CODEWITHSUNDEEP
pythonmatplotlibplotscipyclipboard
EelkeSpaak
EelkeSpaak

Reputation: 2825

How to add clipboard support to Matplotlib figures?

In MATLAB, there is a very convenient option to copy the current figure to the clipboard. Although Python/numpy/scipy/matplotlib is a great alternative to MATLAB, such an option is unfortunately missing.

Can this option easily be added to Matplotlib figures? Preferably, all MPL figures should automatically benefit from this functionality.

I'm using MPL's Qt4Agg backend, with PySide.

Upvotes: 19

Views: 21103

Answers (6)

Doron Behar
Doron Behar

Reputation: 2878

In my workflow, I often create CLI applications that both plot figures and if instructed via a --save argument, also save the plot to a file. While seeking the option to copy these figures straight to the clipboard, without going through the file system, I found the following code flow useful:

import sys
import argparse

parser = argparse.ArgumentParser(...)
parser.add_argument("--save", ...)

args = parser.parse_args()
if args.save = ["-"]:
    # See https://matplotlib.org/stable/gallery/misc/print_stdout_sgskip.html
    args.save = [ sys.stdout.buffer ]
# plot stuff...
if args.save:
    plt.savefig(args.save[0])

Then, from the command line, you can simply pipe this to a clipboard command like this:

./plotter.py --save - | wl-copy --type image/png

Or if you are still on Xorg, use:

./plotter.py --save - | xclip -selection clipboard -target image/png -i

Upvotes: 0

emgf_co
emgf_co

Reputation: 373

The last comment is very useful.

  1. Install the package with

    pip install addcopyfighandler

  2. Import the module after importing matplotlib, for instance:

    import matplotlib.pyplot as plt
    import addcopyfighandler

  3. Use ctr + C to copy the Figure to the clipboard

And enjoy.

Upvotes: 2

kilodalton
kilodalton

Reputation: 121

I enabled ctrl-c to copy figures to the clipboard by adding toolbar: toolmanager to my matplotlibrc.

  1. Locate your rc file by typing
  import matplotlib
  print(matplotlib.matplotlib_fname())

in the python interpreter.

Then paste

toolbar: toolmanager

into the file using your favorite editor.

Alternatively, the following bash one-liner will do the trick:

echo "toolbar: toolmanager" >> $(python -c "import matplotlib;print(matplotlib.matplotlib_fname())")

Note that this needs to be done for each python interpreter. So if you make a new anaconda environment, you will have to repeat the process.

Upvotes: 3

Guimoute
Guimoute

Reputation: 4629

Building on the solution described in EelkeSpaak's answer, we can write a function to do the work instead of monkey-patching matplotlib's Figure class:

import io
import matplotlib.pyplot as plt
import numpy as np

from PyQt5.QtGui import QImage
from PyQt5.QtWidgets import QApplication

# Example figure.
fig, ax = plt.subplots()
X = np.linspace(0, 2*np.pi)
Y = np.sin(X)
ax.plot(X, Y)

def add_figure_to_clipboard(event):
    if event.key == "ctrl+c":
       with io.BytesIO() as buffer:
            fig.savefig(buffer)
            QApplication.clipboard().setImage(QImage.fromData(buffer.getvalue()))

fig.canvas.mpl_connect('key_press_event', add_figure_to_clipboard)

Please note that this variant uses Qt5.

Upvotes: 3

Vaiaro
Vaiaro

Reputation: 1008

EelkeSpaak's solution was packed in a nice module: addcopyfighandler

Simply install by pip install addcopyfighandler, and import the module after importing matplotlib or pyplot.

Upvotes: 8

EelkeSpaak
EelkeSpaak

Reputation: 2825

Yes, it can. The idea is to replace the default plt.figure with a custom one (a technique known as monkey patching) that injects a keyboard handler for copying to the clipboard. The following code will allow you to copy any MPL figure to the clipboard by pressing Ctrl+C:

import io
import matplotlib.pyplot as plt
from PySide.QtGui import QApplication, QImage

def add_clipboard_to_figures():
    # use monkey-patching to replace the original plt.figure() function with
    # our own, which supports clipboard-copying
    oldfig = plt.figure

    def newfig(*args, **kwargs):
        fig = oldfig(*args, **kwargs)
        def clipboard_handler(event):
            if event.key == 'ctrl+c':
                # store the image in a buffer using savefig(), this has the
                # advantage of applying all the default savefig parameters
                # such as background color; those would be ignored if you simply
                # grab the canvas using Qt
                buf = io.BytesIO()
                fig.savefig(buf)
                QApplication.clipboard().setImage(QImage.fromData(buf.getvalue()))
                buf.close()

        fig.canvas.mpl_connect('key_press_event', clipboard_handler)
        return fig

    plt.figure = newfig

add_clipboard_to_figures()

Note that if you want to use from matplotlib.pyplot import * (e.g. in an interactive session), you need to do so after you've executed the above code, otherwise the figure you import into the default namespace will be the unpatched version.

Upvotes: 15

Related Questions