CGAL users discussion list

[Nico Kruithof <>]

The datasets would be great for performance tests (uniform points in a rectangle are so boring). I still had problems with degenerate cases that I need to solve first. I'll contact you when I think that I have something working.

Nico

On Tue, Oct 9, 2012 at 8:22 PM, Gabriel Redner <> wrote:

Hi Nico,

Thanks for your message!  Currently my needs are fairly simple - I
just need to insert points into a Delaunay triangulation and then
iterate over the edges.  The extent of my interaction with CGAL looks
like:

  typedef CGAL::Exact_predicates_inexact_constructions_kernel Kernel;
  typedef CGAL::Triangulation_vertex_base_with_info_2<...stuff...,Kernel>
VertexBase;
  typedef CGAL::Triangulation_face_base_2<Kernel> FaceBase;
  typedef CGAL::Triangulation_data_structure_2<VertexBase,FaceBase>
TriangulationDataStructure;
  typedef CGAL::Delaunay_triangulation_2<Kernel,TriangulationDataStructure>
Triangulation;
  typedef typename Triangulation::Point Point;
  typedef typename Triangulation::Vertex_handle VertexHandle;
  typedef typename Triangulation::Face_handle FaceHandle;
  typedef typename Triangulation::Edge Edge;
  typedef typename Triangulation::Edge_iterator EdgeIterator;

  Triangulation t;
  for (...stuff...) {
    Vertex h = t.insert(Point(...stuff...));
    h->info() = ...stuff...;
  }

  for (EdgeIterator it = t.edges_begin(); it != t.edges_end(); ++it) {
    const Edge e = *it;
    const FaceHandle f = e.first;

    const VertexHandle v1 = f->vertex(t.cw(e.second));
    const VertexHandle v2 = f->vertex(t.ccw(e.second));

    do_stuff(v1->info(), v2->info());
  }

I do not currently anticipate needing anything more complex than this
(i.e. no removal etc.).

I'd be happy to provide data sets if it would be helpful.  Let me know
what range of sizes you are interested in and I can pull something out
of my simulation data.

I don't need the functionality anytime in particular.  I am currently
inserting image points by hand in order to emulate a periodic
triangulation, so whenever your code is available it will just enable
me to rip out this hack.

Thanks,
-Gabe

On Tue, Oct 9, 2012 at 3:50 AM, Nico Kruithof <> wrote:
> Hi,
>
> I'm working on the two dimensional triangulation in my spare time, which is
> rather sparse at the moment. What kind of operations do you need? Instertion
> in a Delaunay triangulation? Do you also want to (re-)move points? I haven't
> thought about a regular triangulation as it is more difficult. Would it be
> possible to get some representative data sets? Can I contact you when I have
> the requested functionality, even if it might take a while or do you need it
> soon?
>
> Nico
>
>
> On Friday, September 21, 2012, Monique Teillaud wrote:
>>
>> Le 21/09/12 09:50, Daniel Duque a écrit :
>>>
>>> On Thursday 20 September 2012 21:59:54 gredner wrote:
>>>>
>>>> Hi folks,
>>>>
>>>> I've been a happy user of CGAL's 2D Delaunay triangulation code for some
>>>> time.  However, I'm now in need of code for creating 2D *periodic*
>>>> triangulations, and I was very surprised to find that CGAL supports such
>>>> constructions in 3D but not in 2D!  So I'm wondering:
>>
>>
>>> Afaik, currently have to implement a 3D periodic triangulation with all
>>> of
>>> your points on one plane. It's a waste of space, I know.
>>
>>
>> Indeed, it is a waste of space, and of time as well: if you use the 3D
>> periodic triangulation for points that all lie in a plane, then you are in a
>> 'bad' case when you can never stop computing in the 27-sheeted covering
>> space.
>> So, you always compute with 27 copies of each point...
>>
>> Also, using the dummy points in this very specific case would have bad
>> effects on the resulting triangulation.
>>
>> --
>> Monique Teillaud
>> INRIA Sophia Antipolis - Méditerranée
>> http://www.inria.fr/sophia/members/Monique.Teillaud/
>>
>> --
>> You are currently subscribed to cgal-discuss.
>> To unsubscribe or access the archives, go to
>> https://sympa.inria.fr/sympa/info/cgal-discuss
>>
>>
>
>
> --
> Nico Kruithof
> nghk.nl
>

--
You are currently subscribed to cgal-discuss.
To unsubscribe or access the archives, go to
https://sympa.inria.fr/sympa/info/cgal-discuss

--
Nico Kruithof

nghk.nl