This little tool provides the means of easily creating TeX code for plotting large 2D arrays, which are provided in as a saved numpy array. This alleviates the drawback, that PGFplots quickly runs out of memory, when plotting large 2D heatmaps. We work around this by just plotting the numpy array to a pdf, which is then embedded into a tikzpicture.
This approach yields the following goodies: 1. We are able to generate large annotated heat plots at all! 2. We save compilation time of the global TeX document, since the heatplots themselves are "cached". 3. The generated *.tex files are very selfcontained and can be placed anywhere in your document without restrictions
In you main document you should make sure that you satisfy all dependencies by adding
\usepackage{pgfplots}
\pgfplotsset{compat=1.15}
\usepgfplotslibrary{colormaps}
in your preamble. The rest is very easy. Invoking python axify -h tells you
what to do. Assume we have two files called 'test1.npy' and 'test2.npy'
containing 2D ndarrays. Then we can axify them via
python axify.py -p test1 test2 -m jet -s heatmap -t simple
from our favourite shell. This will make use of the 'jet' colormap and the
theme in the file 'simple.tex'. After issuing this command, we should find four
new files, namely test1.png, test1.tex, test2.png and test2.tex, where
the images contain heatmap plots of the given data and the TeX files ready to
include TeX-Code, which you can place freely in your document.
One can also make use of axify directly without first writing data to disk,
by using it as a module. To do so, please import via import axify and then
carefully study the output of help(axify).