Trying to use visitor in split() method

[citystrawman]

Please use the following template to help us solving your issue.

Issue Details

I am using split function to intersect two surface meshes and I want to get the newly created points/edges during split so that I can remesh these points. From this link I think I could use visitor to get what I want. the surface meshes, named slope and slide respectively, are shown as follows:

However, I could not get any infomation, before_subface_creations, after_subface_created after_subface_creations all the 3 call back functions do not get called.

/

I hope if anyone could help me solve this. Thank you!

Source Code

#include <CGAL/Exact_predicates_inexact_constructions_kernel.h>
#include <CGAL/Exact_predicates_exact_constructions_kernel.h>
#include <CGAL/Surface_mesh.h>
#include <CGAL/boost/graph/selection.h>
#include <CGAL/Polygon_mesh_processing/remesh.h>

#include <CGAL/Polygon_mesh_processing/corefinement.h>
#include <CGAL/Polygon_mesh_processing/clip.h>
#include <CGAL/Polygon_mesh_processing/IO/polygon_mesh_io.h>

#include <iostream>
#include <string>

typedef CGAL::Exact_predicates_inexact_constructions_kernel K;
typedef CGAL::Exact_predicates_exact_constructions_kernel EK;
typedef CGAL::Surface_mesh<K::Point_3> Mesh;
typedef boost::graph_traits<Mesh>::vertex_descriptor vertex_descriptor;
typedef Mesh::Property_map<vertex_descriptor, EK::Point_3> Exact_point_map;
typedef boost::graph_traits<Mesh>::face_descriptor            face_descriptor;
typedef boost::graph_traits<Mesh>::edge_descriptor            edge_descriptor;
typedef boost::graph_traits<Mesh>::halfedge_descriptor        halfedge_descriptor;


namespace PMP = CGAL::Polygon_mesh_processing;
namespace params = CGAL::parameters;

class Insert_iterator
{
    typedef std::unordered_map<face_descriptor, face_descriptor> Container;
    Container& container;
public:

    Insert_iterator(Container& c)
        : container(c) {}

    Insert_iterator&
        operator=(const std::pair<face_descriptor, face_descriptor>& p)
    {
        container[p.second] = p.first;
        return *this;
    }

    Insert_iterator&
        operator*() { return *this; }

    Insert_iterator
        operator++(int) { return *this; }

};
struct Visitor : public CGAL::Polygon_mesh_processing::Corefinement::Default_visitor<Mesh>
{
    typedef std::unordered_map<face_descriptor, face_descriptor> Container;

    Container& container;
    face_descriptor qfd;

    Visitor(Container& container)
        : container(container)
    {}

    void before_subface_creations(face_descriptor fd)
    {
        std::cout << "split : " << fd << " into:" << std::endl;
        Container::iterator it = container.find(fd);
        qfd = it->second;
        container.erase(it);
    }

    void after_subface_created(face_descriptor fd)
    {
        std::cout << "  " << fd;
        container[fd] = qfd;
    }

    void after_subface_creations()
    {
        std::cout << std::endl;
    }
};

int main(int argc, char* argv[])
{
    const std::string filename1 = (argc > 1) ? argv[1] : CGAL::data_file_path("meshes/slope.off");
    const std::string filename2 = (argc > 2) ? argv[2] : CGAL::data_file_path("meshes/slide.off");

    Mesh mesh1, mesh2;
    if (!PMP::IO::read_polygon_mesh(filename1, mesh1) || !PMP::IO::read_polygon_mesh(filename2, mesh2))
    {
        std::cerr << "Invalid input." << std::endl;
        return 1;
    }

    if (PMP::does_self_intersect(mesh1) || PMP::does_self_intersect(mesh2))
    {
        std::cerr << "self intersect input." << std::endl;
        return 1;
    }

    std::unordered_map<face_descriptor, face_descriptor> t2q;
    Visitor v(t2q);

    PMP::split(mesh1, mesh2, params::do_not_modify(true), params::visitor(v).do_not_modify(true));

    return 0;
}

Environment

• Operating system (Windows/Mac/Linux, 32/64 bits): Windows 11
• Compiler: MSVC
• Release or debug mode:
• Specific flags used (if any):
• CGAL version: 5.6.1
• Boost version:
• Other libraries versions if used (Eigen, TBB, etc.):

[sloriot]

The visitor should be in the first named parameter, not the second one (you have to expand the doc to see the np_tm only). This kind of problem should be fixed when #7966 is merged and used for all functions.

