Surface from 3 and 4 point - data trees problem

Hello everyone!

I'm having little trouble trying to connect the intersection points of a grid to form a surface.

The grid itself is composed by triangular and quadrangular polygons, and the data trees of the intersection points is giving me some problem.

Here is the process i've done to come up with the grid (best viewing the image!):

1. create a base rectangular polygon;

2. find the diagonals and subdivide these, let's say in 10 points;

3. connect the points of one diagonal to the adjacent one in this order:

1 to 9
2 to 8
3 to 7
4 to 6
5 to 5
6 to 4
7 to 3
8 to 2
9 to 1

4. find the intersection points on every line drawn.

Now I would like to create individual surfaces connecting all the points that form the vertices of the polygons, but I cannot find out how...

Can somebody help me?

Thanks so much!

Attachments:

1.JPG, 123 KB
1.png, 551 KB
faurrasco.gh, 13 KB

Replies to This Discussion

Permalink Reply by Matteo Mandrile on May 1, 2013 at 1:51pm

anyone? :)

Permalink Reply by David Stasiuk on May 1, 2013 at 3:10pm

With your setup it will be pretty tricky to try to generate your surfaces from points/edges...but you can definitely use your curves to split a base surface into smaller ones. So if you make a bounding box for all of your lines (use the union box) convert your flat box to a surface and then use your lines to split it up, you'll have what you need. That said, you'll have four extra surfaces that you'll need to extract and your surfaces will be trimmed, but you can deal with this by extracting any of your trimmed surfaces that have more than four edges, and then you can rebuild untrimmed surfaces using points/edges from your trimmed surfaces. I don't know how important ordering is for, but this technique will give you surfaces at least.

Permalink Reply by Matteo Mandrile on May 1, 2013 at 4:10pm

Hi David,

that's a clever solution and works perfectly!

Any way, the next step would be identify the center point as a field attractor and have all the points of the surfaces moved in the vertical direction proportionally, based on the proximity to the center point. So I think that the problem will persist in any case.

If i found the way to connect the points by solving this data tree problem, then the extrusions and the solids will be constrained to that points, allowing me a complete freedom in further developement. That's what I'm looking for!

Do you know a way to solve this data trees?

Permalink Reply by Chris Tietjen on May 1, 2013 at 5:10pm

I would try solving it for a smaller number of points to see if you can identify a pattern.

Permalink Reply by David Rutten on May 1, 2013 at 5:41pm

I couldn't think of a better way either.

David Rutten

david@mcneel.com

Poprad, Slovakia

Attachments:

faurrasco.gh, 15 KB

Permalink Reply by David Rutten on May 1, 2013 at 5:54pm

I did reduce the logic of your file a bit in places. The file is attached but you won't be able to open it until 0.9.0050 is released as there is one component in there that isn't available yet.

David Rutten

david@mcneel.com

Poprad, Slovakia

Attachments:

faurrasco_reduced.gh, 14 KB

Permalink Reply by Vicente Soler on May 2, 2013 at 7:42am

The new list item component looks great. Hopefully it also has a "match outputs" like the explode tree.

I wonder if anyone ever uses on purpose the wrap input as false (in any of the components that has it). It would be great to have all wrap inputs as true by default. Ideally I would remove them all.

Only problems I can think of is either user error creating integers larger than the list length or for some conditional like "if -1 then something" which can still be used.

Permalink Reply by Danny Boyes on May 2, 2013 at 7:49am

OMG I've just ***

A [List Item] with a ZUI!

Permalink Reply by David Rutten on May 3, 2013 at 3:58am

I just changed all wraps defaults to true. You're right, I almost always have to change it or it doesn't matter. Wrapping is a more sensible default.

David Rutten

david@mcneel.com

Poprad, Slovakia

Permalink Reply by Matteo Mandrile on May 2, 2013 at 8:30pm

I've just realized your use of "flip Matrix"... As simple and effective as just a genius can do ;)

Permalink Reply by Matteo Mandrile on May 2, 2013 at 3:15pm

Thanks everyone guys!!

The solution of Systemig work great, but I do need also the central part of the geometry! ahahah

The Stasiuk/Rutten one instead seems to be the nearest one to what I'm looking for, but I'm having trouble to make it work... Do omeone know why?

The reduced one looks way better, so I'll wait the release of 0.9.0050 version to have a try.

Permalink Reply by Vicente Soler on May 2, 2013 at 3:27pm

You can use the reduced version, just replace that particular component with 4 list item components (as you've got in your definition), or use a flip matrix followed by an explode tree component.