Skip to content

BUILDING.md

Latest commit

 

History

History
85 lines (58 loc) · 4.51 KB

BUILDING.md

File metadata and controls

85 lines (58 loc) · 4.51 KB

Building GDMP

Getting Started

  1. Clone the repository, remember to initialize the submodules.

    git submodule update --init --recursive [--depth=1]

  2. Install Bazelisk or Bazel version that meets MediaPipe requirement.

  3. Run setup.py to generate the godot-cpp bindings as well as apply/setup source code to the mediapipe workspace. Run setup.py --help to view the various options.

Example:

./setup.py --godot-binary path/to/godot/binary
  1. Copy addons to your Godot project's root directory.
  2. (Optional) Add calculator dependencies in GDMP/GDMP.bzl if you have additional calculators to compile into GDMP.

General Build Instructions

build.py is a wrapper script that invoke Bazel build commands in mediapipe workspace to build GDMP libraries.

Common usage:

./build.py [target] --type [debug/release] --output [output-dir]

Run build.py --help to see all options.

Building for Android

Windows users: Cross-compiling Android library on Windows is currently not supported, please set up WSL or Linux VM before proceed to the instructions below.

  1. Install Android NDK, then setting ANDROID_NDK_HOME environment variable that point to your install location.

    Refer to MediaPipe docs for more details.

  2. Run build.py android to build Android native libraries.

    Use --arch to specify the architecture. For example, run:

    build.py android --arch arm64-v8a

    to build library for arm64-v8a architecture.

Linking prebuilt OpenCV: GDMP builds OpenCV from source on Android by default, if you want to use pre-built OpenCV libraries, you will need to:

  1. Comment out --define=OPENCV=source from build.py to make Bazel using prebuilt OpenCV rules.
  2. Make sure to add libopencv_java4.so to the dependencies list of Android targets in addons/GDMP/GDMP.gdextension for your project.

Building for iOS

  1. Install Xcode, then switch to installed Xcode by using xcode-select

  2. Run build.py ios to build iOS library

    By default, GDMP build arm64 architecture for iOS device, and both arm64 and x86_64 for iOS simulator. If you would like to customize which platform or architecture to build, modify ios attribute in GDMP/ios/BUILD.

Building for Linux

Flatpak environment: If you intend to distribute the library, you might want to use an environment isolated from host before you proceed. You can use scripts/flatpak_env.sh to enable Flatpak environment.

  1. Run build.py desktop to build desktop library.

Linking prebuilt OpenCV: GDMP builds OpenCV from source on Linux by default, if you want to use pre-compiled OpenCV libraries, you will need to:

  1. Modify mediapipe/third_party/opencv_linux.BUILD to make OpenCV installed on your host visible to Bazel.
  2. Comment out --define=OPENCV=source from build.py to make Bazel using binary OpenCV rules.

When linking prebuilt OpenCV libraries, build script will not copy required libraries to output destination, make sure they can be found by dynamic linker.

Building for macOS

  1. Install Xcode, then switch to installed Xcode by using xcode-select
  2. Run build.py desktop to build desktop library.

Building for Windows

  1. Install MSVC compiler and Windows SDK, then setting BAZEL_VC environment variable pointing to your VC installation.

    Refer to Bazel documentation for more details.

  2. Bash is required for building MediaPipe, make sure bash is in PATH or setting BAZEL_SH environment variable pointing to it.

    Bash can be installed from Git for Windows or MSYS2

    Refer to MediaPipe documentation for more details.

  3. Run build.py desktop to build desktop library.

Linking prebuilt OpenCV: GDMP builds OpenCV from source on Windows by default, if you want to use pre-compiled OpenCV libraries, you will need to:

  1. Modify mediapipe/WORKSPACE for path under windows_opencv if OpenCV is not installed on C:\opencv
  2. Modify OPENCV_VERSION in mediapipe/third_party/opencv_windows.BUILD if OpenCV version is not 3.4.10
  3. Comment out --define=OPENCV=source from build.py to make Bazel using binary OpenCV rules.