Color values in imshow for matplotlib?

Question

I'd like to know the color value of a point I click on when I use imshow() in matplotlib. Is there a way to find this information through the event handler in matplotlib (the same way as the x,y coordinates of your click are available)? If not, how would I find this information?

Specifically I'm thinking about a case like this:

imshow(np.random.rand(10,10)*255, interpolation='nearest')

Thanks! --Erin

Answer 1

The above solution only works for a single image. If you plot two or more images in the same script, the "inaxes" event will not make a difference between both axis. You will never know in which axis are you clicking, so you won't know which image value should be displayed.

Answer 2

If by 'color value' you mean the value of the array at a clicked point on a graph, then this is useful.

from matplotlib import pyplot as plt
import numpy as np


class collect_points():
   omega = []
   def __init__(self,array):
       self.array = array
   def onclick(self,event):
       self.omega.append((int(round(event.ydata)),   int(round(event.xdata))))

   def indices(self):
       plot = plt.imshow(self.array, cmap = plt.cm.hot, interpolation =  'nearest', origin= 'upper')
       fig = plt.gcf()
       ax = plt.gca()
       zeta = fig.canvas.mpl_connect('button_press_event', self.onclick)
       plt.colorbar()
       plt.show()
       return self.omega

Usage would be something like:

from collect_points import collect_points
import numpy as np

array = np.random.rand(10,10)*255   
indices = collect_points(array).indices()

A plotting window should appear, you click on points, and returned are the indices of the numpy array.

Answer 3

Here's a passable solution. It only works for interpolation = 'nearest'. I'm still looking for a cleaner way to retrieve the interpolated value from the image (rather than rounding the picked x,y and selecting from the original array.) Anyway:

from matplotlib import pyplot as plt
import numpy as np

im = plt.imshow(np.random.rand(10,10)*255, interpolation='nearest')
fig = plt.gcf()
ax = plt.gca()

class EventHandler:
    def __init__(self):
        fig.canvas.mpl_connect('button_press_event', self.onpress)

    def onpress(self, event):
        if event.inaxes!=ax:
            return
        xi, yi = (int(round(n)) for n in (event.xdata, event.ydata))
        value = im.get_array()[xi,yi]
        color = im.cmap(im.norm(value))
        print xi,yi,value,color

handler = EventHandler()

plt.show()

Answer 4

You could try something like the below:

x, y = len(df.columns.values), len(df.index.values)

# subplot etc 

# Set values for x/y ticks/labels
ax.set_xticks(np.linspace(0, x-1, x))
ax.set_xticklabels(ranges_df.columns)
ax.set_yticks(np.linspace(0, y-1, y))
ax.set_yticklabels(ranges_df.index)

 for i, j in product(range(y), range(x)):
    ax.text(j, i, '{0:.0f}'.format(ranges_df.iloc[i, j]),
    size='small', ha='center', va='center')