Home > Software engineering >  plotting a boolean array as a translucent overlay over a graph with matplotlib
plotting a boolean array as a translucent overlay over a graph with matplotlib

Time:01-17

I want to plot the True parts of a boolean array as translucent boxes over another plot.

This sketch illustrates what I envision. I know I could do that with Asymptote, but I (among other reasons) need to verify that the data I work with is concise. I can supply example code of a graph and a boolean array if that helps - I don't have an idea yet how to realize the overlays, though. Asymptote might be the best option for producing plots for later publication, though.

sketch of a graph with boolean overlay

CodePudding user response:

To overlay boxes, you can use enter image description here

import matplotlib.pyplot as plt
import numpy as np
from matplotlib.collections import PatchCollection
from matplotlib.patches import Rectangle

# boolean and plot data
n_points = 100
x = np.linspace(0, 2.5 * np.pi, n_points)
my_truth = np.zeros(n_points)
# true regions must be surrounded by false, pad by 1 false if needed
my_truth[20:40] = 1
my_truth[60:70] = 1


def get_truth_intervals(logical_data):
    """ extract 'true' regions embedded in 'false' regions """
    truth_spikes = np.diff(logical_data)
    truth_starts = np.argwhere(truth_spikes == 1)
    truth_ends = np.argwhere(truth_spikes == -1)
    return truth_starts, truth_ends


with plt.xkcd():
    fig = plt.figure()
    ax = plt.gca()
    ax.plot(x, np.sin(x))

    # draw boxes defined by true sections and plot height
    y_start, y_end = ax.get_ylim()
    boxes = [Rectangle((x[x_start[0]], y_start),
                       x[x_end[0]] - x[x_start[0]],
                       y_end - y_start)
             for x_start, x_end in zip(*get_truth_intervals(my_truth))]
    # implement all rectangles as a single collection
    pc = PatchCollection(boxes, facecolor="red", alpha=0.2,
                         edgecolor="red")
    ax.add_collection(pc)
    ax.plot()

plt.show()
  • Related