• GitHub repo
• Latest documentation

Euclid is a header only library for geometry processing and shape analysis.

It contains some utilities and algorithms which extend and cooperate with other popular libraries around there, like Eigen(libigl), CGAL, to name a few.

The purpose of Euclid is not to replace any of the above packages, but to glue things together as well as to provide more algorithms which do not appear in those libraries.

Different packages in Euclid may require a different set of dependencies. Although some are dependency free (like the IO module), but most of them rely on the following packages, which are also marked as 'REQUIRED' if you configure using CMake, including

• Boost
• CGAL
• Eigen
• Libigl

Additional dependencies are required by some packages, including

• Spectra for solving eigenvalue problems.
• Embree for fast cpu ray tracing.
• Vulkan for headless gpu rendering.
• TTK for topological shape analysis.
• Cereal for serialization.

Make sure you properly compile and link to the above libraries when they are used. Also, Euclid uses features in the C++17 standard, so you'll need a C++17 enabled compiler.

Since it's a header only library, it is not mandatory to configure this project with CMake. Although be sure to configure other dependencies properly, as some of them are not header only.

If you are using CMake, there are two ways to go:

First you could use the find script shipped with Euclid and configure other dependencies manually like below. This is preferable if you only need parts of the headers and do not wish to configure all the dependencies.

find_package(Euclid)
target_include_directories(your-target Euclid_INCLUDE_DIR)
// other dependencies, e.g. Eigen

Otherwise, you could configure Euclid using CMake first. It will output an EuclidConfig.cmake file in the build tree for you to use. You can set the variable Euclid_DIR to the path containing this file and then in your own CMakeLists.txt you could do

find_package(Euclid)
target_link_libraries(your-target Euclid::Euclid)

Here's an example which reads a mesh file, converts it to a CGAL::Surface_mesh data structure, computes its discrete gaussian curvatures and ouput the values into mesh colors.

#include <vector>
#include <CGAL/Simple_cartesian.h>
#include <CGAL/Surface_mesh.h>
#include <Euclid/IO/OffIO.h>
#include <Euclid/IO/PlyIO.h>
#include <Euclid/MeshUtil/CGALMesh.h>
#include <Euclid/Geometry/TriMeshGeometry.h>
#include <Euclid/Util/Color.h>

using Kernel = CGAL::Simple_cartesian<float>;
using Point_3 = Kernel::Point_3;
using Mesh = CGAL::Surface_mesh<Point_3>;

int main()
{
    // Read triangle mesh into buffers
    std::vector<float> positions;
    std::vector<unsigned> indices;
    Euclid::read_off<3>("Euclid_root/data/bumpy.off", positions, nullptr, &indices, nullptr);

    // Generate a CGAL::Surface_mesh
    Mesh mesh;
    Euclid::make_mesh<3>(mesh, positions, indices);

    // Compute gaussian curvatures
    auto curvatures = Euclid::gaussian_curvatures(mesh);

    // Turn curvatures into colors and output to a file
    std::vector<unsigned char> colors;
    Euclid::colormap(igl::COLOR_MAP_TYPE_JET, curvatures, colors, true);
    Euclid::write_ply<3>(
        "outdir/bumpy.ply", positions, nullptr, nullptr, &indices, &colors);
}

See the examples folder for more tutorials. However, many modules are not covered yet. For a more complete example, you could check the test cases to see the usage of most functions and classes. More information on how to run the tests could be found here.

MIT for code not related to any third-party libraries.

Otherwise it should follow whatever license the third-party libraries require.