如何使用特定的 colors 为矩阵着色？

Question

Here my minimal working example

arr0 = []
arr1 = []
fig, ax = plt.subplots()

provamatrix =  np.full((750, 750), -1)
arr0 = [random.randint(1,750) for _ in range(100000)]
np.random.shuffle(arr0)
arr1 = [random.randint(1,750) for _ in range(100000)]
np.random.shuffle(arr1)
coordsprova = list(zip(arr0, arr1))
m = 0
for c in coordsprova:
    provamatrix[c] = m
    m += 1
plt.imshow(provamatrix, cmap=plt.cm.Blues)

How can I color the pixels with a -1 value in black?

Answer 1

You can set an "under" color for the colormap. A range goes from vmin to vmax. When vmin to vmax aren't set explicitly (and no norm is used) they get calculated from the data.

Setting vmin to zero, makes all negative values to be 'under'. (There is a similar concept of an "over" color for values higher than vmax.)

A colorbar has an option to show the under color in a triangular arrow via plt.colorbar(extend='min') . Other values are 'max', 'both' and 'neither'. Related options can show the under color as a rectangle ( extendrect=True ) and set the fraction of the colorbar to be used for the extensions (default 5%: extendfrac=0.05 ).

Here is some example code to show how it works. Red is used instead of black to illustrate more clearly what's happening. (The code also tries to write the given code a bit more pythonic.)

from matplotlib import pyplot as plt
import numpy as np

provamatrix = np.full((750, 750), -1)
coordsprova = np.random.randint(0, 750, (30000, 2))
for m, (i, j) in enumerate(coordsprova):
    provamatrix[i, j] = m
cmap = plt.cm.get_cmap('Blues')
cmap.set_under('red')
plt.imshow(provamatrix, cmap=cmap, vmin=0)
plt.colorbar(extend='min', extendrect=True)
plt.show()