C# A triangulation puzzle

Well friends,

This thread is a "subset" of that:

http://www.grasshopper3d.com/forum/topics/c-delaunay-solver-syntax

But you don't actually need to come across that thread: just read the comments inside the definition attached here.

Initially I thought that a Delauney type of triangulation could be the answer, but appears that is not (white lines), nor the ball pivoting script (as it is):

In order to create some proper code I think(?) that I have(???) outlined the right pseudo code ... the "only"(?) problem is to define the "ideal" starting point (for instance in the above image : point 2). Sorting properly points could be a starting approach but multi sort(s) appear not working (screenshot from the actual definition):

Any bright idea, welcome

best, Peter

Attachments:

Tens_from_random_polyline_v4b_the triangulation_puzzle_v3.gh, 55 KB

Replies to This Discussion

Permalink Reply by peter fotiadis on October 6, 2014 at 6:17am

PS: October must be my stupid month : sort works, so I'm going to attempt my triangulation approach

Added some Z "noise" in the test case (ball point goes bananas)

Attachments:

Tens_from_random_polyline_v4b_the triangulation_puzzle_v4.gh, 61 KB

Permalink Reply by Vicente Soler on October 8, 2014 at 11:05am

If you already have the concave hull you can check the face centers of the Delaunay triangulation for inclusion and cull them accordingly.

Permalink Reply by peter fotiadis on October 9, 2014 at 1:52am

Well... as I said I've opted to do my way: In fact, somehow, it's more complex than a Delaunay triangulation ... since the logic behind the triangulation is not only the classic one (proximity, topology, ... checks) but also the "shape" of the resulting triangle matters as well (remember that defines a cone (in a "sub membrane") that should have some kind of "substance").

Anyway my code is almost(*) ready:

(*) some functions checking the validity of choices are still WIP > see the second triangle from the left that is quite "narrow" and must replaced by something more "equalized" as regards the edge angles.

I'll post here the finished C# soon (highly case specific: it doesn't deal with any collection of points, only with points that have a meaning for the membrane design and their random distribution within the "boundaries" of that type shown).

best, Peter

Permalink Reply by peter fotiadis on October 17, 2014 at 1:02am

Update:

Project abandoned for a while, but now is almost ready. It works using solely C# and no GH component of any kind. I'll post here the thing soon.

Weak point: my "uniform" random point generation (for the mast bases) is way inferior to the one used in the PopGeo component (obviously custom code by David).

I would be clad to get some hints about how the PopGeo works (or ... er.... hmm... the C#, if this is not way beyond the meaning of " reasonable assistance").