Testing

Test Driven Development

We strive to follow a Test Driven Development (TDD) pattern. All models or statistical functions that are added to the main code base are to have tests versus an existing statistical package, if possible.

Introduction to pytest

Like many packages, statsmodels uses the pytest testing system and the convenient extensions in numpy.testing. Pytest will find any file, directory, function, or class name that starts with test or Test (classes only). Test function should start with test, test classes should start with Test. These functions and classes should be placed in files with names beginning with test in a directory called tests.

Running the Test Suite

You can run all the tests by:

>>> import statsmodels.api as sm
>>> sm.test()

You can test submodules by:

>>> sm.discrete.test()

Running Tests using the command line

Test can also be run from the command line by calling pytest. Tests can be run at different levels:

  • Project level, which runs all tests. Running the entire test suite is slow and normally this would only be needed if making deep changes to statsmodels.
pytest statsmodels
  • Folder level, which runs all tests below a folder
pytest statsmodels/regression/tests
  • File level, which runs all tests in a file
pytest statsmodels/regression/tests/test_regression.py
  • Class level, which runs all tests in a class
pytest statsmodels/regression/tests/test_regression.py::TestOLS
  • Test level, which runs a single test. The first example runs a test in a class. The second runs a stand alone test.
pytest statsmodels/regression/tests/test_regression.py::TestOLS::test_missing
pytest statsmodels/regression/tests/test_regression.py::test_ridge

How To Write A Test

NumPy provides a good introduction to unit testing with pytest and NumPy extensions here. It is worth a read for some more details. Here, we will document a few conventions we follow that are worth mentioning. Often we want to test a whole model at once rather than just one function, for example. The following is a pared down version test_discrete.py. In this case, several different models with different options need to be tested. The tests look something like

from numpy.testing import assert_almost_equal
import statsmodels.api as sm
from results.results_discrete import Spector

class CheckDiscreteResults(object):
    """
    res2 are the results. res1 are the values from statsmodels
    """

    def test_params(self):
        assert_almost_equal(self.res1.params, self.res2.params, 4)

    decimal_tvalues = 4
    def test_tvalues(self):
        assert_almost_equal(self.res1.params, self.res2.params, self.decimal_tvalues)

    # ... as many more tests as there are common results

class TestProbitNewton(CheckDiscreteResults):
    """
    Tests the Probit model using Newton's method for fitting.
    """

    @classmethod
    def setup_class(cls):
        # set up model
        data = sm.datasets.spector.load()
        data.exog = sm.add_constant(data.exog)
        cls.res1 = sm.Probit(data.endog, data.exog).fit(method='newton', disp=0)

        # set up results
        res2 = Spector()
        res2.probit()
        cls.res2 = res2

        # set up precision
        cls.decimal_tvalues = 3

    def test_model_specifc(self):
        assert_almost_equal(self.res1.foo, self.res2.foo, 4)

The main workhorse is the CheckDiscreteResults class. Notice that we can set the level of precision for tvalues to be different than the default in the subclass TestProbitNewton. All of the test classes have a @classmethod called setup_class. Otherwise, pytest would reinstantiate the class before every single test method. If the fitting of the model is time consuming, then this is clearly undesirable. Finally, we have a script at the bottom so that we can run the tests should be running the Python file.

Test Results

The test results are the final piece of the above example. For many tests, especially those for the models, there are many results against which you would like to test. It makes sense then to separate the hard-coded results from the actual tests to make the tests more readable. If there are only a few results it’s not necessary to separate the results. We often take results from some other statistical package. It is important to document where you got the results from and why they might differ from the results that we get. Each tests folder has a results subdirectory. Consider the folder structure for the discrete models:

tests/
    __init__.py
    test_discrete.py
    results/
        __init__.py
        results_discrete.py
        nbinom_resids.csv

It is up to you how best to structure the results. In the discrete model example, you will notice that there are result classes based around particular datasets with a method for loading different model results for that dataset. You can also include text files that hold results to be loaded by results classes if it is easier than putting them in the class itself.