How can I display a matplotlib interactive window without any of the controls visible?

Question

I need to display a PNG on a users Xfce4 desktop. So far I'm using a python script that shows the PNG in an interactive matplotlib window:

import matplotlib.pyplot as plt
import matplotlib.image as mpimg

img = mpimg.imread("My.png")
plt.imshow(img)
plt.show()

That is very unattractive though. Is there a way to remove all the interactive controls, all the border space (axis?) that gets put around the image, and resize the window to a particular width/height on startup?

Alternatively, is there a better option to provide a lightweight and static display of an image on a desktop? Doesn't have to be python/matplotlib of course.

Answer 1

Sure, but at that point, you might consider using a "bare" gui toolkit instead.

At any rate, here's the matplotlib way:

import matplotlib.pyplot as plt

# Note that the size is in inches at 80dpi. 
# To set a size in pixels, divide by 80.
fig = plt.figure(figsize=(4, 5))

# Now we'll add an Axes that takes up the full figure
ax = fig.add_axes([0, 0, 1, 1])
ax.axis('off') # Hide all ticks, labels, outlines, etc.

# Display the image so that it will stretch to fit the size
# of the figure (pixels won't be square)
ax.imshow(plt.imread('test.png'), aspect='auto')

plt.show()

This does everything except for hiding the toolbar. To hide the toolbar, you'll need to be be backend-specific. You have two options: 1) Handle creating the window manually and embed the matplotlib canvas inside it, 2) hide the toolbar using backend-specific methods.

As an example of hiding the toolbar, with the qt-based backends, you'd do:

import matplotlib
matplotlib.use('qt4agg')
import matplotlib.pyplot as plt

fig = plt.figure(figsize=(4, 5))
ax = fig.add_axes([0, 0, 1, 1])
ax.axis('off')

ax.imshow(plt.imread('test.png'), aspect='auto')

# qt specific!
fig.canvas.toolbar.setVisible(False)

plt.show()

And for the Tk-backend, you'd do:

import matplotlib
matplotlib.use('tkagg')
import matplotlib.pyplot as plt

fig = plt.figure(figsize=(4, 5))
ax = fig.add_axes([0, 0, 1, 1])
ax.axis('off')

ax.imshow(plt.imread('test.png'), aspect='auto')

# Tk specific!
fig.canvas.toolbar.pack_forget()

plt.show()

By contrast, if you wanted to skip matplotlib together and just use Tkinter, you'd do something like:

import Tkinter as tk
from PIL import ImageTk

root = tk.Tk()

im = ImageTk.PhotoImage(file='test.png')
panel = tk.Label(root, image=im)
panel.pack(fill=tk.BOTH, expand=True)

root.mainloop()

This displays the image at one-pixel-to-one-pixel on the screen and doesn't allow for resizing. However, it's about as minimal as you can get.

Answer 2

try this:

import matplotlib.pyplot as plt
import matplotlib.image as mpimg

w = 10
h = 20
fig, ax = plt.subplots(figsize=(10, 20))
ax.axis('off')
img = mpimg.imread("c:\mario.png")
plt.imshow(img)
plt.show()