Gaussian smoothing filter for UniformGrid

[imsodin]

Not much to say here: Adds the Gaussian smoothing filter to uniform grids.

For testing I just check that the data is changed at all and shape has not changed.

Also the scalars, preferences filter arguments and correspondingly setting the input of the vtk algo is duplicated so many times, and not always consistently - this should definitely happen in a wrapper function/decorator/...

[GuillaumeFavelier]

Except the typo, this looks good to me, great work! Also, it would be awesome to have an example to illustrate this feature in the gallery. I was thinking about something like:

import pyvista as pv

data = pv.read("lenna.png")
smoothed_data = data.gaussian_smooth(std_dev=4.)

p = pv.Plotter(shape=(1, 2))
p.subplot(0, 0)
p.add_text("Original Image", font_size=24)
p.add_mesh(data)
p.camera_position = 'xy'
p.subplot(0, 1)
p.add_text("Gaussian smoothing", font_size=24)
p.add_mesh(smoothed_data)
p.camera_position = 'xy'
p.show()

But I can take care of it in a follow-up PR if you want.

[GuillaumeFavelier]

- When scalars is specified, this is the perfered scalar type to
+ When scalars is specified, this is the preferred scalar type to

[imsodin]

That typo was copy-pasted (while it could have very well been mine :) ). I fixed it everywhere in filters.py.

[imsodin]

Given you have already done it, I'd prefer if you could take care of the example - thanks.

[banesullivan]

Great work! Here's a 3D example. In general, this might fit well with the volume rendering examples: https://docs.pyvista.org/examples/02-plot/volume.html

import pyvista as pv
from pyvista import examples

data = examples.download_brain()

smoothed_data = data.gaussian_smooth(std_dev=3.)


dargs = dict(clim=smoothed_data.get_data_range(),
             opacity = [0, 0, 0, 0.1, 0.3, 0.6, 1])

n = [100, 150, 200, 245, 255]

p = pv.Plotter(shape=(1, 2), notebook=0)
p.subplot(0, 0)
p.add_text("Original Image", font_size=24)
# p.add_mesh(data.contour(n), **dargs)
p.add_volume(data, **dargs)
p.subplot(0, 1)
p.add_text("Gaussian smoothing", font_size=24)
# p.add_mesh(smoothed_data.contour(n), **dargs)
p.add_volume(smoothed_data, **dargs)
p.link_views()
p.camera_position = [(-162.0, 704.8, 65.02),
                     (90.0, 108.0, 90.0),
                     (0.0068, 0.0447, 0.999)]
p.show()

The brain is a little more visible. The opacity mapping could probably be improved

[banesullivan]

Looks great to me! We can add an example in the gallery down the road