pubstyle

Matplotlib settings and wrapper functions for close to publication quality figures. Tries to replicate my preferred style of using Adobe Illustrator to beautify figures, and gets 90% of the way there. Sets all font sizes, line widths, tick styles etc to publication quality.

Installation

Enter the repo root directory, and run pip install . (it is recommended to do so in a fresh conda environment if errors occur) - alternatively, you can just copy the functions you care about out of the "core.py" but this will miss some stylistic things. Import like import pubstyle as ps

Use

Settings

There are a few main functions,

1. ps.set_publication_style() - implements the new matplotlib parameters shown in "pubstyle.mplstyle". This only has to be run once, and will apply to all subsequent figures and does 80% of the styling. To revert to the default styling run "ps.reset_style()"
2. ps.format(fig) - takes as input a figure and applies more custom styling to it (rounding of ticks, styling legends, making markers look nicer, etc). If you look in core.py you'll see that ps.format(fig) is a wrapper for a bunch of subfunction calls, which you can turn off/do manually as you see fit
3. ps.size(str,height) - when creating axes, it is convenient to make them align exactly with the sizes of various journals. The first argument of "ps.size" is a journal identified (e.g. "nature", "aps", "science", "nature2", "aps2", "science2", "science3" for various journals and column widths), which will then autoset the width to the appropriate size. The second argument, height, is in mm.

Standard use

It is standard to import pubstyle as ps, then run ps.set_publication_style() once in order to set all the various plotting defaults. Then run ps.format(fig) after the figure is created (right before you do plt.show() or save the figure).

Known problems

I have not tested all possible plot types, so it's always possible something comes up that causes problems. I have endeavored to make it support multiple subplots, but there will always be places where spacings are a little bit weird. You should assume if you are arranging multiple disparate subplots that you will likely want to arrange in Illustrator (or equivalent) after.

Examples

In all plots I apply set_publication_style() and format(fig). See pubstyle/example.py

I do not set any font sizes etc, and let either the standard default settings, or the pubstyle settings take care of everything. (obviously the "before' could be made to look more nice by changing font sizes etc, but that's the point of pubstyle, to take care of all that for standard journal formatting.)

Before	After