[vtkusers] Extracting normals from polydata add points ?

[Nicolas Cedilnik]

Hello,

I'm wondering if I should file a bug for this or if it's a feature I 
don't understand.

I'm computing normals for a triangular mesh and for some reason the new 
polydata (with normals) ends up having more points than the original.

I don't really have time to explore this issue right now but if needed I 
do it later. Here's a quick illustration with my ipython output:

In [58]: polydata.GetPoints().GetNumberOfPoints()
Out[58]: 11290

In [59]:     normals_generator = vtk.vtkPolyDataNormals()
     ...:     normals_generator.SetInputData(polydata)
     ...:     normals_generator.ComputeCellNormalsOff()
     ...:     normals_generator.ComputePointNormalsOn()
     ...:     normals_generator.Update()
     ...:
     ...:     polydata_with_normals = normals_generator.GetOutput()
     ...:

In [60]: polydata_with_normals.GetPoints().GetNumberOfPoints()
Out[60]: 11382

So I have 92 new points, the 11290 first ones being equal:

In [74]: 
(vtk_to_numpy(polydata_with_normals.GetPoints().GetData())[:11290] == 
vtk_to_numpy(polydata.GetPoints().GetData())).all()
Out[74]: True

I'm using VTK with python3.6 from conda-forge:

$ conda list | grep vtk
vtk                       7.1.1                    py36_1 conda-forge

Thoughts?

-- Nicolas