Cross Section, Bisect nodes and objects.

[Durman]

Is it possible to cut objects individually? I think this nodes unite objects in one inside itself. It will be better if one matrix cut first object and two matrix cut second object, for me in my case.))

https://gist.github.com/8b592141eba2da47443ba4e0690fcf6a

[zeffii]

@Durman the node doesn't join the objects, but it does perform all matrix cuts on each incoming object.

[zeffii]

look at the code to understand why

        for cut_mat in cut_mats:
            pp = cut_mat.to_translation()
            pno = Vector((0.0, 0.0, 1.0)) * cut_mat.to_3x3().transposed()
            for obj in zip(verts_ob, edg_pols):
                res = bisect(obj[0], obj[1], pp, pno, self.outer, self.inner, self.fill)
                if not res:
                    return
                verts_out.append(res[0])
                edges_out.append(res[1])
                polys_out.append(res[2])

[zeffii]

it makes sense to add a toggle to use the above algorithm as "All matrices" vs a mode that does "individual matrices".

[zeffii]

[zeffii]

I am reluctant to modify The cross section node too.. can't commit this just yet.

[zeffii]

[zeffii]

i've committed this to Bisect node.

[Durman]

Thanks Zeffii, very fast.))
Can I Use several cutting matrix for one object and several (may be more) for another? :/

[Durman]

I can look at the code but I can not understand nothing. My knowledge about Python too poor.

[zeffii]

@Durman ofcourse it can be done, but not with the current incarnation. Not sure what the interface should be for that (many options are possible). Maybe the most conventional is to take nested lists of Matrices...

matrix = [[v,v,v,v], [v,v,v,v], [v,v,v,v], [v,v,v,v]]

# normal matrix list 
matrices = [matrix1, matrix2, matrix3]

# nested matrix list
matrices = [[matrix1a, matrix1b, matrix1c], [matrix2a, matrix2b, matrix2c], [matrix3a, matrix3b, matrix3c]]

a bool/toggle for "nested" would iterate over objects, and then sub matrix list.

for idx, obj in enumerate(verts, faces):
   for mat in matrices[idx]:
       ... make chops..

[Durman]

When I tested bisect node I supposed that nested list of matrix received right result, for this I used List split node, but it did not help me.))

[Durman]

With Scale node this method has good work.

[zeffii]

only a few lines of code need to be added to do what you want @Durman - but i will think about it first before writing code. -- maybe tomorrow.

[zeffii]

as for nested Matrix lists.. by which means do you intend to produce them? the normal Matrix In doesn't exactly offer an easy way.

[Durman]

Good question.)) I thought something like this but it does not work I am not understanding why.

[zeffii]

it's possible to give Matrix In a "Multi Matrix" mode., which expects "complex" input (nested input). Without a convenient way to make nested Matrix lists... i don't see much reason to push the changes to Bisector node ( I already have a branch that should work on nested matrix lists ..and have not tested it yet)

[Durman]

ok

[Durman]

I did it with Scripted node lite and Python.))

[zeffii]

good work on solving your own problem!

[nortikin]

Yes, go ahead for more results. Can do!

[Durman]

I have reconstructed Bisect node. I have changed input matrix on two vectors inputs, first for moving section plane second for rotation. Now it is possible to mange with section plane for each input object.

'''
in vertices v
in edg_pols s
in move_matrix v d=[[(0,0,0)]] n=0
in rotate_matrix v d=[[(0,0,1)]] n=0
out verts_out v
out edges_out s
out polys_out s
'''
from mathutils import Vector
from sverchok.data_structure import Matrix_generate, Vector_generate, match_long_repeat
from sverchok.nodes.modifier_make.bisect import bisect
from sverchok.utils.sv_mesh_utils import mesh_join

def input_mat(io):
    def check_len(data, io):
        if len(data) > io: return io
        else: return -1
    return match_long_repeat([move_matrix[check_len(move_matrix, io)]] + [rotate_matrix[check_len(rotate_matrix, io)]])

verts_ob = Vector_generate(vertices)

for io,obj in enumerate(zip(verts_ob, edg_pols)):
    ln_e,ln_p = 0,0
    v_obj,e_obj,p_obj = [],[],[]

    for i,p_mat in enumerate(zip(*input_mat(io))):
        pp = Vector(p_mat[0])
        pno = Vector(p_mat[1])
        
        res = bisect(obj[0], obj[1], pp, pno, True, True, True)
        
        v_obj.extend([res[0]])
        e_obj.extend([res[1]])
        p_obj.extend([res[2]])
        
    v_obj,e_obj,p_obj = mesh_join(v_obj,e_obj,p_obj)
            
    verts_out.append(v_obj)
    edges_out.append(e_obj)
    polys_out.append(p_obj)

[zeffii]

@Durman i think it is great that you have managed to solve this for yourself. That's what SNLite was made for.

• I'm not going to modify the current bisect node to behave this way (maybe you, or someone else wants to... )
• Maybe there should be a more convenient way to store user-created SNLite scripts locally / online..
• Maybe a more convenient way to include custom nodes in a user's sverchok folder..

[zeffii]

I will think about this when time permits.

[nortikin]

cutter-bisector_2017_07_27_19_53.zip

[Durman]

@zeffii
I think logic of this node too unlike to another nodes of Sverchek.
@nortikin
Something strange.))

Where can I to find code of the delaunay node on the Sverchek store?

[zeffii]

@Durman , the delaunay node is located in the same file as 2dvoronoi ( https://github.com/nortikin/sverchok/blob/master/nodes/modifier_make/voronoi_2d.py )