How to use the mlab iso_surface module in a Mayavi app - enthought

Question

I'm trying to build a simple Mayavi script application which utilises the mlab iso_surface module.
However, when I run my app it throws up two windows, one showing my mayavi iso_surface plot and the other showing a blank "Edit properties" window. It seems that the mayavi scene is not being displayed in the specified view layout for the "Edit properties" window.
So my question is: Why is the mayavi iso_surface scene not appearing in the view layout, and how do I get it in there?
A simple test script which displays this behaviour is pasted below. I am using Canopy version: 2.1.1.3504 (64 bit), python 3.5.2 on a Windows 10 system.
[Note: I have modified my original question to include another question. How do I update the 's' data with the input from a Range object (mult_s)? I have had a go at doing this below, but with no success. It throws up: TraitError: Cannot set the undefined 's' attribute of a 'ArraySource' object.]
class Isoplot1(HasTraits):
scene = Instance(MlabSceneModel, ())
mult_s = Range(1, 5, 1)
#on_trait_change('scene.activated')
def _setup(self):
# Create x, y, z, s data
L = 10.
x, y, z = np.mgrid[-L:L:101j, -L:L:101j, -L:L:101j]
self.s0 = np.sqrt(4 * x ** 2 + 2 * y ** 2 + z ** 2)
# create the data pipeline
self.src1 = mlab.pipeline.scalar_field(x, y, z, self.s0)
# Create the plot
self.plot1 = self.scene.mlab.pipeline.iso_surface(
self.src1, contours=[5, ], opacity=0.5, color=(1, 1, 0)
)
#on_trait_change('mult_s')
def change_s(self):
self.src1.set(s=self.s0 * self.mult_s)
# Set the layout
view = View(Item('scene',
editor=SceneEditor(scene_class=MayaviScene),
height=400, width=600, show_label=False),
HGroup('mult_s',),
resizable=True
)
isoplot1 = Isoplot1()
isoplot1.configure_traits()

Answer 1

If you use self.scene.mlab.pipeline.scalar_field instead of mlab.pipeline.scalar_field this should not happen.
In general, you should avoid creating any visualization in the initializer. Instead you should always setup the scene when the scene.activated event is fired. To be safe for uses with raw mlab you should rewrite your code as follows.
from mayavi import mlab
from traits.api import HasTraits, Instance, on_trait_change
from traitsui.api import View, Item
from mayavi.core.ui.api import MayaviScene, SceneEditor, MlabSceneModel
import numpy as np
class Isoplot1(HasTraits):
scene = Instance(MlabSceneModel, ())
#on_trait_change('scene.activated')
def _setup(self):
# Create x, y, z, s data
L = 10.
x, y, z = np.mgrid[-L:L:101j, -L:L:101j, -L:L:101j]
s = np.sqrt(4 * x ** 2 + 2 * y ** 2 + z ** 2)
# create the data pipeline
self.src1 = mlab.pipeline.scalar_field(x, y, z, s)
# Create the plot
self.plot1 = self.scene.mlab.pipeline.iso_surface(
self.src1, contours=[5, ], opacity=0.5, color=(1, 1, 0)
)
# Set the layout
view = View(Item('scene',
editor=SceneEditor(scene_class=MayaviScene),
height=400, width=600, show_label=False),
resizable=True
)
isoplot1 = Isoplot1()
isoplot1.configure_traits()
You probably already know this but just in case you can also take a look at some of the other mayavi interactive examples in the documentation.