2D colormaps

Python

Use a two-dimensional colormap when one array supplies the lightness and a second array supplies the hue. For two measurements with matching coordinates:

import xarray as xr

import erlab.plotting as eplt

data_a, data_b = xr.align(data_a, data_b, join="exact")
intensity = data_a + data_b
asymmetry = ((data_a - data_b) / intensity).where(intensity > 0)

Plot the two quantities separately first. Check where low total intensity makes the normalized difference unstable:

import matplotlib.pyplot as plt

fig, axes = plt.subplots(
    1,
    2,
    figsize=(7.2, 3.0),
    layout="compressed",
    sharex=True,
    sharey=True,
)
intensity_image = eplt.plot_array(
    intensity,
    ax=axes[0],
    cmap="viridis",
    aspect="equal",
)
asymmetry_image = eplt.plot_array(
    asymmetry,
    ax=axes[1],
    cmap="bwr",
    norm=eplt.CenteredPowerNorm(1.0, vcenter=0.0, halfrange=1.0),
    aspect="equal",
)
eplt.nice_colorbar(ax=axes[0], mappable=intensity_image, width=7)
eplt.nice_colorbar(ax=axes[1], mappable=asymmetry_image, width=7)
eplt.set_titles(axes, ["Total intensity", "Normalized difference"])
eplt.clean_labels(axes)

(Source code)

Map total intensity to lightness and normalized difference to hue:

import matplotlib.pyplot as plt

fig, ax = plt.subplots(figsize=(4.8, 3.4), layout="compressed")

_, colorbar = eplt.plot_array_2d(
    intensity,
    asymmetry,
    ax=ax,
    lnorm=eplt.InversePowerNorm(0.5),
    cnorm=eplt.CenteredInversePowerNorm(0.7, vcenter=0.0, halfrange=1.0),
)
colorbar.ax.set_xticks(colorbar.ax.get_xlim(), labels=["Min", "Max"])
colorbar.ax.set(xlabel="Intensity", ylabel="Asymmetry")

(Source code)

Figure Composer

First calculate intensity and asymmetry with the Python code above. Open both arrays in ImageTool Manager and add them as Figure Composer sources. The separate diagnostic plots then use editable recipe steps:

1. Set Layout to a \(1 \times 2\) grid.

2. Add intensity and asymmetry in Sources.

3. Add one Image Plot step for each source and target one axes per step.

4. Set the intensity colormap to viridis.

5. Set the asymmetry colormap to bwr. Select CenteredPowerNorm, and set Gamma to 1. Under Center/range, set vcenter to 0 and halfrange to 1.

6. Add one ERLab Method step with nice_colorbar after each image.

7. Add set_titles and clean_labels ERLab Method steps.

The combined lightness-and-hue plot does not have an editable recipe step.

Planned Figure Composer support

Figure Composer does not yet have an editable step for this plotting operation. Structured support is planned. Until then, add a Python step to the recipe and use the code in this section.

1. Add data_a and data_b in Sources.

2. Add a Python step to a \(1 \times 1\) figure.

3. Review this code, then enter it in Code:

data_a, data_b = xr.align(data_a, data_b, join="exact")
intensity = data_a + data_b
asymmetry = ((data_a - data_b) / intensity).where(intensity > 0)

_, colorbar = eplt.plot_array_2d(
    intensity,
    asymmetry,
    ax=ax,
    lnorm=eplt.InversePowerNorm(0.5),
    cnorm=eplt.CenteredInversePowerNorm(0.7, vcenter=0.0, halfrange=1.0),
)
colorbar.ax.set_xticks(colorbar.ax.get_xlim(), labels=["Min", "Max"])
colorbar.ax.set(xlabel="Intensity", ylabel="Asymmetry")

Mask or otherwise handle points where the summed intensity is too small for a stable normalized difference. Confirm that data_a and data_b have aligned coordinates before calculating the asymmetry. The two-dimensional colorbar shows the independent lightness and hue mappings.

See erlab.plotting.plot_array_2d() for color normalization and colorbar arguments.