Regarding the computation of intersection of two meshes where more than distinct (watertight) region.

[bardiche89]

When computing intersections, when there are more than one distinct region, are there any clever means to generate separate mesh entities for each region?

Thank you

[Fedr]

Could you please clarify the question and probably show some example?

[bardiche89]

Sorry,

I compute the intersection between two meshes.
This gave me two distinct and disconnected regions.
I wonder if there is a way then to save each distinct region as a individual new mesh.

For example, I notice when I invoke getAllComponents() on my initial intersection mesh, it returns me a list of two BitSet entities. So it has clearly identified the two regions, is there a way to build new mesh from this?

[bardiche89]

I created a minimal working example:

import meshlib.mrmeshpy as mr

mesh = mr.loadMesh("out_boolean.stl")
print(f"{mesh.topology.numValidFaces()} faces")

components = mr.getAllComponents(mesh)
print(f"{len(components)} components")

for i, face_bitset in enumerate(components):
    print(f"{face_bitset.count()} faces")
    region = mr.MeshPart(mesh, face_bitset)
    component_mesh = region.mesh
    print(f"{component_mesh.topology.numValidFaces()} faces")
    mr.saveMesh(component_mesh, f"component_{i}.stl")

out_boolean.zip

344 faces
Found 2 components
172 faces
344 faces
172 faces
344 faces

So it is clear I am not extracting the individual mesh region correctly. Could someone share the correct way to do this?

Thank you.

[Grantim]

Hello!

You need to create new mesh out of face bitsets like this:

import meshlib.mrmeshpy as mr

mesh = mr.loadMesh("out_boolean.stl")
print(f"{mesh.topology.numValidFaces()} faces")

components = mr.getAllComponents(mesh)
print(f"{len(components)} components")

for i, face_bitset in enumerate(components):
    print(f"{face_bitset.count()} faces")
    component_mesh = mr.Mesh()
    component_mesh.addMeshPart(mr.MeshPart(mesh, face_bitset))
    print(f"{component_mesh.topology.numValidFaces()} faces")
    mr.saveMesh(component_mesh, f"component_{i}.stl")

[bardiche89]

Thank you very much once again.

[bardiche89]

A further question,

Could you also give me an example of this loop in C++? Or point me to the relevant APIs?

I know I can get a vector of FaceBitSet using getAllComponents, but are there equivalent C++ APIs for creating new mesh from these facebitsets?

Thank you.

[Grantim]

Yes, all python functions have analogues for c++ and vice versa:

c++ code will look somehow like this

#include "MRMesh/MRMeshLoad.h"
#include "MRMesh/MRMeshComponents.h"
#include "MRMesh/MRMeshSave.h"
#include "MRMesh/MRMesh.h"

int main()
{
    auto meshLoadRes = MR::MeshLoad::fromAnySupportedFormat( "out_boolean.stl" );
    if ( !meshLoadRes.has_value() )
    {
        std::cerr << meshLoadRes.error();
        return 1;
    }
    auto components = MR::MeshComponents::getAllComponents( *meshLoadRes );
    for ( int i = 0; i < components.size(); ++i )
    {
        MR::Mesh component_mesh;
        component_mesh.addMeshPart( {*meshLoadRes,&components[i]} );
        auto saveRes = MR::MeshSave::toAnySupportedFormat( component_mesh, "component_" + std::to_string(i)+".stl" );
        if ( !saveRes.has_value() )
        {
            std::cerr << saveRes.error();
            return 1;
        }
    }
    return 0;
}

[bardiche89]

Thank you very much yet again. Your guidance is much appreciated.