EyeAICore

the core implementation of EyeAI in crossplatform c++

Building EyeAICore from source using Docker container

Build docker image:

cd .. # Dockerfile is located at the root, since it copies MiDaS from EyeAIApp
docker build -t eye-ai .

Now you have a docker image that contains a build of EyeAICore from source.

To run the docker image in a container:

docker run -it eye-ai

Building EyeAICore from source locally

Requirements:

• C++ compiler that supports all features of C++ 20. (for example: GCC 13+, Clang 14+ or MSVC 19.30+)
• CMake: version 3.22 or higher
• Git
• Python 3

Optional programs to install:

• Ninja: build system (used by the cmake presets, package is often named: ninja-build)
• clang-tidy: static code analysis (make sure to use at least clang-tidy-15, clang-tidy-14 has known issues with <chrono>)
• ccache: cache compilation results for faster repeated builds
• doxygen: generate documentation from source code

cmake --preset=release
cmake --build --preset=release

Run EyeAICore Tests

cmake --preset=debug
cmake --build --preset=debug
ctest --preset=default

Profile EyeAICore

• For EyeAIApp: select the "profiling" build type

• For running EyeAICore on your PC:

  cmake --preset=profiling
  cmake --build --preset=profiling

  Then run EyeAICore and listen with the Tracy profiler

Generating EyeAICore docs using doxygen

install doxygen (here: using linux):

sudo apt install doxygen graphviz

generate docs:

cmake --preset=debug
cmake --build --preset=generate_docs

the generated doxygen docs will be in build/docs

See also

Metric Depth