Clean Definition and multiple curve to work building code parametrize

Replies to This Discussion

Permalink Reply by Toussaint Jimenez Rojas on March 28, 2012 at 9:40am

:(

Permalink Reply by Hannes Löschke on March 28, 2012 at 3:35pm

If you want to use multiple inputs, flatten is always a bad idea. You need to maintain each input on a separate branch.

As your def is pretty complex anf has lots of flattening, maybe you should start to restructure it, so you can live without the flatten.

Permalink Reply by Toussaint Jimenez Rojas on March 28, 2012 at 4:37pm

first i need to solve the problem of list item because when i select other  curve in other position doesnt work cuz the segments of lines have other position and are arranged diferent, first i need to evaluate the position of the polygon selected then start from this point to project the intersection corner then subdivide the line to construct the inside polygons.

i need some interrogations,

1- another clean and short way to fillet my polygons corners and find the intersection point 

    to start the subdivision. 

2- theres a way to draw a line through 4 point to make a closed polygon different that what i used to.  

list item doesnt work for me now, maybe i neet to start to evaluate a custom C# or VB# to achieve my goals. 

comments will be helpfull   this is the first issues of my Master Thesis project, and i'm not an expert in GH. 

http://aircrossover.wordpress.com/ 

Permalink Reply by Hannes Löschke on March 29, 2012 at 3:07am

I started to work through your def. but really the first problem is the fillets. I tried to patch and untrim the curves but the original face seems to be skewed.

Maybe you could reconstruct the voronoi points that build the region and cull the farther points (that would account to the fillets.)

Permalink Reply by Toussaint Jimenez Rojas on March 29, 2012 at 6:06am

you are right the curves are skewed, this is because the curves are real city block data from GIS MAPS, so in the real life city blocks are not straight.

i dont know anything about voroni points in GH :(

tonight i will going to check it to see what can i figure out

 

Permalink Reply by Hannes Löschke on March 29, 2012 at 11:30am

Such an intriguing problem... :D

This is how Voronoi reconstruction works...

Attachments:

• GALIBOSDEFINITION_HL_Voronoi_reconstruction.gh, 14 KB

Permalink Reply by Toussaint Jimenez Rojas on March 29, 2012 at 2:56pm

now when i arrive at home i will going to poke into the re-definition of my defintion to see whats going on with the internal logic, 

Thanks Hannes, but you are right that i will going to interrogate you once again when i cant figure out what the componnet its doing 
for example i dont understand well the pathmap   

Permalink Reply by Hannes Löschke on March 29, 2012 at 3:05pm

Path mapper is basically doing what flatten did. Just a level of data further. It flattens the last branch in the tree. Just plug a panel before and after the Mapper.

{A;B} -> {A} takes a tree like

{0,0} (0)

{0,1} (0)

{0,2} (0)

{1,0} (0)

{1,1} (0)

{1,2} (0)

and creates a tree

{0} (0, 1, 2)

{1} (0, 1, 2)

So flatten, but for each base branch seperately...

I use simplify quite a lot to keep trees at the smalles size.

Permalink Reply by Toussaint Jimenez Rojas on March 29, 2012 at 4:24pm

i'm touching the definition why the polygon center shows 3 center points ? this is an error ? 

Permalink Reply by Toussaint Jimenez Rojas on March 29, 2012 at 5:13pm

don't answer now i know what happend until now 

Permalink Reply by Toussaint Jimenez Rojas on March 29, 2012 at 5:59pm

error the voroni componet not working well with skew polygon

the find closet points its failing because just found 1 point and no 4 like the others polygons :(  so easy to be real :@

imagine that i have to evaluate 8 zones with aprox 8 - 20 squares  
i'm death :( 

Permalink Reply by Hannes Löschke on March 30, 2012 at 5:40am

Problems will arise if a long side doesn't simplify to a single line. Then you end up with a 2 or more lines essentially creating the same point/side in the rectangle. You should set the angle tolerance in the simplify to a reasonably high value like 10 degrees to catch slightly jagged edges.

Don't know how the tolerance values for the SimplifyCurve should be set. Just try to play with them...

I inserted a second culling stage, that sorts out short curves so the fillets wouldn't create points to begin with.

You can insert a third stage of checks if you remove duplicate points before the ClosestPts with a reasonable radius of tolerance, because points close together are likely to be created by an edge that wasn't simplified to a single line.

Attachments:

• GALIBOSDEFINITION_HL_Voronoi_reconstruction_2.gh, 17 KB

• ‹ Previous
• 1
• 2
• Next ›
• Page