A demo app for task queueing using FastAPI and Dramatiq with RabbitMQ as the broker.

The different settings for the app, broker and workers are defined in the .env file at the root of the project. The settings are then used in the code through pydantic-settings. For further details, see the fastapi_dramatiq.settings module.

The settings for running the demo in containers are slightly different compared to running the API and workers locally. Mainly because of the differences in the RabbitMQ URL. See the function get_broker_url() in the module fastapi_dramatiq.actors for details.

The demo requires a message broker, Dramatiq workers and an API. Follow these steps to test with the API and workers running locally.

The test broker can be started with a helper script. It uses the Docker image of RabbitMQ and a docker-compose file. Normal Docker flags can be passed to the script (-d, --build, etc.) and uses the environment variables defined in the `.env`` file at the root of the project.

./scripts/test/run_containers.sh up rabbitmq

The worker can be started with the helper script that uses the Dramatiq CLI utility dramatiq. It spawns worker processes to pop the messages off the queue and sending them to the actor functions (tasks) defined (fastapi_damatiq.actors.tasks).

./scripts/entrypoint_dramatiq.sh

The API can be started for testing with the entrypoint_api.sh script. The settings are loaded from the .env file (any settings with the API_ prefix).

./scripts/entrypoint_api.sh

To run the demo with all 3 components in containers, use the following command:

./scripts/test/run_containers.sh up

The run_containers.sh script accepts Docker/compose arguments, things like --build can be passed as well. The script does not do much more than calling docker-compose with those arguments, using the docker-compose.yaml file in the docker folder.

This will spin up 3 containers:

1. rabbitmq
2. tasks_api
3. workers

The first one is based on the offcial RabbitMQ image and uses the .env file settings for RabbitMQ.

The last two are based on the Dockerfile in the docker directory. Both use the same base layer as they both use functionality of the package partially. Changes in the base layer of that Dockerfile might affect both services.

To see the worker/actor in action or better understand how workers work, run the following command once you have the RabbitMQ broker and Dramatiq workers running (the run_containers.sh script can be used for that).

python3 -m examples.example

NOTE: The settings below are required in the .env for the example to work.

CONTAINERIZED=False
RABBITMQ_HOST="localhost"

It will call the count_words function asynchronously using Dramatiq, as shown in the Dramatiq docs.