What's the fastest way to calculate the midpoints of a sequence of edges?

Question

I'm trying to create a list that calculates the midpoints of a sequence of edges.

Right now I'm doing it like this:

eList = []
for e in [e for e in bm.edges if e.select]:
    pt_1, pt_2 = (ob.matrix_world * e.verts[0].co), (ob.matrix_world * e.verts[1].co)
    addPt = pt_1 + pt_2
    co = mathutils.Vector((addPt[0]/2,addPt[1]/2,addPt[2]/2 ))
    eList.append((e, co))

I multiply each vert in the edge by the object's world matrix, then add the verts together. I divide each axis in the vert by 2 and covert them to a vector and then put the edge and the vector in a tuple and append that to a list.

It's probably not a very efficient process but I'm still pretty new to python and can't think of a faster way.

Is this something that could be done faster in numpy?

What would be the fastest way to perform this function over a large sequence of edges?

Answer 1

Some improvements.

• Use list comprehension
• Only need to multiply the local average edge coord by matrix world once.
• Multiply / divide a vector by a scalar, rather than on a per component basis.

Test script

import bpy
import bmesh

ob = bpy.context.edit_object
me = ob.data
bm = bmesh.from_edit_mesh(me)

mw = ob.matrix_world
elist = [(e, mw * (e.verts[0].co + e.verts[1].co) / 2) 
        for e in bm.edges if e.select]

could make a little helper routine.

def mid_edge(e):       
    return (e.verts[0].co + e.verts[1].co) / 2

elist = [(e, mw * mid_edge(e)) 
        for e in bm.edges if e.select]