Optimize Curve Division

Replies to This Discussion

Permalink Reply by nikos tzar on August 19, 2015 at 1:17pm

Hello Jose,

[DivideDistance] only works with a single Distance input. So when you use a list of numbers it actually divides the curve many times (that's why you get a tree as output) and that's why you get those "strange" distances betwwen points.

See this post as well: http://www.grasshopper3d.com/forum/topics/divide-curve-multiple

I am afraid the only way to do this would be to use a loop plugin such as anemone or hoopsnake, or maybe you could use kangaroo...

(something similar to this but on the same curve)

Permalink Reply by David Rutten on August 19, 2015 at 1:21pm

If you're willing to settle for dividing by length rather than distance, you can use Dash Pattern, it accepts a list of dash and gap lengths.

Permalink Reply by Jose Miguel Armijo on August 19, 2015 at 2:04pm

Dear David and Nikos

When I use Dash pattern I got a lot of approximates, this could work for now.

But I will try Nikos Suggestion after installing Hoopsnake

Thank you both

Permalink Reply by nikos tzar on August 20, 2015 at 4:08am

Hello Jose,

I believe the approximates are because of the difference of dividing by distance vs dividing by length.

When you divide a curve by a distance d, then the straight lines from each point to the next will have a length of d.

When you divide a curve by a length d, then the curve fragments between two consecutive points will have a length of d.

Divide by length is an easier operation and you can actually find the optimal solution just by using math operations (you can use [mass addition] and use the "Partial Results" output to find the largest number that is smaller than the curve's length).

Permalink Reply by Vicente Soler on August 20, 2015 at 10:41am

Since you are trying to find the best seed value, as I understand, there is no advantage in using Galapagos over just generating a number of random sequences and keeping the best one.

See the attached definition for a script of this. It should also work way faster than using Galapagos.

Permalink Reply by Vicente Soler on August 20, 2015 at 11:15am

The previous script didn't account for the case when more than one point lies on the same polyline segment.

Attachments:

• post02gh.gh, 11 KB

Permalink Reply by Vicente Soler on August 22, 2015 at 3:24am

...was this useful?

Permalink Reply by Jose Miguel Armijo on September 14, 2015 at 12:57pm

Hello Vicente

I didn't have the chance to use it, I took a break from this subject, but right now I giving a try and it looks very promising,

I Think it's great, but it's ,more helpful if it's an induced pattern, like when it's has more curvature the segments have lower lengths. 

On a side note, how do you learn c#? I definitively want to learn to code like that!

Thank you for your help!