Skip to content

Latest commit

 

History

History
224 lines (170 loc) · 6.66 KB

README.md

File metadata and controls

224 lines (170 loc) · 6.66 KB

Latest Version Build Status Code Coverage Code Climate

Django Mock Queries

A library for mocking Django queryset functions in memory for testing

Features

  • QuerySet style support for method chaining
  • Filtering with Q objects
  • Aggregates generation
  • CRUD functions
  • Field lookups
  • django-rest-framework serializer asserts

Examples

from django.db.models import Avg, Q
from django_mock_queries.query import MockSet, MockModel

qs = MockSet(
    MockModel(mock_name='john', email='[email protected]'),
    MockModel(mock_name='jeff', email='[email protected]'),
    MockModel(mock_name='bill', email='[email protected]'),
)

print [x for x in qs.all().filter(email__icontains='gmail.com').select_related('address')]
# Outputs: [john, bill]

qs = MockSet(
    MockModel(mock_name='model s', msrp=70000),
    MockModel(mock_name='model x', msrp=80000),
    MockModel(mock_name='model 3', msrp=35000),
)

print qs.all().aggregate(Avg('msrp'))
# Outputs: {'msrp__avg': 61666}

qs = MockSet(
    MockModel(mock_name='model x', make='tesla', country='usa'),
    MockModel(mock_name='s-class', make='mercedes', country='germany'),
    MockModel(mock_name='s90', make='volvo', country='sweden'),
)

print [x for x in qs.all().filter(Q(make__iexact='tesla') | Q(country__iexact='germany'))]
# Outputs: [model x, s-class]

qs = MockSet(cls=MockModel)
print qs.create(mock_name='my_object', foo='1', bar='a')
# Outputs: my_object

print [x for x in qs]
# Outputs: [my_object]

Test function that uses Django QuerySet:

"""
Function that queries active users
"""
def active_users(self):
    return User.objects.filter(is_active=True).all()

"""
Test function applies expected filters by patching Django's user model Manager or Queryset with a MockSet
"""
from django_mock_queries.query import MockSet, MockModel


class TestApi(TestCase):
    users = MockSet()
    user_objects = patch('django.contrib.auth.models.User.objects', users)

    @user_objects
    def test_api_active_users_filters_by_is_active_true(self):
        self.users.add(
        	MockModel(mock_name='active user', is_active=True),
        	MockModel(mock_name='inactive user', is_active=False)
        )

        for x in self.api.active_users():
        	assert x.is_active

Test django-rest-framework model serializer:

"""
Car model serializer that includes a nested serializer and a method field
"""
class CarSerializer(serializers.ModelSerializer):
    make = ManufacturerSerializer()
    speed = serializers.SerializerMethodField()

    def get_speed(self, obj):
        return obj.format_speed()

    class Meta:
        model = Car
        fields = ('id', 'make', 'model', 'speed',)

"""
Test serializer returns fields with expected values and mock the result of nested serializer for field make
"""
def test_car_serializer_fields(self):
    car = Car(id=1, make=Manufacturer(id=1, name='vw'), model='golf', speed=300)

    values = {
        'id': car.id,
        'model': car.model,
        'speed': car.formatted_speed(),
    }

    assert_serializer(CarSerializer) \
        .instance(car) \
        .returns('id', 'make', 'model', 'speed') \
        .values(**values) \
        .mocks('make') \
        .run()

Full Example

There is a full Django application in the examples/users folder. It shows how to configure django_mock_queries in your tests and run them with or without setting up a Django database. Running the mock tests without a database can be much faster when your Django application has a lot of database migrations.

To run your Django tests without a database, add a new settings file, and call monkey_patch_test_db(). Use a wildcard import to get all the regular settings as well.

# settings_mocked.py
from django_mock_queries.mocks import monkey_patch_test_db

from users.settings import *

monkey_patch_test_db()

Then run your Django tests with the new settings file:

./manage.py test --settings=users.settings_mocked

Here's the pytest equivalent:

pytest --ds=users.settings_mocked

That will run your tests without setting up a test database. All of your tests that use Django mock queries should run fine, but what about the tests that really need a database?

ERROR: test_create (examples.users.analytics.tests.TestApi)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/.../examples/users/analytics/tests.py", line 28, in test_create
    start_count = User.objects.count()
  [...]
NotSupportedError: Mock database tried to execute SQL for User model.

If you want to run your tests without a database, you need to tell Django to skip the tests that need a database. You can do that by putting a skip decorator on the test classes or test methods that need a database.

@skipIfDBFeature('is_mocked')
class TestApi(TestCase):
    def test_create(self):
        start_count = User.objects.count()

        User.objects.create(username='bob')
        final_count = User.objects.count()

        self.assertEqual(start_count + 1, final_count)

Installation

$ pip install django_mock_queries

Contributing

Anything missing or not functioning correctly? PRs are always welcome! Otherwise, you can create an issue so someone else does it when time allows.

You can follow these guidelines:

  • Fork the repo from this page
  • Clone your fork:
$ git clone https://github.com/{your-username}/django-mock-queries.git
$ cd django-mock-queries
$ git checkout -b feature/your_cool_feature
  • Implement feature/fix
  • Add/modify relevant tests
  • Run tox to verify all tests and flake8 quality checks pass
$ tox
  • Commit and push local branch to your origin
$ git commit . -m "New cool feature does this"
$ git push -u origin HEAD
  • Create pull request

TODO

  • Add docs as a service like readthedocs with examples for every feature
  • Add support for missing QuerySet methods/Field lookups/Aggregation functions:
    • Methods that return new QuerySets: annotate, reverse, none, extra, raw

    • Methods that do not return QuerySets: bulk_create, in_bulk, as_manager

    • Field lookups: search

    • Aggregation functions: StdDev, Variance