django-test-plus 1.0.5

django-test-plus provides useful additions to Django's default TestCase

Useful additions to Django’s default TestCase from Revolution Systems

Rationale

Let’s face it, writing tests isn’t always fun. Part of the reason for that is all of the boilerplate you end up writing. django-test-plus is an attempt to cut down on some of that when writing Django tests. We guarantee it will increase the time before you get carpal tunnel by at least 3 weeks!

Support

Supports: Python 2 and Python 3

Supports Django Versions: 1.4, 1.5, 1.6, 1.7, and 1.8

Usage

Using django-test-plus is pretty easy, simply have your tests inherit from test_plus.test.TestCase rather than the normal django.test.TestCase like so:

from test_plus.test import TestCase

class MyViewTests(TestCase):
    ...

This is sufficient to get things rolling, but you are encouraged to create your own sub-class on a per project basis. This will allow you to add your own project specific helper methods.

For example, if you have a django project named ‘myproject’, you might create the following in myproject/test.py:

from test_plus.test import TestCase as PlusTestCase

class TestCase(PlusTestCase):
    pass

And then in your tests use:

from myproject.test import TestCase

class MyViewTests(TestCase):
    ...

Note that you can also option to import it like this if you want, which is more similar to the regular importing of Django’s TestCase:

from test_plus import TestCase

Methods

reverse(url_name, *args, **kwargs)

When testing views you often find yourself needing to reverse the URL’s name. With django-test-plus there is no need for the from django.core.urlresolvesr import reverse boilerplate. Instead just use:

def test_something(self):
    url = self.reverse('my-url-name')
    slug_url = self.reverse('name-takes-a-slug', slug='my-slug')
    pk_url = self.reverse('name-takes-a-pk', pk=12)

As you can see our reverse also passes along any args or kwargs you need to pass in.

get(url_name, *args, **kwargs)

Another thing you do often is HTTP get urls. Our get() method assumes you are passing in a named URL with any args or kwargs necessary to reverse the url_name. If needed, place kwargs for TestClient.get() in an ‘extra’ dictionary.:

def test_get_named_url(self):
    response = self.get('my-url-name')
    # Get XML data via AJAX request
    xml_response = self.get(
        'my-url-name',
        extra={'HTTP_X_REQUESTED_WITH': 'XMLHttpRequest'})

When using this get method two other things happen for you, we store the last response in self.last\_response and the response’s Context in self.context. So instead of:

def test_default_django(self):
    response = self.client.get(reverse('my-url-name'))
    self.assertTrue('foo' in response.context)
    self.assertEqual(response.context['foo'], 12)

You can instead write:

def test_testplus_get(self):
    self.get('my-url-name')
    self.assertInContext('foo')
    self.assertEqual(self.context['foo'], 12)

It’s also smart about already reversed URLs so you can be lazy and do:

def test_testplus_get(self):
    url = self.reverse('my-url-name')
    self.get(url)
    self.response_200()

If you need to pass query string parameters to your url name, you can do so like this. Assuming the name ‘search’ maps to ‘/search/’ then:

def test_testplus_get_query(self):
    self.get('search', data={'query': 'testing'})

Would GET /search/?query=testing

post(url_name, data, *args, **kwargs)

Our post() method takes a named URL, the dictionary of data you wish to post and any args or kwargs necessary to reverse the url_name. If needed, place kwargs for TestClient.post() in an ‘extra’ dictionary.:

def test_post_named_url(self):
    response = self.post('my-url-name', data={'coolness-factor': 11.0},
                         extra={'HTTP_X_REQUESTED_WITH': 'XMLHttpRequest'})

get_context(key)

Often you need to get things out of the template context, so let’s make that easy:

def test_context_data(self):
    self.get('my-view-with-some-context')
    slug = self.get_context('slug')

assertInContext(key)

You can ensure a specific key exists in the last response’s context by using:

def test_in_context(self):
    self.get('my-view-with-some-context')
    self.assertInContext('some-key')

assertContext(key, value)

