ChatNoir web UI frontend and search backend.

The search backend is written in Python3 with Django, the frontend in JavaScript with VueJS.

Install dependencies:

npm install

For development purposes, it is most convenient to run a Vite dev server with auto-reload:

npm run serve

The dev server will run at http://localhost:5173 and send requests to the backend server running at http://localhost:8000 (see next section).

A static version of the frontend can be built with:

npm run build

The compiled files are written to chatnoir_ui/dist/ui. This static version can be used for production deployments, where you want to run only the Django backend server (ideally via uWSGI). See the next section for more information.

The backend uses Poetry as a package manager. If you haven't installed it yet, do that first:

# Option 1: pip
python3 -m pip install poetry

# Option 2: system package manager
sudo apt install python3-poetry

Then create a Poetry env and install all necessary Python dependencies:

poetry install

After a fresh installation, you need to create a local configuration file at chatnoir/chatnoir/local_settings.py. You can copy the provided local_settings.example.py and adjust the required settings.

Once the config has been created, initialize the database and the backend superuser (this has to be done only once):

poetry run chatnoir-manage migrate
poetry run chatnoir-manage createsuperuser

The chatnoir-manage command behaves the same way as Django's default manage.py script and should be used in its stead.

If you're not using the Vite dev server, compile a static version of the frontend and collect the static files for Django to serve:

npm run build
yes yes | poetry run chatnoir-manage collectstatic --clear 

Finally, start the backend server:

poetry run chatnoir-serve

The chatnoir-serve command runs any pending migrations and then loads the default chatnoir Django app. If no other port is given, the server will run at http://localhost:8000 and serve the static version of the frontend (if compiled), as well as the API backend.

• If you use the Vite dev server, make sure you have configured Django's CORS_ALLOWED_ORIGINS and CSRF_TRUSTED_ORIGINS properly (see local_settings.example.py).
• The built-in Django server and the Vite dev server should be used for development only. Production deployments should use uWSGI instead. A Dockerfile for a production-ready ChatNoir image is provided in this repository.
• Instead of using poetry run, you can also start an interactive Poetry shell in which you can invoke chatnoir-serve or chatnoir-manage directly:

  poetry shell
  chatnoir-serve

To run any other installed Django app or to pass additional parameters to the server, specify the app name after the chatnoir-serve command (note: you have to pass explicit port names to run multiple apps concurrently):

poetry run chatnoir-serve chatnoir
poetry run chatnoir-serve web_cache 8001
poetry run chatnoir-serve chatnoir_admin 8002
poetry run chatnoir-serve ir_anthology 8003

The parameters are the same as for Django's runserver command (run chatnoir-serve APPNAME --help for more information).

You can also run the ChatNoir front- and backend(s) in Docker:

docker run --rm --init -p 8000:8000 ghcr.io/chatnoir-eu/chatnoir-web

This will run a production-ready uWSGI server inside the container and expose it on port 8000. Run without additional arguments, the ChatNoir server will have no local_settings.py, so requests to it will fail with an error. You can mount a local config into the container like so:

docker run --rm --init -p 8000:8000 \
    -v $(pwd)/local_settings.py:/opt/chatnoir-web/chatnoir/chatnoir/local_settings.py \
    ghcr.io/chatnoir-eu/chatnoir-web

Set the CHATNOIR_APP environment variable to start any of the non-default ChatNoir Django apps:

docker run --rm --init -p 8000:8000 -e CHATNOIR_APP=web_cache ghcr.io/chatnoir-eu/chatnoir-web