Paper PLT Documentation¶

Paper PLT is a simple matplotlib renderer wrapper to build plots from the commandline.

Contents:

Installation¶

pip install git+https://github.com/nils-werner/pplt.git

Setup¶

PPLT expects there to be a pplt/ directory in your $PYTHONPATH or your current directory.

Inside this directory there must be an __init__.py file and all renderer modules you need. e.g.

|- paper.tex
`- pplt/
   |- __init__.py
   |- input_signal.py
   `- result_plots.py

With this structure you can then render your plots using

pplt input_signal.pdf
pplt result_plots.pdf

Makefiles¶

One key aspect is that each output file is represented by one renderer module Python file. This makes it possible to have a Makefile rule for each plot you need and only re-render the ones that were actually changed.

# Render plots automatically using PPLT
fig/%.pdf: pplt/%.py
        pplt $@

# Build plots when building paper.pdf
paper.pdf: fig/input_signal.pdf fig/result_plots.pdf
    latex paper.tex

Configuration¶

You may set the following values in a pplt/conf.py file:

Processing¶

aliases¶

A dictionary of aliases for your render modules. The key of each entry is the alias name, the value the actual linked-to module.

When using tuples here, the first value is the module and all following values are values passed to main()

aliases = {
    "alias":            "logspec",              # logspec.main()
    "logspec_real":     ("logspec", "real"),    # logspec.main("real")
    "logspec_synth":    ("logspec", "synth"),   # logspec.main("synth")
}

Styling¶

stylesheet¶

The Matplotlib style you wish to use. Use plt.style.available to see what styles you have available.

>>> plt.style.available
[u'dark_background', u'bmh', u'grayscale', u'ggplot', u'fivethirtyeight']

columnwidth¶

the width of your columns. You may resize the figure in LaTeX later on, but the resulting text size depends on a correct setting.

columnwidth = 244.6937  # Get this from LaTeX using \showthe\columnwidth

rc_params¶

A dictionary of values passed on to plt.rcParams.update()

rc_params = {
    'backend': 'ps',
    'axes.labelsize': 9,
    'legend.fontsize': 9,
    'xtick.labelsize': 8,
    'ytick.labelsize': 8,
    'text.usetex': True,
}

Renderer Modules¶

Your code must at least implement a main() function that accepts the matplotlib instance as its only paramter and returns a figure

def main(plt):
    f, ax = plt.subplots(1, 1, figsize=(6, 2))

    ax.plot(...)

    return f

Artists¶

Your function may return the figure alongside a list/tuple of additional artists:

def main(plt):
    f, ax = plt.subplots(1, 1, figsize=(6, 2))

    ax.plot(...)
    lgd = ax.legend(...)

    return f, (lgd,)

RC Settings¶

Your module may define additional RC settings for Matplotlib and Seaborn as well as set a stylesheet and a pre_hook and a post_hook which will run before and after the plotting.

from contextlib import contextmanager

def main(plt):
    f, ax = plt.subplots(1, 1, figsize=(6, 2))

    ax.plot(...)
    lgd = ax.legend(...)

    return f, (lgd,)

sns_params = {
    'font': 'serif',
}

rc_params = {
    'font.size': 9,
}

stylesheet = 'grayscale'

def pre_hook(plt):
    plt.style.use('grayscale')

def post_hook(plt):
    pass

Paper PLT Documentation¶

Installation¶

Setup¶

Makefiles¶

Configuration¶

Processing¶

Styling¶

Renderer Modules¶

Artists¶

RC Settings¶

Indices and tables¶