I published a reusable Django application today: django-http-debug, which lets you define mock HTTP endpoints using the Django admin - like /webhook-debug/ for example, configure what they should return and view detailed logs of every request they receive.
Since it's a resuable app, you can add it to any Django project like so:
pip install django-http-debugThen add the following to INSTALLED_APPS in your Django settings:
INSTALLED_APPS = [
# ...
'django_http_debug',
# ...
]Add this to MIDDLEWARE:
MIDDLEWARE = [
# ...
"django_http_debug.middleware.DebugMiddleware",
# ...
]And run ./manage.py migrate to create the necessary database tables.
I used Claude 3.5 Sonnet to build most of this project (prompts in this issue) and then did that again via LLM to have it help me write the tests (full dialog). I then poked around with the tests until I got them to work. Here's what I learned about using pytest-django in the context of a reusable Django application.
My reusable app had the following directory structure inside django-http-debug - initially created using my python-lib cookiecutter template, then I added the Django models.py and views.py and suchlike:
pyproject.toml
README.md
LICENSE
django_http_debug/__init__.py
django_http_debug/models.py
django_http_debug/views.py
django_http_debug/middleware.py
django_http_debug/admin.py
django_http_debug/migrations/__init__.py
django_http_debug/migrations/0001_initial.py
I like pytest, which encourages adding tests in a dedicated tests/ directory.
My pyproject.toml file ended up looking like this, after I added both pytest and pytest-django as test dependencies:
[project]
name = "django-http-debug"
version = "0.2"
description = "Django app for creating database-backed HTTP debug endpoints"
readme = "README.md"
requires-python = ">=3.8"
authors = [{name = "Simon Willison"}]
license = {text = "Apache-2.0"}
classifiers = [
"License :: OSI Approved :: Apache Software License"
]
dependencies = [
"filetype",
"django"
]
[build-system]
requires = ["setuptools"]
build-backend = "setuptools.build_meta"
[project.urls]
Homepage = "https://github.com/simonw/django-http-debug"
Changelog = "https://github.com/simonw/django-http-debug/releases"
Issues = "https://github.com/simonw/django-http-debug/issues"
CI = "https://github.com/simonw/django-http-debug/actions"
[project.optional-dependencies]
test = ["pytest", "pytest-django"]I could then install everything into a virtual environment for my project folder like so:
pip install -e '.[test]'The big challenge here is that Django apps need a fully configured Django project in which to execute! With more help from Claude I created the following structure:
tests/test_django_http_debug.py
tests/test_project/__init__.py
tests/test_project/settings.py
tests/test_project/urls.py
That settings.py file ended up looking like this, after I removed as much stuff from it as possible (the main goal is to test my new middleware):
SECRET_KEY = "django-insecure-test-key"
DEBUG = True
ALLOWED_HOSTS = ["*"]
INSTALLED_APPS = [
"django.contrib.admin",
"django.contrib.auth",
"django.contrib.contenttypes",
# "django.contrib.sessions",
# "django.contrib.messages",
# "django.contrib.staticfiles",
"django_http_debug",
]
MIDDLEWARE = [
"django_http_debug.middleware.DebugMiddleware",
]
ROOT_URLCONF = "tests.test_project.urls"
DATABASES = {
"default": {
"ENGINE": "django.db.backends.sqlite3",
"NAME": ":memory:",
}
}
USE_TZ = TrueAnd urls.py contained this:
from django.contrib import admin
from django.urls import path
urlpatterns = [
path("admin/", admin.site.urls),
]Here's the first of the tests in tests/test_django_http_debug.py - the rest are here:
import pytest
from django_http_debug.models import DebugEndpoint, RequestLog
from django.test.client import Client
@pytest.mark.django_db
def test_debug_view():
assert Client().get("/test/endpoint").status_code == 404
DebugEndpoint.objects.create(
path="test/endpoint",
status_code=200,
content="Test content",
content_type="text/plain",
)
response = Client().get("/test/endpoint")
assert response.status_code == 200
assert response.content == b"Test content"
assert response["Content-Type"] == "text/plain"This uses @pytest.mark.django_db to ensure the Django database is configured (and then cleared) for that test. It uses the Django test client to exercise an endpoint, and the Django ORM to populate the database.
The part that took the longest was figuring out how to make it so that when pytest ran it could find my test_project Django project.
I saw this error a lot:
django.core.exceptions.ImproperlyConfigured: Requested setting INSTALLED_APPS, but settings are not configured. You must either define the environment variable DJANGO_SETTINGS_MODULE or call settings.configure() before accessing settings.
Eventually I figured out the recipe to fix that. I added the following to my pyproject.toml file:
[tool.pytest.ini_options]
DJANGO_SETTINGS_MODULE = "tests.test_project.settings"
pythonpath = ["."]The pythonpath line tells pytest to add the current directory to the Python path before running. Then the DJANGO_SETTINGS_MODULE variable tells it where to find the test project.
Here's the relevant pytest-django documentation.
This all worked! Now when I run pytest in my root folder I see the following:
======================== test session starts =========================
platform darwin -- Python 3.10.10, pytest-8.3.2, pluggy-1.5.0
django: version: 5.1, settings: tests.test_project.settings (from ini)
rootdir: /Users/simon/Dropbox/Development/django-http-debug
configfile: pyproject.toml
plugins: django-4.8.0
collected 5 items
tests/test_django_http_debug.py ..... [100%]
========================= 5 passed in 0.10s ==========================
Created 2024-08-07T16:14:12-07:00, updated 2024-08-13T14:04:54-07:00 · History · Edit