Convert matplotlib data units to normalized units

Question

Does anyone know how to convert matplotlib data units into normalized units?

The reason that I need it is that I need to create a subplot on top of another plot. An

Answer 1

Here's one way to do it:

inner axes printed at 0.5, 2.5, 1.0, 0.3 (in outer axes coords)

You basically need two transformations -- one from src-coords to display, and one from display to dest-coord. From the docs there seems to be no direct way:
http://matplotlib.org/users/transforms_tutorial.html

bb_data = Bbox.from_bounds(0.5, 2.5, 1.0, 0.3)
disp_coords = ax.transData.transform(bb_data)
fig_coords = fig.transFigure.inverted().transform(disp_coords)

ax and fig both carry transformer with them -- to display-coords!
If you call inverted on them, you get an transformer for the inverse direction.

Here's the full code for the above example:

import matplotlib.pyplot as plt
from matplotlib.transforms import Bbox

plt.plot([0,2], [2,4])
fig = plt.gcf()
ax = plt.gca()

bb_data = Bbox.from_bounds(0.5, 2.5, 1.0, 0.3)
disp_coords = ax.transData.transform(bb_data)
fig_coords = fig.transFigure.inverted().transform(disp_coords)

fig.add_axes(Bbox(fig_coords))

plt.show()