What happens when you call IPython with the `--matplotlib` command-line argument?

Question

I could only find this from their help section.

Configure matplotlib for interactive use with the default matplotlib

I have been having performance issues plotting using matplotlib.pyplot with the IPython command-line until I tried the --matplotlib option.

Example

Without --matplotlib

 $ ipython
 In [1]: import matplotlib as mpl

 In [2]: import matplotlib.pyplot as plt

 In [3]: mpl.get_backend()
 Out[3]: u'Qt4Agg'

 In [4]: plt.plot([0, 1, 2], [0, 1, 2])
 Out[4]: [<matplotlib.lines.Line2D at 0xb473198>]
 # IPython command-line becomes entirely unresponsive, must restart IPython to become usable again

With --matplotlib

 $ ipython --matplotlib
 In [1]: import matplotlib as mpl

 In [2]: import matplotlib.pyplot as plt

 In [3]: mpl.get_backend()
 Out[3]: u'Qt4Agg'

 In [4]: plt.plot([0, 1, 2], [0, 1, 2])
 Out[4]: [<matplotlib.lines.Line2D at 0xcbe1d68>]
 # IPython command-line remains responsive

I suspect a side-effect of using the --matplotlib argument is boosting my performance, but I'd like to know how.

Setup

IPython: 3.0.0
matplotlib: 1.4.3
Python: 2.7.9 :: Anaconda 2.2.0 (64-bit)
Windows 7 (64-bit)

tmdavison · Accepted Answer · 2015-07-24 09:46:54Z

1

I think it is the equivalent of setting plt.interactive(True) (i.e. turning interactive mode on, or equivalently running plt.ion()), so that when a figure instance is created, you still have control of the terminal. See here for more information.

For example:

$ ipython --matplotlib

import matplotlib.pyplot as plt
plt.isinteractive() 
# True

As opposed to:

$ ipython

import matplotlib.pyplot as plt
plt.isinteractive() 
# False

As a side note, in your first example, the command line will remain unresponsive until you close the figure that is created by your plt.plot command. Once you close that window, you should regain control of the command line.

edited Jul 24, 2015 at 9:46

answered Jul 24, 2015 at 9:13

tmdavison

69.7k13 gold badges204 silver badges182 bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

Alex Over a year ago

Thanks Tom, although I think something additional is happening as I still see the performance issues after using plt.ion(). In my first example, the figure is never displayed as I don't call plt.show(). I'm sorry if I wasn't clear enough originally, but by 'unresponsive' I meant each keystroke takes about 1/4 second to appear, rather than simply blocking input.

Collectives™ on Stack Overflow

What happens when you call IPython with the `--matplotlib` command-line argument?

1 Answer 1

1 Comment

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

1 Comment

Your Answer

Sign up or log in

Post as a guest

Related