Software Testing and Testing Automation with Python

PyTest-MPL

Purpose

• Learn how to use pytest-mpl for testing plots.

Testing Plots

Often we want to ensure that our methods to plot data continue working in the same way. This is especially useful for ensuring that we catch when other libraries change defaults. To do this we create and store baseline images that we will compare against.

• Add an import for pytest

• Create a test function that returns a figure.

• We then run the test function and store the output in a baselines directory in our tests directory.

Instructor Code

import pytest

@pytest.mark.mpl_image_compare(remove_text=True)
def test_plotting_meteogram_defaults():
    """Test default meteogram plotting."""
    # Setup
    url = meteogram.build_asos_request_url('AMW',
                                           start_date=datetime.datetime(2018, 3, 26),
                                           end_date=datetime.datetime(2018, 3, 27))
    df = meteogram.download_asos_data(url)


    # Exercise
    fig, _, _, _ = meteogram.plot_meteogram(df)

    # Verify - Done by decorator when run with -mpl flag

    # Cleanup - none necessary

    return fig

pytest -k test_plotting_meteogram_defaults --mpl-generate-path=tests/baseline

• Run the test with our nominal suite as pytest --mpl

Exercise 5

• Modify the plotting function such that it has a keyword argument to plot horizontal lines (dashed) at N, S, E, W wind directions (i.e. 0, 180, 90, 270 degrees, respectively) This keyword argument should default to false.
• Create a new image test, baseline image, and make sure the test suite passes.

Solution

# Add direction lines if requested
if direction_markers:
    for value_degrees in [0, 90, 180, 270]:
        ax2b.axhline(y=value_degrees, color='k', linestyle='--', linewidth=0.25)

@pytest.mark.mpl_image_compare(remove_text=True)
def test_plotting_meteogram_direction_fiducials():
    """Test meteogram plotting with fiducial lines."""
    # Setup
    url = meteogram.build_asos_request_url('AMW',
                                           start_date=datetime.datetime(2018, 3, 26),
                                           end_date=datetime.datetime(2018, 3, 27))
    df = meteogram.download_asos_data(url)


    # Exercise
    fig, _, _, _ = meteogram.plot_meteogram(df, direction_markers=True)

    # Verify - Done by decorator when run with -mpl flag

    # Cleanup - none necessary

    return fig

Home

This site is open source. Improve this page.