Urgent diagrid help

Replies to This Discussion

Permalink Reply by peter fotiadis on June 20, 2014 at 2:20am

Hmm...

...on first sight that's a task for LunchBox (you can do it "manually" of course but why bother?). If you decompose the brep you'll see that some surfaces are a bit "faulty" as well (tolerance issue?)

but ... the thing is that LunchBox doesn't work on trimmed surfaces - thus we need to "reconstruct" the whole thing into a set of untrimmed surfaces (we can't avoid some manual work, mind). On the other hand we have triangles that ... well you understand the issue.

allow me to think a bit more on this but if we had the definition that caused that thing it could be far easier

On the other hand the easiest way to do some untrimmed stuff ... doesn't work (points are in the correct order result is twisted, Karma without doupt).

more (and the definition) in a while . in the mean time get that Lunchbox

best, Peter

Permalink Reply by peter fotiadis on June 20, 2014 at 2:31am

BTW: Progress is made: Topologizer cleaned the "faulty" thing (compare number of lines per face on B6/7)

I think that putting this chaos into an Anemone loop could be the best way.

Permalink Reply by peter fotiadis on June 20, 2014 at 4:52am

OK, here's some WIP thing (data are internalized):

The general scope is to separate untrimmed/trimmed/bad/good faces in order to use Lunchbox.

Thus:

1. Some faces are bad (see bad girls group).

2. Topologizer is used (but could been omitted) in order to sort sardines from sardines.

3. Triangles and LunchBox don't mix very well (a rather easy thing to do some diagrid manually).

4. Most important: you can't apply the same (UV) paneling policy in all faces (BUT you can choose a "common" U or V value).

5. Due to 4 a loop is required in order to allow you to manage stuff interactively (Anemone - not added yet).

best, Peter

Attachments:

• diagrid_on_mass.gh, 62 KB

Permalink Reply by peter fotiadis on June 21, 2014 at 8:33am

Almost ready ... but the Anemone loop it doesn't behave as expected since it doesn't allow "live" U/V alterations (on a per "active" face basis - Item x from the face data tree).

My fault ... but where is the mistake? On the other hand I'm not a big fan of Hoopsnake for a variety of reasons (but I'll give it a try as Plan B).

Lunchbox has a very strange subdivision policy as well : try the various options provided (each choice triggers the loop to restart and redo the paneling up to the "current" face - this works as expected).

PS: internalize data option (input brep) occasionally it doesn't work. Have no idea why.

Moral: back to the drawing board.

Attachments:

• diagrid_on_mass5.gh, 131 KB
• diagrid_on_mass.3dm, 69 KB

Permalink Reply by KarlL on June 22, 2014 at 5:04am

Thanks peter ! This is rather a pickle !

Permalink Reply by peter fotiadis on June 22, 2014 at 8:38am

Indeed

I think that I know what's wrong with the faulty loop > tomorrow the final cut (*) - you'll need to treat the triangles individually for more than obvious reasons.

(*) a "manual/trad" option could be available as well (see this WIP thing operating in a rect grid: by combining RelItem A's and B's with a myriad of ways you could have any triangular panel "design"/combo possible)