STL import

[mkeilman]

This may end up being part of the Phase 2a but the initial steps are not too complicated so I'll make this issue for it. The plan is to copy the warpvnd interface

(edited to mention that approach won't work here, because we have to let Radia create the object and then display it, not feed an arbitrary stl file to Radia)

• simple import (Import simple convex stl file  #4829)
• GUI (20221013 GUI for STL objects  #4848)
• complex import with user feedback for invalid files or concave objects
• tests for good and bad objects (20221019 STL import tests  #4849)

[mkeilman]

This issue should include using STL imports as individual magnet components, not just an entire simulation

[mkeilman]

I don't see a specific "import stl" function in Radia - the most generic thing is RadObjMltExtPgn which builds an object up from stacked convex polygons. Actually RadObjPolyhdr is a better fit for this problem

[mkeilman]

We have had problems getting ObjPolyhdr to work. For example here is some code to test the simplest possible polyhedron (a tetrahedron):

import radia
import numpy
p = [[0, 0, 0], [1, 1, 0], [0, 1, 0], [0.5, 0.5, 1]]
f = [[0, 1, 2], [0, 1, 3], [0, 3, 2], [3, 1, 2]]
g_id = radia.ObjPolyhdr(p, f)
d = radia.ObjDrwVTK(g_id, "Axes->No")
numpy.array([round(x, 6) for x in d["polygons"]["vertices"]]).reshape(-1, 3).tolist()
[
    [1.0, 0.0, 0.0], [0.0, 0.0, 0.0], [1.0, 1.0, 0.0],
    [0.0, 1.0, 0.0], [0.0, 0.0, 0.0], [1.0, 0.0, 0.0],
    [1.0, 1.0, 0.0], [0.0, 1.0, 0.0], [1.0, 0.0, 0.0],
    [1.0, 1.0, 0.0], [0.0, 0.0, 0.0], [0.0, 1.0, 0.0]
]

The generated vertices are missing points supplied to the function and include points not supplied. They are also all in the same plane. The function also sometimes causes a segfault. We are exploring using ObjMltExtPgn - which would be an approximation - as an alternative.

@cchall or @j-edelen, have you had success using ObjPolyhdr?

[cchall]

I've been able to use ObjPolyhdr in the past to do very similar things.
I'll take a look at this example and see what I can find.

[cchall]

I think it was just the face indexing. Radia uses 1-based indexing.

import radia
import numpy
p = [[0, 0, 0], [1, 1, 0], [0, 1, 0], [0.5, 0.5, 1]]
f = numpy.array([[0, 1, 2], [0, 1, 3], [0, 3, 2], [3, 1, 2]]) + 1
g_id = radia.ObjPolyhdr(p, f.tolist())
d = radia.ObjDrwVTK(g_id, "Axes->No")
numpy.array([round(x, 6) for x in d["polygons"]["vertices"]]).reshape(-1, 3).tolist()
[[0.0, 1.0, 0.0],
 [1.0, 1.0, 0.0],
 [0.0, 0.0, 0.0],
 [0.0, 0.0, 0.0],
 [1.0, 1.0, 0.0],
 [0.5, 0.5, 1.0],
 [0.0, 0.0, 0.0],
 [0.5, 0.5, 1.0],
 [0.0, 1.0, 0.0],
 [0.5, 0.5, 1.0],
 [1.0, 1.0, 0.0],
 [0.0, 1.0, 0.0]]

[mkeilman]

Oh, for goodness sake. @Benbenbenin0 does that info help?

[Benbenbenin0]

Yup, feature works...
Thank you @cchall

[mkeilman]

All sub-tasks completed