Skip to main content

convert matplotlib figures into TikZ/PGFPlots

Project description

Build Status Code Health Documentation Status PyPi Version PyPi Downloads

This is matplotlib2tikz, a Python tool for converting matplotlib figures into PGFPlots (TikZ) figures like

for native inclusion into LaTeX.

The output of matplotlib2tikz is in PGFPlots, a LaTeX library that sits on top of TikZ and describes graphs in terms of axes, data etc. Consequently, the output of matplotlib2tikz retains more information, can be more easily understood, and is more easily editable than raw TikZ output. For example, the matplotlib figure

from matplotlib import pyplot as pp
from matplotlib import style
import numpy as np
fig = pp.figure()
t = np.arange(0.0, 2.0, 0.1)
s = np.sin(2*np.pi*t)
s2 = np.cos(2*np.pi*t)
pp.plot(t, s, 'o-', lw=4.1)
pp.plot(t, s2, 'o-', lw=4.1)
pp.ylabel('Voltage (mV)')
pp.title('Simple plot $\\frac{\\alpha}{2}$')

(see above) gives

% This file was created by matplotlib2tikz.


title={Simple plot $\frac{\alpha}{2}$},
ylabel={Voltage (mV)},
xmin=0, xmax=2,
ymin=-1, ymax=1,
tick align=outside,
x grid style={white},
y grid style={white},
axis line style={white},
axis background/.style={fill=white!89.803921568627459!black}
\addplot [line width=1.64pt, color0, mark=*, mark size=3, mark options={draw=black}]
coordinates {
% [...]
\addplot [line width=1.64pt, color1, mark=*, mark size=3, mark options={draw=black}]
coordinates {
% [...]


Tweaking the plot is straightforward and can be done as part of your LaTeX work flow. The fantastic PGFPlots manual contains great examples of how to make your plot look even better.


Python Package Index

matplotlib2tikz is available from the Python Package Index, so simply type

pip install matplotlib2tikz

Manual installation

Download matplotlib2tikz from the Python Package Index. Place the matplotlib2tikz script in a directory where Python can find it (e.g., $PYTHONPATH). You can install it system-wide with

python install

or place the script into the directory where you intend to use it.


matplotlib2tikz needs matplotlib and NumPy to work. matplotlib2tikz works both with Python 2 and Python 3.

To use the resulting TikZ/PGFPlots figures, your LaTeX installation needs

  • TikZ (aka PGF, >=2.00), and

  • PGFPlots (>=1.3).


  1. Generate your matplotlib plot as usual.

  2. Instead of, invoke matplotlib2tikz by


    to store the TikZ file as mytikz.tex. Load the library with:

    from matplotlib2tikz import save as tikz_save

    Optional: The scripts accepts several options, for example height, width, encoding, and some others. Invoke by

    tikz_save('mytikz.tex', figureheight='4cm', figurewidth='6cm')

    Note that height and width must be set large enough; setting it too low may result in a LaTeX compilation failure along the lines of Dimension Too Large or Arithmetic Overflow; see information about these errors in the PGFPlots manual. To specify the dimension of the plot from within the LaTeX document, try

        figureheight = '\\figureheight',
        figurewidth = '\\figurewidth'

    and in the LaTeX source

  3. Add the contents of mytikz.tex into your LaTeX source code; a convenient way of doing so is via \input{/path/to/mytikz.tex}. Also make sure that in the header of your document the packages for PGFPlots and proper Unicode support and are included:


    Additionally, with LuaLaTeX


    is needed to typeset Unicode characters. Optionally, to use the latest PGFPlots features, insert



If you experience bugs, would like to contribute, have nice examples of what matplotlib2tikz can do, or if you are just looking for more information, then please visit matplotlib2tikz’s GitHub page.


matplotlib2tikz has automatic unit testing to make sure that the software doesn’t accidentally get worse over time. In test/testfunctions/, a number of test cases are specified. Those are

  • run through matplotlib2tikz,

  • the resulting LaTeX file is compiled into a PDF (pdflatex),

  • the PDF is converted into a PNG (pdftoppm),

  • a perceptual hash is computed from the PNG and compared to a previously stored version.

To run the tests, just check out this repository and type



nose2 -s test

The final pHash may depend on any of the tools used during the process. For example, if your version of Pillow is too old, the pHash function might operate slightly differently and produce a slightly different pHash, resulting in a failing test. If tests are failing on your local machine, you should first make sure to have an up-to-date Pillow, .

If you would like to contribute a test, just take a look at the examples in test/testfunctions/. Essentially a test consists of three things:

  • a description,

  • a function that creates the image in matplotlib, and

  • a pHash.

Just add your file, add it to test/testfunction/, and run the tests. A failing test will always print out the pHash, so you can leave it empty in the first run and fill it in later to make the test pass.


To create a new release

  1. bump the __version__ number,

  2. create a Git tag,

    $ git tag v0.3.1
    $ git push --tags


  3. upload to PyPi:

    $ make upload


matplotlib2tikz is published under the MIT license.

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

matplotlib2tikz-0.5.4.tar.gz (22.8 kB view hashes)

Uploaded source

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page