matevz-ap/galeriaa

Django Conductor

A simple Railway template project

Features

• Bootstrap
• Dockerfile
• Docker Compose
• Django
• HTMX

Usage

Local

For local development, docker compose up will configure and launch all necessary services.
Additional packages can be added to the requirements/local.txt. Rebuild image if necessary.
Local development depends on two settings files. These are:

website/settings/base.py
website/settings/local.py

Modify base.py when you want those settings reflected on the local and production images, and local.py when you only want to see those settings locally.

Environment customization is made possible through the docker-compose.yaml file, or through an .env file at the
root. When a .env file exists at the root, docker compose will automatically load the variables. Default variables are provided, but they can easily be overridden.

# docker-compose.yaml
# ...
local:
  # ...
  environment:
    DATABASE_URL: postgresql://user:password@postgres/database
    DJANGO_SUPERUSER_EMAIL: [email protected]
    DJANGO_SUPERUSER_PASSWORD: password
    DJANGO_SUPERUSER_USERNAME: local
# ...

By default, a DATABASE_URL is configured that connects to the PostgreSQL Docker service.
To automate superuser creation, the DJANGO_SUPERUSER_EMAIL DJANGO_SUPERUSER_PASSWORD and DJANGO_SUPERUSER_USERNAME environment variables are defined and will be used by the service on launch.
Additional environment variables can also be defined so that they override the defaults found in website/settings/local.py\

A few volumes are also defined so that the server can restart itself when it detects changes in those directories.
Add and remove volumes as needed.

# ...
volumes:
  - ./static:/railway/static
  - ./templates:/railway/templates
  - ./website:/railway/website
# ...

No Docker

If Docker is not an option, you can still run the project, but additional configuration will be necessary.

To begin, create and activate a virtual environment

# System dependent; will not work on every OS
python -m venv venv
source venv/bin/activate

Install local requirements

pip install -r requirements/local.txt

Migrate -- if DATABASE_URL is not supplied as an environment variable, project will default to a SQLite DB

python manage.py migrate

Start server

python manage.py runserver_plus --nostatic

Production

Railway

The simplest way is to deploy the project is to create a New Project through the Railway Dashboard, or CLI, and selecting Deploy a Template
After a few brief dialogs, project should be fully configured and online.

You can also clone the GitHub repository and launch the project using the following steps:

railway link
railway up

Once the project has been deployed to Railway, you can override many of the settings found in
website/settings/base.py or website/settings/production.py through the Railway interface.
Simply define a new environment variable on Railway and wait for the project to be redeployed.

Dockerfile CMD

For every deploy on Railway, or docker compose up, the Dockerfile dynamically builds a sh script that will
then be executed by the relevant service's Docker CMD

Locally, when this script is executed, it begins by executing manage.py migrate, followed by manage.py createsuperuser, if the environment variables are defined, and ultimately manage.py runserver_plus

In production, manage.py migrate will be called first, followed by manage.py createsuperuser and then manage.py collecstatic, manage.py check --deploy, and lastly gunicorn website.wsgi

---

What if you wanted to edit, or add to, the existing commands?

Locally, you have two options:

Execute a command inside a running Docker service:

docker compose exec local ...

Or by modifying the railway.sh script that is generated by the Dockerfile build stages explained below

Since connecting to a running container on Railway is not possible, you must alter the Dockerfile

For example:

If you wanted to add a command to output all the Django settings being used in deployment, then you could alter
the Dockerfile's production build stage RUN command as so:

FROM base as production

# ...

RUN <<CAT cat >> ./railway.sh
  python manage.py collectstatic --no-input
  python manage.py check --deploy
  # ---
  python manage.py diffsettings --all
  # ---
  gunicorn website.wsgi
CAT

# ...

CMD ./railway.sh

The same change could also be performed on the Dockerfile's base build stage, but be aware that

python manage.py migrate --no-input

should generally be the first command that is executed since createsuperuser, collectstatic, etc... will be dependent on it.

You could then execute railway up or git add ... git commit ... git push, to have Railway redeploy the project. If you wanted to see these changes locally, then rebuild the image and rerun docker compose up