Align the size of the colorbar with matplotlib

What is this

When I make a graph with matplotlib, the color bar sticks out of the graph and it's awkward. I will show you how to align the color bars on the graph.

Various methods

1: Plot normally

colorbar1.py

import numpy
import matplotlib.pyplot

x = numpy.arange(10)
y = numpy.arange(10)
X, Y = numpy.meshgrid(x, y)
Z = X**2. + Y**2.

fig = matplotlib.pyplot.figure()
ax = fig.add_subplot(111)
im = ax.imshow(Z, interpolation='none')
fig.colorbar(im)

fig.savefig('colorbar1.png')

An example where the height of the main graph and the color bar happened to be the same.

2: Plot normally (sorry version)

colorbar2.py

import numpy
import matplotlib.pyplot

x = numpy.arange(15)
y = numpy.arange(10)
X, Y = numpy.meshgrid(x, y)
Z = X**2. + Y**2.

fig = matplotlib.pyplot.figure()
ax = fig.add_subplot(111)
im = ax.imshow(Z, interpolation='none')
fig.colorbar(im)

fig.savefig('colorbar2.png')

Once the aspect ratio of the main graph collapses, the color bars stick out up and down. sad. .. ..

3: Use mpl_toolkits.axes_grid1

colorbar3.py

import numpy
import matplotlib.pyplot
import mpl_toolkits.axes_grid1

x = numpy.arange(15)
y = numpy.arange(10)
X, Y = numpy.meshgrid(x, y)
Z = X**2. + Y**2.

fig = matplotlib.pyplot.figure()
ax = fig.add_subplot(111)
divider = mpl_toolkits.axes_grid1.make_axes_locatable(ax)
cax = divider.append_axes('right', '5%', pad='3%')
im = ax.imshow(Z, interpolation='none')
fig.colorbar(im, cax=cax)

fig.savefig('colorbar3.png')

The most common example on the net. How to use make_axes_locatable and append_axes in mpl_toolkits.axes_grid1. This is the solution. It was good.

4: But if you specify projection, ...

colorbar4.py

import numpy
import matplotlib.pyplot
import mpl_toolkits.axes_grid1

x = numpy.arange(15)
y = numpy.arange(10)
X, Y = numpy.meshgrid(x, y)
Z = X**2. + Y**2.

fig = matplotlib.pyplot.figure()
ax = fig.add_subplot(111, projection='polar')
divider = mpl_toolkits.axes_grid1.make_axes_locatable(ax)
cax = divider.append_axes('right', '5%', pad='3%')
im = ax.imshow(Z, interpolation='none')
fig.colorbar(im, cax=cax)

fig.savefig('colorbar4.png')

The color bar is not displayed! ?? Because type (cax) is matplotlib.projections.polar.PolarAxes. append_axes () does not have a projection argument and seems to use the projection of ax as it is. Is it possible to change the projection somehow? .. ..

5: Change cax position by yourself

colorbar4.py

import io
import numpy
import matplotlib.pyplot

x = numpy.arange(15)
y = numpy.arange(10)
X, Y = numpy.meshgrid(x, y)
Z = X**2. + Y**2.

fig = matplotlib.pyplot.figure()
ax = fig.add_subplot(111)
im = ax.imshow(Z, interpolation='none')
cax = fig.colorbar(im)

fig.savefig(io.BytesIO())

ax_pos = ax.get_position()
cax_pos0 = cax.ax.get_position()
cax_pos1 = [cax_pos0.x0, ax_pos.y0, cax_pos0.x1 - cax_pos0.x0, ax_pos.y1 - ax_pos.y0]
cax.ax.set_position(cax_pos1)

fig.savefig('colorbar5.png')

I was able to get the position of ax with get_position () and align it vertically using set_position () of cax. However, since ax.get_position () has a specification (?) That is not updated until fig.savefig () is executed, fig.savefig () will be called twice. This method is used in fig.savefig (bbox_inches ='tight').