matplotlib-figure-testing

Test that two matplotlib figures are similar, in the sense that a list of attributes of the figures are the same.

Usage

Figure Comparison

To use these tests, you need only import one function:

from test_figures import assert_similar_figures

Then compare ref_fig and other_fig with

assert_similar_figures(ref_fig, other_fig[, attrs])

Where attrs is a tuple of attributes you would like to compare. Possible entries are:

• "x_data"
• "y_data"
• "legend"
• "linewidth"
• "linestyle"
• "x_scale"
• "y_scale"
• "legend_entries"
• "grid_spec"
• "title"
• "suptitle"
• "x_label"
• "y_label"
• "marker"
• "width" (for rectangle patches)
• "height" (for rectangle patches)
• "position_x" (for rectangle patches)
• "position_y" (for rectangle patches)
• "theta1" (for wedge patches)
• "theta2" (for wedge patches)
• "theta" (for wedge patches)
• "center_x" (for wedge patches)
• "center_y" (for wedge patches)
• "r" (for wedge patches)

If a provided attribute doesn't make sense for the particular object being compared, it will be ignored. If attrs is not provided, all the relevent attributes will be tested.

Figure Capture from functions

Sometimes function generate figures, but don't return them. Using the state based approach to matplotlib, we can still get a handle to these figures:

from test_figures import capture_figures

figs = capture_figures(func, *args, **kwargs)

will call the function func (which generates the figures), and aquire handles to all the figures generated in the function. These figures can then be compared with assert_similar_figures.