mpastell
commented
8 years ago You can set matplotlib to use pgf as backend for pdf files using a hidden chunk in the start of the document if you want. I will look into adding it as optional format as well. It looks like the currently used "agg" backend doesn't support unicode so in order to get full support one would need to set pgf as backend.
<<echo=False>>=
import matplotlib
from matplotlib.backends.backend_pgf import FigureCanvasPgf
matplotlib.backend_bases.register_backend('pdf', FigureCanvasPgf)
from matplotlib.externals import six
import matplotlib as mpl
pgf_with_custom_preamble = {
"font.family": "serif", # use serif/main font for text elements
"text.usetex": True, # use inline math for ticks
"pgf.rcfonts": False, # don't setup fonts from rc parameters
"pgf.preamble": [
"\\usepackage{units}", # load additional packages
"\\usepackage{metalogo}",
"\\usepackage{unicode-math}", # unicode math setup
r"\setmathfont{xits-math.otf}",
r"\setmainfont{DejaVu Serif}", # serif font via preamble
]
}
mpl.rcParams.update(pgf_with_custom_preamble)
@
As of version 1.2, matplotlib has the ability to write pgf files for rendering in LaTeX documents (see http://matplotlib.org/users/whats_new.html#pgf-tikz-backend). This format is useful for scientific documents where the text and measurement used in the plot need to match the text and measurements in the LaTeX document. However, they cannot be included using the standard LaTeX \includegraphics command, and require the \input command instead.
I propose adding a simple check in PwebTexFormatter that checks if figfmt is "pgf", then inserts the appropriate \input statement rather than \includegraphics.