Margins removed in baseline

[pylang]

When I run or save a modified README code in Jupyter, I get a different image than the generated baseline.

Example code run in a Jupyter cell:

#%%file test.py
%matplotlib inline

import matplotlib.pyplot as plt
import pytest

@pytest.mark.mpl_image_compare
def test_succeeds():
    fig = plt.figure()
    ax = fig.add_subplot(1,1,1)
    ax.plot([1,2,3])
    fig.savefig("test")
    return fig

Running test_succeeds(); in Jupyter gives resulting image and saved test.png with margins:

Next, I generate a test.py file by uncommenting the first line, commenting the second line
and then running the cell

%%file test.py
#%matplotlib inline

...

From a command prompt, in the appropriate directory I run the following to generate the baseline image:

> py.test --mpl-generate-path=baseline test.py

In the baseline folder, a comparable image called test_succeeds.png is created without margins:

Aside from the difference in sizes, I notice the baseline image uses the classic mpl style. Perhaps this is the cause for the absence in margins. I report this as an issue as it is ideal that when testing the baseline image, it should be the same image as shown and saved by Jupyter.

I tested this with Anaconda 4.2.

[astrofrog]

Do you have a matplotlibrc file in ~/.matplotlib? I wonder if you have the bbox_inches='tight' option set by default?

[pylang]

No, I don't have any customization in that file. I tried that latter with fig.tight_layout() as well. I got the same results.

I am unable to confirm this with that latest version of matplotlib (due to this issue while upgrading). I am interested if someone can reproduce these results with matplotlib 2.0.0.

[ConorMacBride]

I believe the differences are due to pytest-mpl using the classic style by default. (Matplotlib changed their default style in v2.0) To use Matplotlib's current default style in pytest-mpl, test functions should be decorated with @pytest.mark.mpl_image_compare(style="default") instead.

---

Most new users today would probably prefer to use Matplotlib's current default style, so maybe we should consider only setting the style if it's explicitly passed to the pytest-mpl decorator? (This is assuming the default style remains stable between major releases of Matplotlib.) This could be a breaking change in pytest-mpl v1.0.0.

[dopplershift]

Jupyter by default uses bbox_inches='tight' with the matplotlib inline support.