CDF, matplotlib - not enough colors for plot, python

问题

Here is needed to plot CDF for 8 different functions in one plot. The problem that it gives just 7 different colors and the 8 one gives just first blue color again. How to make 8 different colors?

Here is the script:

locerror_2d=[Scan_Around[1],Triangle_Around[1],M_shape_Around[1],Hilbert_Around[1],Scan_SbS[1],Triangle_SbS[1],M_shape_SbS[1],Hilbert_SbS[1]]


# N = len(locerror_2d[0]) #same for all ( here, I hope so... )
# N1=len(locerror_2d[2])
H_cent,h_cent1 = np.histogram( locerror_2d[0], bins = 10, normed = True ) # Random Walk Centroid
hy_cent = np.cumsum(H_cent)*(h_cent1[1] - h_cent1[0])

H_1st,h_1st = np.histogram( locerror_2d[1], bins = 10, normed = True ) # Random Walk Weighterd
hy_1st = np.cumsum(H_1st)*(h_1st[1] - h_1st[0])

H_2nd,h_2nd = np.histogram( locerror_2d[2], bins = 10, normed = True ) # Circle Walk Centroid
hy_2nd = np.cumsum(H_2nd)*(h_2nd[1] - h_2nd[0])

H_3rd,h_3rd = np.histogram( locerror_2d[3], bins = 10, normed = True ) # Circle Walk Weighterd
hy_3rd = np.cumsum(H_3rd)*(h_3rd[1] - h_3rd[0])

H_mm,h_mm = np.histogram( locerror_2d[4], bins = 10, normed = True ) # G Walk Centroid
hy_mm = np.cumsum(H_mm)*(h_mm[1] - h_mm[0])

H_shr,h_shr = np.histogram( locerror_2d[5], bins = 10, normed = True ) # G Walk Weighterd
hy_shr = np.cumsum(H_shr)*(h_shr[1] - h_shr[0])

H_s,h_s = np.histogram( locerror_2d[6], bins = 10, normed = True ) # G Walk Weighterd
hy_s = np.cumsum(H_s)*(h_s[1] - h_s[0])

H_sh,h_sh = np.histogram( locerror_2d[7], bins = 10, normed = True ) # G Walk Weighterd
hy_sh = np.cumsum(H_sh)*(h_sh[1] - h_sh[0])


plt.hold(True)
ddd_hist_cent, = plt.plot(h_cent1[1:], hy_cent,label="Scan_Around")        # centroid
ddd_hist_1st, = plt.plot(h_1st[1:], hy_1st,label='Triangle_Around')       #Gradient
ddd_circ_cent, = plt.plot(h_2nd[1:], hy_cent,label="M_shape_around")        # centroid
ddd_circ_wei, = plt.plot(h_3rd[1:], hy_1st,label='Hilbert_Around')       #Gradient
ddd_g_cent, = plt.plot(h_mm[1:], hy_cent,label="Scan_SbS")        # centroid
ddd_g_wei, = plt.plot(h_shr[1:], hy_1st,label='Triangle_SbS')       #Gradient
ddd_g_w, = plt.plot(h_s[1:], hy_cent,label='M_shape_SbS')
ddd_g_we, = plt.plot(h_sh[1:], hy_1st,label='Hilbert_SbS')

plt.hold(False)

plt.rc('legend',**{'fontsize':10})
plt.legend(handles=[ddd_hist_cent, ddd_hist_1st, ddd_circ_cent, ddd_circ_wei, ddd_g_cent,ddd_g_wei, ddd_g_w],loc='center left', bbox_to_anchor=(0.75, 0.18)) #no trilateration here
plt.ylabel('Probability')
plt.xlabel('Localization Error, m')
plt.ylim(ymax = 1.1, ymin = 0)
plt.title('Path Planning Algorithms')
plt.grid()
plt.show()

Thank you

回答1:

I love to read my colors directly from a colormap with this code

def getColor(c, N, idx):
    import matplotlib as mpl
    cmap = mpl.cm.get_cmap(c)
    norm = mpl.colors.Normalize(vmin=0.0, vmax=N - 1)
    return cmap(norm(idx))

Here, c is the name of the colormap (see https://matplotlib.org/examples/color/colormaps_reference.html for a list), N is the number of colors you want in total, and idx is just an index that will yield the specific color.

Then when calling the plot function, just add the color=getColor(c, N, idx) option.

回答2:

ok. I got it. In the end of plot I just need to show the color.

ddd_hist_cent, = plt.plot(h_cent1[1:], hy_cent,label="Scan_Around", c='yellow') 

回答3:

Easiest solution: Give the last curve a different color:

plt.plot(h_sh[1:], hy_1st,label='Hilbert_SbS', color="orange")

Matplotlib version 1.5 or below has 7 different colors in its color cycle, while matplotlib 2.0 has 10 different colors. Hence, updating matplotlib is another option.

In general, you may of course define your own color cycle which has as many colors as you wish.

• Build a cycler from a colormap, as shown in this question:

  import matplotlib.pyplot as plt
  from cycler import cycler
  import numpy as np
  
  N = 8 # number of colors
  plt.rcParams["axes.prop_cycle"] = cycler('color', plt.cm.jet(np.linspace(0,1,N)) )
  

• Build a cycler from a list of colors:

  import matplotlib.pyplot as plt
  from cycler import cycler
  
  colors=["aquamarine","crimson","gold","indigo",
          "lime","orange","orchid","sienna"]
  plt.rcParams["axes.prop_cycle"] = cycler('color',colors)
  

来源：https://stackoverflow.com/questions/45612129/cdf-matplotlib-not-enough-colors-for-plot-python