Test Procrastinate interactions in a Django project

Unit tests

If you want to be able to run your code that uses Procrastinate in a unit test, you can use the procrastinate.testing.InMemoryConnector.

Here’s an example pytest fixture you can write in your conftest.py:

import pytest
from procrastinate import testing
from procrastinate.contrib.django import procrastinate_app

@pytest.fixture
def in_memory_app(monkeypatch):
    app = procrastinate_app.current_app.with_connector(
        testing.InMemoryConnector()
    )
    monkeypatch.setattr(procrastinate_app, "current_app", app)
    return app

Note

procrastinate.contrib.django.procrastinate_app.current_app is not exactly documented but whatever instance this variable points to is what will be returned as procrastinate.contrib.django.app. It’s probably a bad idea to use this outside of tests.

Integration tests

You can use Procrastinate normally with procrastinate.contrib.django.app in your tests, though registering new tasks or loading tasks from blueprints within your tests might lead to test isolation issues. If you need to do that, you may want to use the trick described above to have a dedicated app for each relevant test.

You can use the procrastinate models to check if the jobs have been created as expected.

from procrastinate.contrib.django.models import ProcrastinateJob

def test_my_task(app):
    # Run the task
    app.defer("my_task", args=(1, 2))

    # Check the job has been created
    assert ProcrastinateJob.objects.filter(task_name="my_task").count() == 1

Making the models writable in tests

If you need to write to the procrastinate models in your tests, a setting is provided: PROCRASTINATE_READONLY_MODELS. If set to False, the models will be writable. You may want to use the settings fixture from pytest-django to set this setting only for the tests that need it.

@pytest.fixture
def procrastinate_writable_models(settings):
    settings.PROCRASTINATE_READONLY_MODELS = False

Warning

This setting is only intended for tests.