05: Unit Tests and
Provide unit testing for our project's Python code.
As the mantra says, "Untested code is broken code." The Python community has had a long culture of writing test scripts which ensure that your code works correctly as you write it and maintain it in the future. Pyramid has always had a deep commitment to testing, with 100% test coverage from the earliest pre-releases.
Python includes a unit testing framework in its standard library. Over the years a
number of Python projects, such as pytest, have
extended this framework with alternative test runners that provide more
convenience and functionality. The Pyramid developers use
we'll use in this tutorial.
Don't worry, this tutorial won't be pedantic about "test-driven development" (TDD). We'll do just enough to ensure that, in each step, we haven't majorly broken the code. As you're writing your code, you might find this more convenient than changing to your browser constantly and clicking reload.
We'll also leave discussion of pytest-cov for another section.
- Write unit tests that ensure the quality of our code.
- Install a Python package (
pytest) which helps in our testing.
First we copy the results of the previous step, as well as install the
$ cd ..; cp -r debugtoolbar unit_testing; cd unit_testing $ $VENV/bin/pip install -e . $ $VENV/bin/pip install pytest
Now we write a simple unit test in
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
import unittest from pyramid import testing class TutorialViewTests(unittest.TestCase): def setUp(self): self.config = testing.setUp() def tearDown(self): testing.tearDown() def test_hello_world(self): from tutorial import hello_world request = testing.DummyRequest() response = hello_world(request) self.assertEqual(response.status_code, 200)
Now run the tests:
$ $VENV/bin/py.test tutorial/tests.py -q . 1 passed in 0.14 seconds
tests.py imports the Python standard unit testing framework. To make
writing Pyramid-oriented tests more convenient, Pyramid supplies some
pyramid.testing helpers which we use in the test setup and teardown. Our
one test imports the view, makes a dummy request, and sees if the view returns
what we expect.
tests.TutorialViewTests.test_hello_world test is a small example of a
unit test. First, we import the view inside each test. Why not import at the
top, like in normal Python code? Because imports can cause effects that break a
test. We'd like our tests to be in units, hence the name unit testing. Each
test should isolate itself to the correct degree.
Our test then makes a fake incoming web request, then calls our Pyramid view. We test the HTTP status code on the response to make sure it matches our expectations.
Note that our use of
pyramid.testing.tearDown() aren't actually necessary here; they are only
necessary when your test needs to make use of the
config object (it's a
Configurator) to add stuff to the configuration state before calling the view.
- Change the test to assert that the response status code should be
404(meaning, not found). Run
py.testagain. Read the error report and see if you can decipher what it is telling you.
- As a more realistic example, put the
tests.pyback as you found it, and put an error in your view, such as a reference to a non-existing variable. Run the tests and see how this is more convenient than reloading your browser and going back to your code.
- Finally, for the most realistic test, read about Pyramid
Responseobjects and see how to change the response code. Run the tests and see how testing confirms the "contract" that your code claims to support.
- How could we add a unit test assertion to test the HTML value of the response body?
- Why do we import the
hello_worldview function inside the
test_hello_worldmethod instead of at the top of the module?