point clustering along surface

Hi Folks,

I realize there are much easier methods for this (script, Lunchbox), but something has been bugging me. I wanted to test out a simple logic using native GH components.

The idea is to use the center points from the Isotrim surfaces along with the vertices (after the duplicates have been culled) as the searchable cloud, against the center point of the vertical edges from the Iso.

Searching for (4) closest points should give me the info I need to build the panels (well, half of them).

It works to a certain degree if the proportions are right. But why would only a portion of the points be sorted in a different order than the rest? The image below shows the only two (out of 200 in this case) inverted surfaces due to incorrect sorting. Definition is also attached for reference. Curious.

Attachments:

pt_cluster_test.gh, 16 KB

Replies to This Discussion

Permalink Reply by mire on November 21, 2014 at 5:17am

Hi Nick, not sure about the default order, but you can do the sorting yourself:

Permalink Reply by Nick Bruscia on November 22, 2014 at 1:34pm

Thanks mire. Nice idea to sort the sets of points by their interior angles when directed toward their center. This seems to work nicely.

It seems self evident how GH sorts the points by default - likely in order from closest to farthest, which cannot be the same for all cells by default. I was just wondering if this were true, but I suppose after sorting it doesn't matter.

RSS