[citystrawman]

The visitor should be in the first named parameter, not the second one (you have to expand the doc to see the np_tm only). This kind of problem should be fixed when #7966 is merged and used for all functions.

Thank you for your kind remind, but after changing the visitor to the first name parameter, the call backs are still not called, and the visitor v is still empty after running split method.

[sloriot]

You cannot put do_not_modify(true) for the first mesh. Then the signature of the visitor functions are incorrect, the last argument is a mesh. container.find(fd); might return container.end() which will lead to segfault if you access the iterator. On CGAL side, you'll need #8276 to get it working after fixing the other issues.

[citystrawman]

You cannot put do_not_modify(true) for the first mesh. Then the signature of the visitor functions are incorrect, the last argument is a mesh. container.find(fd); might return container.end() which will lead to segfault if you access the iterator. On CGAL side, you'll need #8276 to get it working after fixing the other issues.

Sorry I may not catch your words.
You mean Split() method should be written like this: PMP::split(mesh1, mesh2, params::visitor(v))? (I tried this, and it still does not work.)
Or do you mean that, the visitor is not working correctly until the issue is fixed?

[citystrawman]

You cannot put do_not_modify(true) for the first mesh. Then the signature of the visitor functions are incorrect, the last argument is a mesh. container.find(fd); might return container.end() which will lead to segfault if you access the iterator. On CGAL side, you'll need #8276 to get it working after fixing the other issues.

Hi, after reading this post, I changed split to clip method, and now before_subface_creations get called. At first before_subface_creations has code as follows:

void before_subface_creations(face_descriptor fd)
    {
        std::cout << "split : " << fd << " into:" << std::endl;
        Container::iterator it = container.find(fd);
        qfd = it->second;
        container.erase(it);
    }

which is copied from official code example.
However when I run the code, a face_descriptor which has the same idx_(in my example, f1134) run into this callback twice.
I am confused about it because I think face_descriptor should be unique and if a face_descriptor get called, the container will erase it, then why a face_descriptor that has the same idx_ run into the callback twice?

[sloriot]

Copy/paste of my answer here: You are mixing things up. The example you are pointing out is for triangulating a mesh, and the visitor must model the concept PMPTriangulateFaceVisitor. If you are using corefinement, the visitor must model the concept PMPCorefinementVisitor. In clip (and corefinement) you have 2 meshes so the same id might appear twice but with different meshes.

[sloriot]

Or do you mean that, the visitor is not working correctly until the issue is fixed?

Yes you need the aforementioned patch so that the visitor is correctly forwarded

[citystrawman]

Or do you mean that, the visitor is not working correctly until the issue is fixed?

Yes you need the aforementioned patch so that the visitor is correctly forwarded

Thank you, BTW I see that you just changed few code in #8276 as well as in #7966 , If I want to use this new feature now, can I just use the 5.5.x branch and just add the same code ?

[citystrawman]

Copy/paste of my answer here: You are mixing things up. The example you are pointing out is for triangulating a mesh, and the visitor must model the concept PMPTriangulateFaceVisitor. If you are using corefinement, the visitor must model the concept PMPCorefinementVisitor. In clip (and corefinement) you have 2 meshes so the same id might appear twice but with different meshes.

OK. So params::visitor(v) actually deals with both the tm as well as clipper/splitter even though it is associated as np_tm?