CGAL users discussion list

["Sebastien Loriot (GeometryFactory)" <>]

I didn't look at you code but the best way to do that is
to use the draw_dual member functions.

You can copy for example what is done in
include/CGAL/CGAL_Ipelet_base_v7.h
in draw_dual_in_ipe line 729 which is drawing the dual
of a triangulation restricted to an iso rectangle (you can even simplify
this code to do everything in the stream functor).

Sebastien.

memecs wrote:
> Hi, 
>
> I am trying to generate a bounded (by a rectangle) voronoi diagram. Here is
> my code: 
>
>
>         DT dt; 
>         dt.insert(points.begin(), points.end()); 
>
>         typedef typename K::Point_2 Point_2; 
>  	typename K::Ray_2 ray; 
>         typename K::Line_2 line; 
>         typename K::Segment_2 segment; 
>
>         vector<Polygon_2> bounded_voronoi_faces; 
>
>
>         // Find to which voronoi face the boundary points belong to. 
>         list<std::pair&lt;DT::Vertex_handle, Point_2> > nearest_vertices; 
>         for(int j = 0; j < 4; ++j) { 
>                 DT::Vertex_handle nearest =
> dt.nearest_vertex(bbox.vertex(j)); 
>                 nearest_vertices.push_back(std::pair<DT::Vertex_handle,
> Point_2>(nearest, bbox.vertex(j))); 
>         } 
>
>         // Iterate over all delaunay vertices. Each vertex is a site of a
> voronoi face. 
>         unsigned i = 0; 
>         for(DT::Finite_vertices_iterator fvit = dt.finite_vertices_begin(); 
>                         fvit != dt.finite_vertices_end(); ++fvit, ++i) { 
>
>                 
>                 // Circulate over all edges incident to the current vertex. 
>                 vector<Point_2> unique_points; 
>                 DT::Edge_circulator ec = dt.incident_edges(fvit); 
>                 do { 
>                         if(!dt.is_infinite(ec)) { 
>                                 CGAL::Object voronoi_obj, intersection_obj; 
>                                 voronoi_obj = dt.dual(ec); 
>                                 if(CGAL::assign(segment,voronoi_obj)) { 
>                                         intersection_obj =
> CGAL::intersection(segment,bbox); 
>                                 } 
>                                 else if(CGAL::assign(ray,voronoi_obj)) { 
>                                         intersection_obj =
> CGAL::intersection(ray,bbox); 
>                                 } 
>
>                                 else if(CGAL::assign(line,voronoi_obj)) { 
>                                         intersection_obj =
> CGAL::intersection(line,bbox); 
>                                 } 
>                                 else { 
>                                         // Something is wrong??? 
>                                         assert(false); 
>                                 } 
>                                 typename K::Segment_2 intersected_segment; 
>                                 Point_2 intersected_point; 
>                                
> if(CGAL::assign(intersected_segment,intersection_obj)) { 
>                                         // The object intersected 
>                                        
> unique_points.push_back(intersected_segment.source()); 
>                                        
> unique_points.push_back(intersected_segment.target()); 
>                                 } 
>                                 else { 
>                                         // If it's a point or didn't
> intersect do nothing... 
>                                 } 
>                         } 
>                 } while(++ec != dt.incident_edges(fvit));     
>
> The problem is: 
>
> sometimes I get  voronoi segments with source or target points equals to
> NaN. 
>
> For example: 
>
> if(CGAL::assign(segment,voronoi_obj)) { 
>   Delaunay segment: 490 39 495 29 
>   Voronoi segment: 493.5 34.5 -nan -nan 
> } 
>
>
>
>
>
> I really don't understand how is this possible. The input points I use to
> construct the delaunay triangulation don't overlap. 
>
> Does someone have an idea? 
>
> Thanks
>
> --
> View this message in context: 
> http://cgal-discuss.949826.n4.nabble.com/Voronoi-Diagram-and-NaN-tp4042568p4042568.html
> Sent from the cgal-discuss mailing list archive at Nabble.com.