subplots-from-axsize 0.1.9

subplots_from_axsize

This package provides a single function, subplots_from_axsize(), which is based on matplotlib's subplots() and adjust_subplots() functions, but:

• figsize is replaced by axsize
• left, bottom, right, top use inches instead of fractions
• wspace, hspace use inches instead of fractions
• axsize, wspace, hspace can take lists as arguments (see example #2 below)

rationale

Specifying sizes directly makes it easier to achieve consistent ax sizes across figures. Suppose you're plotting various time series data and you would like consistent inches / week on your x axis. Or you carefully crafted your figures but suddenly you need a little extra space for you labels. Or you would want to move the axes a little bit apart and not rescale everything else in the whole world simultaneously.

getting started

The package is available on PyPi.

example #1

import matplotlib.pyplot as plt
from subplots_from_axsize import subplots_from_axsize

fig, ax = subplots_from_axsize(
    axsize=(4, 3),
    left=0.9, bottom=0.5, top=0.3, right=0.2,
)

ax.set_xlabel('x label')
ax.set_ylabel('y label\nbut much\nmuch longer')
ax.set_title('important!')
fig.patch.set_facecolor('#ffbbff')

Since the default dpi in matplotlib is 100, the axis is exactly 400 by 300 pixels. The left margin, for example, is 0.9 × 100 = 90 pixels. Changing the margins does not change the axis size (you have my promise).

example #2

Additionally, subplots_from_axsize makes it easy to create multiple axes.

import matplotlib.pyplot as plt
from subplots_from_axsize import subplots_from_axsize

fig, axs = subplots_from_axsize(
    axsize=([1, 3], 1),
    hspace=[0.5, 1.0],
    left=0.4, bottom=0.3,
)
fig.patch.set_facecolor('#ffbbff')

The number of columns (or rows) is inferred automatically based on three arguments: nrows (if given), length of axsize[0] (if it is a list), length of wspace + 1 (if it is a list). If none of the conditions is met, the default is 1. If they disagree you get an error (xor create an Issue).