Logo Questions Linux Laravel Mysql Ubuntu Git Menu
 

matplotlib how to fill_between step function

I am trying to shade regions of a plot where an input signal is high (value = 1). The region should remain shaded until the signal goes low (value = 0). I have gotten pretty close, following a number of examples: http://matplotlib.org/examples/pylab_examples/axhspan_demo.html In a matplotlib plot, can I highlight specific x-value ranges? How do I plot a step function with Matplotlib in Python?

The problem is that right now it is only shading directly under where the signal = 1, rather than to the next change to signal = 0 (step function). For example, in the image / code below, I would like the plot to be filled between 20-40 and 50-60 (rather than 20-30, and a spike under 40). How can I modify my code to achieve this? Thanks. output plot showing incorrect shading

import numpy as np
import matplotlib.pyplot as plt

x = np.array([0,10,20,30,40,50,60])
s = np.array([0,0,1,1,0,1,0])
t = np.array([25,24,25,25,24,25,24])

fig, ax = plt.subplots()

ax.plot(x,t)
ax.step(x,s,where='post')

# xmin xmax ymin ymax
plt.axis([0,60,0,30])

ymin, ymax = plt.ylim()
# want this to fill until the next "step"
# i.e. should be filled between 20-40; 50-60
ax.fill_between(x, ymin, ymax, where=s>0, facecolor='green', alpha=0.5)

plt.show()
like image 914
Roberto Avatar asked Nov 02 '22 08:11

Roberto


1 Answers

Define a generator giving the intervals to fill.

def customFilter(s):
    foundStart = False
    for i, val in enumerate(s):
        if not foundStart and val == 1:
            foundStart = True
            start = i
        if foundStart and val == 0:
            end = i
            yield (start, end+1)
            foundStart = False
    if foundStart:
        yield (start, len(s))  

The use this to get the intervals to fill in.

for start, end in customFilter(s):
    print 1
    mask = np.zeros_like(s)
    mask[start: end] = 1
    ax.fill_between(x, ymin, ymax, where=mask, facecolor='green', alpha=0.5)
like image 89
M4rtini Avatar answered Nov 15 '22 03:11

M4rtini