matplotlib: specify default colors explicitly

Question

When I do

ax.plot(...)
ax.plot(...)
...

the color of each successive plot is in the default color cycle (blue, orange, green &c).

Alas, sometimes I need to specify the colors explicitly in each call to ax.plot because each logical plot is generated by several such calls.

So, suppose I need to plot num_plots curves, and I want the color selection to be identical to the default one.

How do I get the list of colors that matplotlib would use if I call ax.plot exactly num_plots times?

I tried

import matplotlib.cm as colormaps
colors = colormaps.get_cmap(None, num_plots)(np.linspace(0,1,num_plots))

(and various names like "Accent" and "winter" instead of None that appear in the error message when I use random junk instead of None).

Answer 1

I am not sure if I understand your question correclty, but it seems like you are asking for the default color cycle which can be found here. You say that

@JodyKlymak: but then if I have many plots, some of them will be in the same color. I don't think this happens when I do not specify the color explicitly.

This is not a correct assumption. It is a color cycle, so the colors will repeat.

import matplotlib.pyplot as plt

num_plots = 20
for i in range(num_plots):
    plt.plot([0,1], [(i+1)/num_plots, (i+1)/num_plots])

plt.show()

If you want to cycle through the colors manually, just get the list of default colors and use the modulo operator as follows.

import matplotlib.pyplot as plt

prop_cycle = plt.rcParams['axes.prop_cycle']
colors = prop_cycle.by_key()['color']

num_plots = 20

for i in range(num_plots):
    plt.plot([0,1], [(i+1)/num_plots, (i+1)/num_plots], color=colors[i % len(colors)])

plt.show()