Slices in yt-napari

In addition to sampling region objects, the yt_napari plugin can sample 2D slices normal to the yt dataset axes. To get started, initialize an empty napari.Viewer, load in a yt dataset and initialize our yt_napari Scene:

import napari
import yt
from yt_napari.viewer import Scene

viewer = napari.Viewer()

ds = yt.load("IsolatedGalaxy/galaxy0030/galaxy0030")

yt : [INFO     ] 2023-07-07 13:35:37,106 Parameters: current_time              = 0.0060000200028298
yt : [INFO     ] 2023-07-07 13:35:37,107 Parameters: domain_dimensions         = [32 32 32]
yt : [INFO     ] 2023-07-07 13:35:37,108 Parameters: domain_left_edge          = [0. 0. 0.]
yt : [INFO     ] 2023-07-07 13:35:37,108 Parameters: domain_right_edge         = [1. 1. 1.]
yt : [INFO     ] 2023-07-07 13:35:37,109 Parameters: cosmological_simulation   = 0

yt_scene = Scene()

The add_slice method will sample a slice normal to one of the dataset axes. The arguments are similar to yt.SlicePlot:

yt_scene.add_slice(viewer,
                   ds,
                   'z',
                   ('enzo', 'Density'),
                   center = ds.domain_center,
                   resolution = (600, 600),
                   width = ds.quan(25, 'kpc'),
                   height = ds.quan(25, 'kpc')
                   )

Parsing Hierarchy : 100%|█████████████████████████████████████████████████| 173/173 [00:00<00:00, 17558.31it/s]
yt : [INFO     ] 2023-07-07 13:35:37,143 Gathering a field list (this may take a moment.)
yt : [INFO     ] 2023-07-07 13:35:38,527 Making a fixed resolution buffer of (('enzo', 'Density')) 600 by 600

from napari.utils import nbscreenshot
nbscreenshot(viewer)

useful applications

For simple visualizations, it’s likely easier to simply use the standard yt plotting API (yt.SlicePlot), but there are a number of uses for slices in napari.

field composites

yt_napari makes it easy to create composite images with different fields:

viewer.layers.clear() # first clear existing layers

yt_scene.add_slice(viewer,
                   ds,
                   'z',
                   ('enzo', 'Density'),
                   center = ds.domain_center,
                   resolution = (600, 600),
                   width = ds.quan(25, 'kpc'),
                   height = ds.quan(25, 'kpc'),
                   colormap='red',
                   contrast_limits=(3.4, 6.4)
                   )

yt_scene.add_slice(viewer,
                   ds,
                   'z',
                   ('enzo', 'Temperature'),
                   center = ds.domain_center,
                   resolution = (600, 600),
                   width = ds.quan(25, 'kpc'),
                   height = ds.quan(25, 'kpc'),
                   colormap = 'yellow',
                   contrast_limits=(4., 6.),
                   blending='additive',
                   )

yt : [INFO     ] 2023-07-07 13:35:38,991 Making a fixed resolution buffer of (('enzo', 'Temperature')) 600 by 600

from napari.utils import nbscreenshot
nbscreenshot(viewer)

variable resolution composites

You can also easily create composite images of different resolutions.

viewer.layers.clear() # first clear existing layers

First, let’s add a slice covering the whole domain:

yt_scene.add_slice(viewer,
                   ds,
                   'z',
                   ('enzo', 'Density'),
                   center = ds.domain_center,
                   resolution = (600, 600),
                   name = 'density_whole',
                   )

yt : [INFO     ] 2023-07-07 13:35:39,210 Making a fixed resolution buffer of (('enzo', 'Density')) 600 by 600

and zoom in:

viewer.camera.zoom = 10

and we see a pixelated view of the center of the density field:

from napari.utils import nbscreenshot
nbscreenshot(viewer)

Now we can add a higher resolution slice just at the center:

yt_scene.add_slice(viewer,
                   ds,
                   'z',
                   ('enzo', 'Density'),
                   center = ds.domain_center,
                   resolution = (600, 600),
                   width = ds.quan(25, 'kpc'),
                   height = ds.quan(25, 'kpc'),
                   name = 'density_high_res'
                   )

and normalize the color limits across layers (without linking them):

yt_scene.normalize_color_limits(["density_whole", "density_high_res"], viewer.layers)

to create a composite image with higher resolution in the region of interest without requiring a high resolution across the whole domain:

from napari.utils import nbscreenshot
nbscreenshot(viewer)

Limitations

At present, yt_napari will not attempt to align slices along different axes, so if you add a second slice along a different axis or add a 3D region, it will simply overlay as a new 2D image layer. You’ll be able to work with the layers in napari, but compositing or 3D views may be difficult or incorrectly aligned.