Fast update of single points in a plot

Question

I have a matplotlib figure in which I have several plots. One has an image with a large number of points in it. The second has an axis in which I want to update a point location with fast updates.

The piece of code below is condensed down to show effectively what I would like to do. As I move the mouse around the ax_large plot I want it to update quite quickly.

I found numerous examples on SO and other places and was trying some different options. But none seem to fit the bill quite correctly (or at least as I would hope / expect, so maybe my expectations need to change).

The code:

class Test:

    def __init__(self):
        self.fig = plt.figure(1)

        # Axis with large plot
        self.ax_large = plt.subplot(121)
        self.ax_large.imshow(np.random.random((5000,5000)))

        # Follow the point
        self.ax = plt.subplot(122)
        self.ax.grid('on')
        self.fig.canvas.callbacks.connect('motion_notify_event', self.callback)

        self.point = self.ax.plot(0,0, 'go')

        plt.show()

    def callback(self, event):
        if event.inaxes == self.ax:
            print('Updating to {} {}'.format(event.xdata, event.ydata))

            self.point[0].set_data(event.xdata, event.ydata)

            # Option 1. Works, bu super slow if there are other large sub-plots
            plt.draw()

            # Option 2. Doesn't update
            # self.fig.canvas.blit(self.ax.bbox)

            # Option 3. Works but then grid goes away
            # self.ax.redraw_in_frame()
            # self.fig.canvas.blit(self.ax.bbox)

            # Option 4. Doesn't update
            # self.ax.draw_artist(self.point[0])

            # Option 5. Draws new point but does not remove the "old" one
            # self.ax.draw_artist(self.point[0])
            # self.fig.canvas.blit(self.ax.bbox)

if __name__ == '__main__':
    tt = Test()

As you move around the ax_large I was hoping it would be a fast update for the point's location.

Any ideas on how to do this would be helpful.

Thanks...

If you look at the examples, you're implementing only half of what is needed for blitting. — ImportanceOfBeingErnest
– ImportanceOfBeingErnest, Commented Mar 9, 2018 at 23:30

ImportanceOfBeingErnest · Accepted Answer · 2018-03-09 23:47:32Z

You are essentially ignoring most of what is needed for blitting. See e.g.

As usual you need to

Draw the canvas, fig.canvas.draw()
Store the background for later, fig.canvas.copy_from_bbox()
Update the point, point.set_data
Restore the background, fig.canvas.restore_region
draw the point, ax.draw_artist
blit the axes, fig.canvas.blit

Hence

import matplotlib.pyplot as plt
import numpy as np

class Test:
    def __init__(self):
        self.fig = plt.figure(1)
        # Axis with large plot
        self.ax_large = plt.subplot(121)
        self.ax_large.imshow(np.random.random((5000,5000)))
        # Follow the point
        self.ax = plt.subplot(122)
        self.ax.grid(True)

        # set some limits to the axes
        self.ax.set_xlim(-5,5)
        self.ax.set_ylim(-5,5)
        # Draw the canvas once
        self.fig.canvas.draw()
        # Store the background for later
        self.background = self.fig.canvas.copy_from_bbox(self.ax.bbox)
        # Now create some point
        self.point, = self.ax.plot(0,0, 'go')
        # Create callback to mouse movement
        self.cid = self.fig.canvas.callbacks.connect('motion_notify_event', 
                                                     self.callback)
        plt.show()

    def callback(self, event):
        if event.inaxes == self.ax:
            # Update point's location            
            self.point.set_data(event.xdata, event.ydata)
            # Restore the background
            self.fig.canvas.restore_region(self.background)
            # draw the point on the screen
            self.ax.draw_artist(self.point)
            # blit the axes
            self.fig.canvas.blit(self.ax.bbox)


tt = Test()

Collectives™ on Stack Overflow

Fast update of single points in a plot

1 Answer 1

Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related