​
Build tips

For the smallest footprint, disable all optional features:
cmake -B build -DMLPP_BUILD_TESTS=OFF \
               -DMLPP_BUILD_EXAMPLES=OFF \
               -DMLPP_WITH_PYPLOT=OFF \
               -DMLPP_WITH_OPENCV=OFF
Enable all features for development and testing:
cmake -B build -DCMAKE_BUILD_TYPE=Debug \
               -DMLPP_BUILD_TESTS=ON \
               -DMLPP_BUILD_EXAMPLES=ON
cmake --build build
ctest --test-dir build
Use CMake presets for reproducible multi-platform builds:
{
  "version": 3,
  "configurePresets": [
    {
      "name": "default",
      "generator": "Ninja",
      "binaryDir": "${sourceDir}/build",
      "cacheVariables": {
        "CMAKE_BUILD_TYPE": "Release",
        "MLPP_BUILD_TESTS": "OFF"
      }
    }
  ]
}

​
Troubleshooting

If Eigen is not found automatically, set the Eigen3_DIR CMake variable to point to your Eigen installation’s CMake directory.

​
Common issues

Eigen not detected: Ensure Eigen 3.4+ is installed and discoverable via CMake. Set Eigen3_DIR if needed: 
cmake -B build -DEigen3_DIR=/path/to/eigen3/share/eigen3/cmake
Python/NumPy errors: The Python3 development headers and NumPy must be available. Disable pyplot if not needed: 
cmake -B build -DMLPP_WITH_PYPLOT=OFF
C++20 not supported: Update your compiler or explicitly specify the compiler version: 
cmake -B build -DCMAKE_CXX_COMPILER=g++-11

Build docs developers (and LLMs) love

Get started for freeTalk to us