Utility to generate PGF vector files from Python's Matplotlib plots to use in LaTeX documents.
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Failed to load latest commit information.




Python's Matplotlib can save plots as PGF vector files, with text containing LaTeX code and all. LaTeX can render PGF files. Sounds like a perfect match, doesn't it? (It does. See my bachelor thesis for example plots.)

Example for a PGF plot in LaTeX

Why PGF?

  • It's a vector format. No blurry pixel graphics anymore!
  • Any text in your plots will be typeset with your document's native font and style.
  • You can use your custom LaTeX \newcommand macros in your plots after defining a dummy implementation for them in Python. They will be typeset correctly in your document.


  1. Download the repository and run

    python example.py

    to test your Python and LaTeX setup.

  2. Copy the texfig.py file next to your Python script. It is just some simple setup and utility code for matplotlib to get you started, so feel free to edit any configurations within.

  3. Use texfig to generate PGF plots:

    # import texfig first to configure Matplotlib's backend
    import texfig  # assuming texfig.py and an __init__.py is in the directory
    # then, import PyPlot as usual
    import matplotlib.pyplot as plt
    # obtain a nicely configured figure from texfig
    fig = texfig.figure()
    # plot as usual
    # save your plot as both a PGF file (for inclusion in LaTeX) and a PDF file (for preview only)

    You can adjust the settings in texfig.py to your liking. In particular, you may want to set the default_width to your LaTeX \textwidth (in inches). In the pgf.preamble list you can also define any macros that you have implemented in your LaTeX document and that you wish to use in your plots. Don't worry about implementing them correctly here, since they are rendered by your LaTeX document later anyway.

    Make sure that no call to change matplotlib backends precedes the import texfig statement, and do not change backends afterwards either. In particular, refrain from using %matplotlib inline configurations in Jupyter Notebooks.

  4. Now \usepackage{pgf} and \input the PGF file in your LaTeX document:

    % in the preamble
    % somewhere in your document
  5. Admire the beauty of LaTeX vector plots.

Hint: If your PGF plots end up slowing down your LaTeX compile times too much, consider omitting the PGF rendering in draft mode like so:

% in the preamble
% somewhere in your document

Now when you give your \documentclass the option draft, a placeholder of the correct size is rendered instead of the PGF plot. Note that for this to work you need to keep the preview PDF generated by texfig.savefig in addition to the corresponding PGF file in the same directory.


TexFig was created and is maintained by Nils Leif Fischer.


TexFig is released under the MIT license. See LICENSE.md for details.