Keep zoom and ability to zoom out to current data extent in matplotlib.pyplot-CodePudding

I want to be able to replot data in an interactive matplotlib figure, keeping the zoom that is present. The Internet says to use set_data. This works.

What doesn't work is then being able to click the home button to see the entire extent of the new data I've just replotted.

How can I force the home button to show me all of the data currently plotted? This all happens inside a tKinter window.

Some example code:

import matplotlib.pyplot as plt

x1 = [1, 2, 3, 4, 5, 6]
y1 = [3, 4, 5, 6, 7, 8]

x2 = [2, 3, 4, 5, 6, 7]
y2 = [8, 7, 6, 5, 4, 3]

# plot data
line, = plt.plot(x1, y1, marker="o")

# zoom in - this actually happens by user interaction with the zoom tool
plt.xlim(2.5, 5.5)
plt.ylim(3, 6)

# replace data, maintaining zoom. This actually happens by the user choosing new data to plot
line.set_data(x2, y2)

plt.show()
# now, I want to unzoom to show all of x2,y2 by pressing the "home" button
pass

CodePudding user response：

The "Home" button takes you back to the zoom that has been registered as default. You should set it programmatically.

I do not know a direct interface for setting this default, but you can register the current status of the Canvas as the default, using ToolBar's push_current method.

In your case, you should make this in the place where you change the data. Instead of simply changing the data

line.set_data(x2, y2)

you should do some bookkeeping:

fig = plt.gcf() # I assume the OP already has this data
ax = fig.gca() # I assume the OP already has this data

# save the current zoom for restoring later
old_x_lim = ax.get_xlim()
old_y_lim = ax.get_ylim()

# let matplotlib calculate the optimal zoom for the new data
ax.relim()
ax.autoscale()

# Now the tricky part... I did not find much documentation on this
toolbar = fig.canvas.manager.toolbar
toolbar.update() # Clear the axes stack
toolbar.push_current()  # save the current status as home
ax.set_xlim(old_x_lim)  # and restore zoom
ax.set_ylim(old_y_lim)

I hope this works in you setup. Otherwise you need to poke into matplotlib's internals.