This whole tool is a basic front end for using Python's matplotlib
in a
moderately interactive and robust manner to do MATLAB-like number crunching
and (more critically) plot generation for papers.
While MATLAB has routines to save figures, the graphics back-end routinely runs into issues with NVIDIA GPU based systems, and I'm sick and tired of being tied to a tool that has a heavy resource footprint and only moderate documentation. The licensing restrictions (though not fundamentally debilitating) are a secondary reason I'm moving away from MATLAB. Finally, as I expect to graduate soon, the $50 (or $130 for my toolboxes) annual cost is going to rise to a debilitating point for things as mundane as personal projects. So I might as well kick an expensive habit while it's easy to fall back when needed.
There are a few tricks to help configuring matplotlib
. I'll update this
document to describe the commands and tools to help decipher the
information required to produce plots in a repeatable and tidy way.
Plot defaults are managed by the matplotlib
import matplotlib.font_manager
print(matplotlib.font_manager.fontManager.afmlist)
print(matplotlib.font_manager.fontManager.ttflist)
I search for fonts using the following method:
import matplotlib.font_manager as FM
import re
fcFontList = FM.get_fontconfig_fonts()
# Search only for fonts that have name matches similar to this
fontsDesired = ['Helvetica', 'Times', 'Arial']
fontsDesiredRe = re.compile('|'.join(fontsDesired), flags=re.IGNORECASE)
# Create a unique set of the fonts selected out of all of the system fonts
fontsAvailable = set([FM.FontProperties(fname=fcFont).get_name()\
for fcFont in fcFontList if fontsDesiredRe.search(fcFont) != None])
- make pySmithPlot a git sub-module
- think of a smarter way to refactor things (this is an ever evolving goal)