Visualizing data using the plot_slice function

plot_slice is a function that is capable of plotting 2-D datasets on axes which need not be orthogonal. This was designed originally for plotting X-ray scattering datasets much in the way that the NeXpy package does.

Importing plot_slice

from nxs_analysis_tools import plot_slice, load_data
from nexusformat.nexus import NXdata, NXfield
import numpy as np

Import data

data_cubic = load_data('example_data/plot_slice_data/cubic_hkli.nxs')
data_hex = load_data('example_data/plot_slice_data/hex_hkli.nxs')

data:NXdata
  @axes = ['H', 'K', 'L']
  @signal = 'counts'
  H = float64(100)
  K = float64(150)
  L = float64(200)
  counts = float64(100x150x200)
data:NXdata
  @axes = ['H', 'K', 'L']
  @signal = 'counts'
  H = float64(100)
  K = float64(150)
  L = float64(200)
  counts = float64(100x150x200)

Basic plotting

Plot slice accepts an NXdata object which hold 2-D data. Thus, if working with a 3D scattering dataset, you must index one of the axes.

Here we plot the K=0 plane.

plot_slice(data_cubic[:,:,0.0])

<matplotlib.collections.QuadMesh at 0x7fa29e132650>

Transposing the X and Y axes

plot_slice(data_cubic[:,:,0.0], transpose=True)

<matplotlib.collections.QuadMesh at 0x7fa29c0f1b40>

Plotting on non-orthogonal axes

When working in a non-orthorhombic cell, it may be necessary to specify the skew angle between the axes of interest. Here, we consider a hexagonal crystal structure, for which H and K are 60$^\circ$ apart.

If we plot the data using no additional arugments, we see that the Bragg reflections are skewed.

plot_slice(data_hex[:,:,0])

<matplotlib.collections.QuadMesh at 0x7fa29be84ca0>

Use the skew_angle parameter to correct the angle between the plotted X and Y axes.

plot_slice(data_hex[:,:,0], skew_angle=60)

<matplotlib.collections.QuadMesh at 0x7fa29a509690>

Adjusting axis limits

Note that the x-axis limits are interpreted in a way that displays the full area covered by the specified x and y limits. Thus, when working with a skewed dataset, the actual tick marks on the x-axis will extend longer than the specified limits.

plot_slice(data_hex[:,:,0.0], skew_angle=60, xlim=(0,1), ylim=(0,1))

<matplotlib.collections.QuadMesh at 0x7fa29a408880>

Adjusting color mapping

plot_slice(data_cubic[:,:,0.0], vmin=0,vmax=1)

<matplotlib.collections.QuadMesh at 0x7fa29a2ec610>

plot_slice(data_cubic[:,:,0.0], vmin=0,vmax=0.5)

<matplotlib.collections.QuadMesh at 0x7fa29a1dcfa0>

plot_slice(data_cubic[:,:,0.0], cmap='magma')

<matplotlib.collections.QuadMesh at 0x7fa29a0dd330>

Using logscale

plot_slice(data_cubic[:,:,0.0], logscale=False)

<matplotlib.collections.QuadMesh at 0x7fa299fe0490>

plot_slice(data_cubic[:,:,0.0], logscale=True)

<matplotlib.collections.QuadMesh at 0x7fa299e6f880>

Using symlogscale

First, let’s create a test dataset with both positive and negative values

data_sym = load_data('example_data/plot_slice_data/sym_hkli.nxs')

data:NXdata
  @axes = ['H', 'K']
  @signal = 'counts'
  H = float64(100)
  K = float64(150)
  counts = float64(100x150)

Without the symlogscale, the data is plotted on a linear colormap. Both a vmin and vmax can be provided for the limits of the colormap.

plot_slice(data_sym, cmap='seismic', symlogscale=False, vmin=-10, vmax=10)

<matplotlib.collections.QuadMesh at 0x7fa29a28b6d0>

With symlogscale enabled, only the vmax parameter is used the colorbar is symmetric about zero.

plot_slice(data_sym, cmap='seismic', symlogscale=True, vmax=10)

<matplotlib.collections.QuadMesh at 0x7fa299d654e0>