Interaction with python's matplotlib figure: assign value to selected features

问题

Is it possible to select an area inside a matplotlib's figure window to which assign the value, say, 0? For example, let's say I want to write a script that, at a certain point, shows the image inside a figure window (pyplot.imshow) and asks the user to select an area to which assign the value 0? Hope that was clear enough.

回答1:

This works well. Here you have a pcolormesh in which you can click and the onclick function that catches the click event will handle it and set the selected square to zero. The mpl_connect function connects the onclick function to the button_press_event event. You can see the update directly after the click.

import numpy as np
import pylab as pl

pl.ioff()

rand_field = np.random.rand(10,10)

fig = pl.figure()
cm = pl.pcolormesh(rand_field, vmin=0, vmax=1)
pl.colorbar()

def onclick(event):
    indexx = int(event.xdata)
    indexy = int(event.ydata)
    print("Index ({0},{1}) will be set to zero".format(indexx, indexy))
    rand_field[indexy, indexx] = 0.
    cm.set_array(rand_field.ravel())
    event.canvas.draw()

cid = fig.canvas.mpl_connect('button_press_event', onclick)

pl.show()

Here you find a more advanced version that can drag a region and handles the error in case somebody clicks outside of the figure. I leave the drawing of the rectangle up to you:

import numpy as np
import pylab as pl

pl.ioff()

rand_field = np.random.rand(10,10)

fig = pl.figure()
cm = pl.pcolormesh(rand_field, vmin=0, vmax=1)
pl.colorbar()

x_press = None
y_press = None

def onpress(event):
    global x_press, y_press
    x_press = int(event.xdata) if (event.xdata != None) else None
    y_press = int(event.ydata) if (event.ydata != None) else None

def onrelease(event):
    global x_press, y_press
    x_release = int(event.xdata) if (event.xdata != None) else None
    y_release = int(event.ydata) if (event.ydata != None) else None

    if (x_press != None and y_press != None and x_release != None and y_release != None):
        (xs, xe) = (x_press, x_release+1) if (x_press <= x_release) else (x_release, x_press+1)
        (ys, ye) = (y_press, y_release+1) if (y_press <= y_release) else (y_release, y_press+1)
        print("Slice [{0}:{1},{2}:{3}] will be set to zero".format(xs, xe, ys, ye))
        rand_field[ys:ye, xs:xe] = 0.
        cm.set_array(rand_field.ravel())
        event.canvas.draw()

    x_press = None
    y_press = None

cid_press   = fig.canvas.mpl_connect('button_press_event'  , onpress  )
cid_release = fig.canvas.mpl_connect('button_release_event', onrelease)

pl.show()

来源：https://stackoverflow.com/questions/33231120/interaction-with-pythons-matplotlib-figure-assign-value-to-selected-features