We can get context values and ensure they exist, but so let’s also test equality while we’re at it. This asserts that key == value:

def test_in_context(self):
    self.get('my-view-with-some-context')
    self.assertContext('some-key', 'expected value')

response_XXX(response) - status code checking

Another test you often need to do is check that a response has a certain HTTP status code. With Django’s default TestCase you would write:

from django.core.urlresolvers import reverse

def test_status(self):
    response = self.client.get(reverse('my-url-name'))
    self.assertEqual(response.status_code, 200)

With django-test-plus you can shorten that to be:

def test_better_status(self):
    response = self.get('my-url-name')
    self.response_200(response)

django-test-plus provides the following response method checks for you:

- response_200()
- response_201()
- response_302()
- response_403()
- response_404()

All of which take an option Django test client response as their only argument. If it’s available, the response_XXX methods will use the last response. So you can do:

def test_status(self):
    self.get('my-url-name')
    self.response_200()

Which is a bit shorter.

get_check_200(url_name, *args, **kwargs)

GETing and checking views return status 200 is so common a test this method makes it even easier:

def test_even_better_status(self):
    response = self.get_check_200('my-url-name')

make_user(username)

When testing out views you often need to create various users to ensure all of your logic is safe and sound. To make this process easier, this method will create a user for you:

def test_user_stuff(self)
    user1 = self.make_user('u1')
    user2 = self.make_user('u2')

NOTE: This work properly with version of Django prior to 1.6 and will use your own User class if you have created your own User model.

If creating a User in your project is more complicated, say for example you removed the username field from the default Django Auth model you can provide a Factory Boy factory to create it or simply override this method on your own sub-class.

To use a Factory Boy factory simply create your class like this:

from test_plus.test import TestCase
from .factories import UserFactory


class MySpecialTest(TestCase):
    user_factory = UserFactory

    def test_special_creation(self):
        user1 = self.make_user('u1')

NOTE: Users created by this method will always have their password set to the string ‘password’ in order to ease testing.

Authentication Helpers

assertLoginRequired(url_name, *args, **kwargs)

It’s pretty easy to add a new view to a project and forget to restrict it to be login required, this method helps make it easy to test that a given named URL requires auth:

def test_auth(self):
    self.assertLoginRequired('my-restricted-url')
    self.assertLoginRequired('my-restricted-object', pk=12)
    self.assertLoginRequired('my-restricted-object', slug='something')

login context

Along with ensuing a view requires login and creating users, the next thing you end up doing is logging in as various users to test our your restriction logic. This can be made easier with the following context:

def test_restrictions(self):
    user1 = self.make_user('u1')
    user2 = self.make_user('u2')

    self.assertLoginRequired('my-protected-view')

    with self.login(username=user1.username, password='password'):
        response = self.get('my-protected-view')
        # Test user1 sees what they should be seeing

    with self.login(username=user2.username, password='password'):
        response = self.get('my-protected-view')
        # Test user2 see what they should be seeing

Since we’re likely creating our users using make_user() from above, we can assume the password is ‘password’ so the login context assumes that for you as well, so you can do:

def test_restrictions(self):
    user1 = self.make_user('u1')

    with self.login(username=user1.username):
        response = self.get('my-protected-view')

We can also derive the username if we’re using make_user() so we can shorten that up even further like this:

def test_restrictions(self):
    user1 = self.make_user('u1')

    with self.login(user1):
        response = self.get('my-protected-view')

Ensuring low query counts

assertNumQueriesLessThan(number) - context

Django provides assertNumQueries which is great when your code generates generates a specific number of queries. However, if due to the nature of your data this number can vary you often don’t attempt to ensure the code doesn’t start producing a ton more queries than you expect:

def test_something_out(self):

    with self.assertNumQueriesLessThan(7):
        self.get('some-view-with-6-queries')

NOTE: This isn’t possible in versions of Django prior to 1.6, so the context will run your code and assertions and issue a warning that it cannot check the number of queries generated.

assertGoodView(url_name, *args, **kwargs)

This method does a few of things for you, it:

• Retrieves the name URL

• Ensures the view does not generate more than 50 queries

• Ensures the response has status code 200

• Returns the response

Often a wide sweeping test like this is better than no test at all. You can use it like this:

def test_better_than_nothing(self):
    response = self.assertGoodView('my-url-name')