Scalar ranges for colors on a 3-D object. about pyvista-support

[banesullivan]

Hi @gunjan71011! For this situation, you would need to make your own colormap using Matplotlib and pass that colormap as the cmap argument.

Perhaps try an example like this: https://matplotlib.org/gallery/images_contours_and_fields/custom_cmap.html

And then pass it to PyVista similar to this example: https://docs.pyvista.org/examples/02-plot/cmap.html

FYI: there is a clim argument for the add_mesh method so you don’t need to use update_scalar_bar_range()

_{Sent with GitHawk}

[gunjan71011]

Thanks for the clim argument tip! So I payed around with customizing and making my own cmap and putting it into pyvista, however, I still can't get the colors to exactly be between a range of values, more specifically I want it to be between certain ranges of the scalar values used to scale my figure. I can create a colorbar with the exact values I want using matplotlib just fine, but i think you can only use color maps on pyvista using the cmap argument. Would you have any more feedback? Thank you!

[banesullivan]

Could you provide your code to make the matplotlib colormap?

[gunjan71011]

from matplotlib.colors import ListedColormap
from matplotlib import cm
import numpy as np

viridis = cm.get_cmap('viridis', 256)
newcolors = viridis(np.linspace(0, 1, 256))
blue = np.array([12/256, 238/256, 246/256, 1])
black = np.array([11/256, 11/256, 11/256, 1])
grey = np.array([189/256, 189/256, 189/256, 1])
yellow = np.array([255/256, 247/256, 0/256, 1])
red = np.array([1,0,0,1])
newcolors[0] = black                    #If my scaler is 0 i want it to be black in color
newcolors[1:30] = blue                 #If my scaler is between 1 and 30 I want it to be blue in color
newcolors[30:55] = yellow
newcolors[55:80] = grey
newcolors[80:] = red                      #Above 80 I want it to be red. The thing is I don't think these numbers actually represent the scalar values. But I would like it to.
newcmp = ListedColormap(newcolors)

Got this idea from:
https://matplotlib.org/3.1.1/tutorials/colors/colormap-manipulation.html

[banesullivan]

Here's an example - the colormap needed a bit more work:

import pyvista as pv
from pyvista import examples
from matplotlib.colors import ListedColormap
from matplotlib import cm
import matplotlib.pyplot as plt
import numpy as np
pv.set_plot_theme('doc')

mesh = examples.download_st_helens()
mesh['values'] = pv.plotting.normalize(mesh['Elevation']) * 100

# Make the custom colormap
blue = np.array([12/256, 238/256, 246/256, 1])
black = np.array([11/256, 11/256, 11/256, 1])
grey = np.array([189/256, 189/256, 189/256, 1])
yellow = np.array([255/256, 247/256, 0/256, 1])
red = np.array([1,0,0,1])

mapping = np.linspace(mesh['values'].min(), mesh['values'].max(), 256)
newcolors = np.empty((256, 4))

newcolors[mapping >= 80] = red
newcolors[mapping < 80] = grey
newcolors[mapping < 55] = yellow
newcolors[mapping < 30] = blue
newcolors[mapping < 1] = black
mycolormap = ListedColormap(newcolors)

# Plot with Matplotlib for a sanity check
image = mesh['values'].reshape(mesh.dimensions[:2], order='f')
plt.pcolormesh(image, cmap=mycolormap)
plt.colorbar()

# Plot with PyVista in 3D!
mesh.warp_by_scalar('Elevation').plot(scalars='values', cmap=mycolormap)

[banesullivan]

Please note that you likely won't be able to plot your mesh with Matplotlib - that only works if the input dataset is a 2D pyvista.UniformGrid object

[banesullivan]

FYI: I added this to the example gallery: https://docs.pyvista.org/examples/02-plot/cmap.html#custom-made-colormaps

[gunjan71011]

would you know how can fix the issue to normalize my data? The line "mesh['values'] = pv.plotting.normalize(mesh['Elevation']) * 100" keeps giving the error: module 'pyvista.plotting' has no attribute 'normalize'

[banesullivan]

Hm - you must be on an older version of PyVista. That's a simple function that you could just copy/paste:

def normalize(x, minimum=None, maximum=None):
    if minimum is None:
        minimum = np.nanmin(x)
    if maximum is None:
        maximum = np.nanmax(x)
    return (x - minimum) / (maximum - minimum)

But note that you will not need that normalizing function when you do this with your data - I simply used that to make an example with similar ranges to yours.

[gunjan71011]

Thanks Bane, that was really helpful!