Object-oriented pyplot

Question

I need to handle the pyplot objects, like figures and axes. Here is a simplified example of what I want:

In [1]: import matplotlib.pyplot as mp

In [2]: fig = mp.figure()             # create a figure

In [3]: mp.show()                     # and immediately show it. And close.

In [4]: ax = fig.add_subplot(111)     # Then I create a plot on that figure

In [5]: ax.plot([1, 2, 3]) 
Out[5]: [<matplotlib.lines.Line2D at 0x104e29a50>]

In [6]: mp.get_fignums()              # But I already released the figure, so it doesn't appear in the list of available figures
Out[6]: []

In [7]: fig.axes[0].lines[0].get_data()   # The data is there, on the plot
Out[7]: (array([ 0.,  1.,  2.]), array([1, 2, 3]))

In [8]: mp.show()                     # But mp.show() shows nothing.

The fig.show() doesn't work too. How to show a figure after releasing it?

UPD: There was a similar question: Matplotlib: re-open a closed figure?, but with no answer.

Answer 1

Which is the recommended way to plot: matplotlib or pylab? question is relevant to this question.

The pyplot interface is a convenience module that keeps track of a) open figures and b) the 'current figure' and 'current axes'. Underneath it is using the OO interface.

To have an open figure and a be able to enter new commands at the repl you need to be in 'interactive' mode which integrates the python repl loop with the GUI event loop.

From your question it looks like you are using IPython so use the %matplotlib magic:

16:31 $ ipython
Python 3.5.1 |Continuum Analytics, Inc.| (default, Dec  7 2015, 11:16:01) 
Type "copyright", "credits" or "license" for more information.

IPython 4.2.0 -- An enhanced Interactive Python.
?         -> Introduction and overview of IPython's features.
%quickref -> Quick reference.
help      -> Python's own help system.
object?   -> Details about 'object', use 'object??' for extra details.

In [1]: %matplotlib
Using matplotlib backend: Qt4Agg

In [2]: import matplotlib.pyplot as plt

In [3]: fig, ax = plt.subplots()  # prompt returns immediatly leaving open figure

In [4]: ln, = ax.plot(range(15), label='test')  # draws line and updates figure

In [5]: ln.set_linewidth(5)  # changes lw and updates screen

In [6]: 

Answer 2

Try this:

import matplotlib.pyplot as mp

fig = mp.figure()

plt.show() # empty figure appears, close it

fig = plt.gcf() # get current figure, this is the key piece.

ax = fig.add_subplot(111) # added axes object

ax.plot([1,2,3])

plt.show()

When I did this, I was able to get the plot to show with a diagonal line plotted.

Answer 3

I found it! Let's create a mp.Figure()

import matplotlib.pyplot as mp
fig = mp.Figure()

Now it is not connected to pyplot, so we can't show it. It is equivalent to what happens when you close a figure. The fact that you can't show a figure which is not connected to pyplot is well documented. Just try

In []: fig.show?
Docstring:
If using a GUI backend with pyplot, display the figure window.
For non-GUI backends, this does nothing.

(I reduced the text of the help message.) But it is possible to trick the pyplot. Let's create a figure:

temp_fig = mp.figure()

Steal the figure manager from temp_fig and assign it to our fig:

m = mp.get_current_fig_manager()
fig.canvas.manager = m

Now we can show it:

mp.show() # Shows the fig figure.

Of course, it is a good practice to delete the temp_fig:

del temp_fig