List of all available matplotlib backends

Question

The current backend name is accessible via

>>> import matplotlib.pyplot as plt
>>> plt.get_backend()
'GTKAgg'

Is there a way to get a list of all backends that can be used on a particular machine?

FYI as of `matplotlib` version 3.x the default pip install does not install an interactive backend so you need to also install one of the interative backends like `PyQt4`, or `PyQt5` see https://matplotlib.org/3.3.3/users/installing.html#install-requirements for more info about supported interactive backends. — Trevor Boyd Smith, Jan 27 '21 at 16:35

score 62 · Accepted Answer · answered Feb 23 '11 at 14:32

You can access the lists

matplotlib.rcsetup.interactive_bk
matplotlib.rcsetup.non_interactive_bk
matplotlib.rcsetup.all_backends

the third being the concatenation of the former two. If I read the source code correctly, those lists are hard-coded though, and don't tell you what backends are actually usable. There is also

matplotlib.rcsetup.validate_backend(name)

but this also only checks against the hard-coded list.

score 53 · Answer 2 · edited Jan 10 '21 at 17:32

Here is a modification of the script posted previously. It finds all supported backends, validates them and measures their fps. On OSX it crashes python when it comes to tkAgg, so use at your own risk ;)

from __future__ import print_function, division, absolute_import
from pylab import *
import time

import matplotlib.backends
import matplotlib.pyplot as p
import os.path


def is_backend_module(fname):
    """Identifies if a filename is a matplotlib backend module"""
    return fname.startswith('backend_') and fname.endswith('.py')

def backend_fname_formatter(fname): 
    """Removes the extension of the given filename, then takes away the leading 'backend_'."""
    return os.path.splitext(fname)[0][8:]

# get the directory where the backends live
backends_dir = os.path.dirname(matplotlib.backends.__file__)

# filter all files in that directory to identify all files which provide a backend
backend_fnames = filter(is_backend_module, os.listdir(backends_dir))

backends = [backend_fname_formatter(fname) for fname in backend_fnames]

print("supported backends: \t" + str(backends))

# validate backends
backends_valid = []
for b in backends:
    try:
        p.switch_backend(b)
        backends_valid += [b]
    except:
        continue

print("valid backends: \t" + str(backends_valid))


# try backends performance
for b in backends_valid:

    ion()
    try:
        p.switch_backend(b)


        clf()
        tstart = time.time()               # for profiling
        x = arange(0,2*pi,0.01)            # x-array
        line, = plot(x,sin(x))
        for i in arange(1,200):
            line.set_ydata(sin(x+i/10.0))  # update the data
            draw()                         # redraw the canvas

        print(b + ' FPS: \t' , 200/(time.time()-tstart))
        ioff()

    except:
        print(b + " error :(")

To just see supported interactive backends see:

#!/usr/bin/env python
from __future__ import print_function
import matplotlib.pyplot as plt
import matplotlib

backends = matplotlib.rcsetup.interactive_bk
# validate backends
backends_valid = []
for b in backends:
    try:
        plt.switch_backend(b)
        backends_valid += [b]
    except:
        continue
print(backends_valid)

The scripts crashes for me (https://gist.github.com/palmstrom/6039823), but works well when run under Spyder IDE. — Matěj Šmíd, Jul 19 '13 at 15:12
What are typical results of this benchmark? FPS numbers anybody? — rc0r, May 03 '17 at 11:25
so awesome! live drawing! calculates framerate and prints it! really useful to figure out which backend is fastest on your machine. — Trevor Boyd Smith, May 31 '18 at 16:32
to run the script under `python2`, make the first line after shebang the `from __future__ import print_function, division, absolute_import`. (i was going to edit the code in the question... but `python2` is so bloody old at this point... i feel it is bad manners to add that sort of cruft... i feel like it would just be encouraging bad behavior.) — Trevor Boyd Smith, Jul 11 '19 at 17:46

score 8 · Answer 3 · answered Jun 23 '19 at 04:54

You can pretend to put a wrong backend argument, then it will return you a ValueError with the list of valid matplotlib backends, like this:

Input:

import matplotlib
matplotlib.use('WRONG_ARG')

Output:

ValueError: Unrecognized backend string 'test': valid strings are ['GTK3Agg', 'GTK3Cairo', 'MacOSX', 'nbAgg', 'Qt4Agg', 'Qt4Cairo', 'Qt5Agg', 'Qt
5Cairo', 'TkAgg', 'TkCairo', 'WebAgg', 'WX', 'WXAgg', 'WXCairo', 'agg', 'cairo', 'pdf', 'pgf', 'ps', 'svg', 'template']

score 6 · Answer 4 · answered May 07 '11 at 10:12

There is the hard-coded list mentioned by Sven, but to find every backend which Matplotlib can use (based on the current implementation for setting up a backend) the matplotlib/backends folder can be inspected.

The following code does this:

import matplotlib.backends
import os.path

def is_backend_module(fname):
    """Identifies if a filename is a matplotlib backend module"""
    return fname.startswith('backend_') and fname.endswith('.py')

def backend_fname_formatter(fname): 
    """Removes the extension of the given filename, then takes away the leading 'backend_'."""
    return os.path.splitext(fname)[0][8:]

# get the directory where the backends live
backends_dir = os.path.dirname(matplotlib.backends.__file__)

# filter all files in that directory to identify all files which provide a backend
backend_fnames = filter(is_backend_module, os.listdir(backends_dir))

backends = [backend_fname_formatter(fname) for fname in backend_fnames]

print backends

Leandro · Answer 5 · 2013-03-13T12:22:33.113

4

You can also see some documentation for a few backends here:

http://matplotlib.org/api/index_backend_api.html

the pages lists just a few backends, some of them don't have a proper documentation:

matplotlib.backend_bases
matplotlib.backends.backend_gtkagg
matplotlib.backends.backend_qt4agg
matplotlib.backends.backend_wxagg
matplotlib.backends.backend_pdf
matplotlib.dviread
matplotlib.type1font

edited Mar 13 '13 at 12:22

answered Feb 18 '13 at 12:37

Leandro

814
7
7

I'm sorry, I thought that the above answers covered the subject, but should have the link as reference, that's why I posted. I mean no harm, I'm a noob. – Leandro Mar 12 '13 at 19:18

score 3 · Answer 6 · answered Oct 02 '20 at 19:07

3

What about this?

%matplotlib --list
Available matplotlib backends: ['tk', 'gtk', 'gtk3', 'wx', 'qt4', 'qt5', 'qt', 'osx', 'nbagg', 'notebook', 'agg', 'svg', 'pdf', 'ps', 'inline', 'ipympl', 'widget']

answered Oct 02 '20 at 19:07

Gadi Oron

880
8
6

score 2 · Answer 7 · answered Feb 23 '11 at 14:44

2

You could look at the following folder for a list of possible backends...

/Library/Python/2.6/site-packages/matplotlib/backends
/usr/lib64/Python2.6/site-packages/matplotlib/backends

answered Feb 23 '11 at 14:44

Bill Lynch

80,138
16
128
173

List of all available matplotlib backends

7 Answers7

Linked