Smallest Angle Between closest neighbors?

Hi All,

I'm a bit stuck on this definition. Basically I'd like to find the smallest of the two angles between each strut(line) and it's two adjacent neighbors as they exist coming from each common node or base point. Having this angle would then ideally drive a distance parameter for each line - i.e. if a strut coming from index node (3) had neighboring struts at 20 degrees and 75 degrees to it's vector, then the 20 degree measurement would be associated with/drive a parameter for that specific strut (line). This would need to happen for each line as they come from each node. I used the Sandbox line topology component to get started, but I'm stuck on isolating for each specific strut and it's neighbors.

Any, help/advice would be appreciated!

Attachments:

Replies to This Discussion

Permalink Reply by ng5 Alex on October 21, 2014 at 12:42pm

Hello Matt

this creates a a data tree with branches that contain all neighboring lines. i believe that is what you are after. problem is i cant think how to measure angles for each group lines.

if you figure it out please post it here.

cheers

alex

*edit these groups are based on the nodes.

Permalink Reply by Matt Hutchinson on October 21, 2014 at 1:01pm

Thanks Alex, this definitely helps get the individual lines from Sandbox. I'll try to distill this further.....until then, anyone else?

Permalink Reply by ng5 Alex on October 21, 2014 at 1:24pm

needs further examination.

brb

Permalink Reply by Matt Hutchinson on October 21, 2014 at 1:45pm

that last image you posted gets close! I recreated it (several components I haven't used before!) but it seems some of the angles are on the large side (most should be smaller than 90) I think this is happening because not all the lines 'start' at the node/vertice - they might be starting at the neighboring point and ending at the node.

Permalink Reply by ng5 Alex on October 21, 2014 at 2:04pm

here it is. confirmed.

gives neighbor angles. the tree structure 0;0;0, 0;0;1,0;0;2, 0;0;3, 1;0;0, etc matches the structure of the L from the Line filter component ( {0} for the list of 4 lines, {1} for the list of 8 lines etc .

cheers

alex

Permalink Reply by Matt Hutchinson on October 23, 2014 at 11:32am

Alex (or anyone else for that matter)

I'm afraid this last solution is giving some mixed results. I'm not sure how to tweak it to get the desired results. I think that:

(A) that the lines coming from the nodes are not necessarily oriented in consistent directions - i.e. some point towards the common node while others point away.

and

(B) the index order of some of the lines coming from some of the nodes isn't necessarily ideal. For node (0) each line 0;0;0, 0;0;1,0;0;2, 0;0;3 is in order relative to it's neighbor, BUT if you go to node (1) the order of lines skips around. I've put in some reference angles within Rhino at some of the simpler intersections to check against the resulting GH angles and I think the previous reasons are causing the errors........

These two factors seem to be making it difficult, if not impossible to get the correct angle between each line and its neighbor. Is there another approach?

Attachments:

Permalink Reply by ng5 Alex on October 23, 2014 at 12:46pm

i believe now its sorted

the curves have been sorted radially and for each list of lines the start point is the same

cheers

alex

Attachments:

SCREEN_nodes_01_auto_variable(3).gh, 34 KB

Permalink Reply by Matt Hutchinson on October 23, 2014 at 2:18pm

WOW! I think think this should do the trick Alex! Thank you again!

Permalink Reply by ng5 Alex on October 23, 2014 at 2:36pm

You are welcome!

cheers

alex

ps if this is for a project i would love to see it when finished.

Permalink Reply by taz on October 23, 2014 at 2:52pm

Yes, nice work guys!

Matt, what direction are you thinking of taking this?

Permalink Reply by ng5 Alex on October 21, 2014 at 2:11pm

yes in that case i think the reflex angle is needed.

Permalink Reply by Matt Hutchinson on October 21, 2014 at 4:40pm

Thanks for the help Alex - greatly appreciated! I'll give this a try later tonight or tomorrow (working on something else right now!) Hopefully I can get this to drive the right parameters in a productive way!!