Django Project with Docker, Pipenv, Pre-Commit, and a Dockerized Postgres

Configure Pipenv and Pre-Commit

init project repo

mkdir django-project-configuration
cd !$
git init

Create a project directory and initialize a git repository in it.

**/__pycache__/**
*pyc
*env
**/.vscode/**

Create a project .gitignore file. This excludes python cache files and directories, as well as an .env files and the .vscode directory.

instantiate pipenv virtualenv and install dev packages

pipenv install --python 3.8

pipenv install --dev pre-commit pytest-django

Create a pipenv virtual environment with a python 3.8 interpreter.
Install the pre-commit and pytest-django packages as dev dependencies.

The pre-commit package allows us to easily install third party git hooks. The pytest-django package allows us to use pytest with django.

create a pre-commit config for your project: .pre-commit-config.yaml

repos:
-   repo: https://github.com/psf/black
    rev: 19.3b0
    hooks:
    - id: black
-   repo: https://github.com/pre-commit/pre-commit-hooks
    rev: v1.2.3
    hooks:
    - id: flake8
      args: [--max-line-length=120]

This pre-commit config file will run hooks on every commit.
The first hook applies black code formatting to the staged files. If there is a diff after the hook runs, meaning that the code was not formatted according to black’s guidelines, it will fail and we will need to stage the changes it made.

The second hook applies linting to our staged code with flake8, a popular python code linter. This helps us catch things like typos before committing them.

pipenv shell

pre-commit install

pre-commit run

Enter the pipenv virtual environment and install the pre-commit hooks to your local git repo.

you can confirm things are configured correctly by typing pre-commit run.

commit initial pre-commit config

git add .pre-commit-config.yaml
git commit -v

Commit your config. I always commit with the verbose -v flag so I can see what is staged for commit before actually making the commit.

It requires you to be familiar with a text-based editor like vim though, so if you aren’t comfortable with that, forget about the -v flag for now and simply use:

git commit --message "Add pre-commit config"

Install Django, Create Project, and Configure Settings

install django and start project

pipenv install django

mkdir src

cd src

django-admin.py startproject webapp .

Install django and use the django-admin command to start a project called webapp in a src directory.

install additional packages

pipenv install dj-database-url conversion psycopg2-binary django_command_overrides

Let’s install a few additional packages we will need for our next steps.

  • We will use dj-database-url to configure our database
  • conversion is a really helpful package written by Roberto Aguilar that converts environment variables into python types
  • psycopg2-binary is the driver that will allow us to work with a postgres database
  • django_command_overrides implements a custom template when we eventually run our startapp command

src/webapp/settings.py

import conversion
import dj_database_url
...
INSTALLED_APPS.append("django_command_overrides")
...
SECRET_KEY = os.environ["DJANGO_SECRET_KEY"]

DEBUG = conversion.convert_bool(os.environ.get("DJANGO_DEBUG", "False"))
...
# Database
DATABASES = {"default": {}}

database_info = dj_database_url.config()

if database_info:
    DATABASES["default"] = database_info

Now open your settings.py file and add some configuration info.

  1. Derive your SECRET_KEY from an environment variable. This aligns with a best practice of never committing secrets to a git repository.
  2. Key django’s DEBUG mode off on an environment variable as well. Set the default to False. Notice how we’re using the conversion package to convert the string False into a python bool object.
  3. Add django_command_overrides to INSTALLED_APPS so we can take advantage of a custom startapp command later on.
  4. Use dj_database_url to define your database connection. This package derives an entire django database configuration from a single DATABASE_URL environment variable. I’ll show you what this environment variable should look like in a moment.

commit a pristine copy of your django project

cd ..
git add src/
git commit -v

Note that when we commit this, our black pre-commit hook will apply black code for matting to our staged files and fail because there’s now a diff between what is staged and what the files actually contain.

To fix this we need to re-stage our files, now correctly formatted by black, for commit.

Now that we have added some basic configuration to our settings module, let’s make a pristine commit of our project.

Dockerize Application For Development and Utilize Dockerized Postgres

Dockerfile

FROM python:3.8 as base

RUN pip install pipenv

ENV PROJECT_DIR /usr/local/src/webapp
ENV SRC_DIR ${PROJECT_DIR}/src

COPY Pipfile Pipfile.lock ${PROJECT_DIR}/

WORKDIR ${PROJECT_DIR}

ENV PYTHONUNBUFFERED=1

FROM base as dev

# this is a dev image build, so install dev packages
RUN pipenv install --system --dev

COPY ./src ${SRC_DIR}/

WORKDIR ${SRC_DIR}

CMD ["python", "manage.py", "runserver", "0.0.0.0:8000"]

Create a Dockerfile. Here we use a multi-stage build that generates a dev image which we can use for local development.

.env

DJANGO_SECRET_KEY=local
DATABASE_URL=postgres://test:test@postgres/test
POSTGRES_DB=test
POSTGRES_USER=test
POSTGRES_PASSWORD=test

Create a .env file and populate it. Our docker-compose file will take advantage of these variables in a moment.
Note the DATABASE_URL environment variable which the dj_database_url package will look for to build a database connection.

docker-compose.local.yml

version: "3.4"
services:
  app:
    build:
      context: .
        target: dev
    working_dir: /opt/project/src
    ports:
      - 8000:8000
    environment:
      - DJANGO_SECRET_KEY
      - DATABASE_URL
      - PYTHONPATH=/opt/project
    volumes:
      - .:/opt/project
  postgres:
    image: postgres:9.6
    volumes:
      - pg-data:/var/lib/postgresql/data
    environment:
      - POSTGRES_DB
      - POSTGRES_USER
      - POSTGRES_PASSWORD

volumes:
  pg-data:
    driver: local

Create a docker-compose.local.yml file that defines two services: a postgres service, and an app service.

The postgres service will run a containerized postgres instance for us that we can use like any other database with our django applicaiton.

The app service will build the image defined in our Dockerfile and inject the DJANGO_SECRET_KEY and DATABASE_URL environment variables.

Rather than using the application files in the docker image, we mount our project root into the container and set its src as our working directory.

This allows us to edit and run files on the fly without rebuilding our image every time.

We add our new working directory to our python path via the PYTHONPATH environment variable.

run local docker compose bash

docker-compose -f docker-compose.local.yml run --rm app bash

python manage.py startapp ebook_store

Now test that we can run everything as expected by getting a shell inside our container.To test that we can run our image locally as expected.

exit container
exit

confirm app was created

ls
ls src

Because we mounted in our local directory to the docker container we can see the files were created on our local machine.

commit Dockerfile, docker-compose.local.yml, and Pipfiles

git add Pipfile*
git commit -v

git add Dockerfile docker-compose.local.yml
git commit -v

To wrap up, commit your Pipfiles, Dockerfile, and docker compose configuration.

In summary, you now have a django project that is itself containerized for local development, talks to a containerized instance of postgres, and runs git pre-commit hooks to enforce code formatting and linting on every commit.

Nice work!

Leave a Reply

Your email address will not be published.