Fastapi service for SimPy as Background tasks

The goal here is to create a FastAPI service that can execute long-running SimPy simulations.

The simulations are executed using Celery workers, therefore not blocking the API itself and set up to be scalable in the future. Initiating a simulating through the API service will start a task on the worker, and return the task_id to the user. When the Task is finished, it can be collected through the task_id.

Additionally, while the Task is running, we will want to publish simulation events to a pub/sub stream. Currently, this example can publish the events during simulation but has not been tested with a consumer.

Next steps are to think about how to publish simulation events properly and persistent.

Setup

Create a virtual environment with conda and activate it:

conda env create -f environment.yml
conda activate simpy-fastapi-service

uvicorn src.simpy_fastapi_service.main:app

Docker compose

Deploy the whole stack with docker-compose.

docker-compose up

Then access the components through these links:

• API docs Swagger UI to execute endpoints: http://127.0.0.1:8080/docs
• RabbitMQ to view all message queues: http://127.0.0.1:15672
  • user/password = guest/guest
• Flower UI to track tasks: http://127.0.0.1:5555

Development

Install pre-commit

Using pre-commit we make sure to always apply checks locally before commits can be pushed to git. Check the .pre-commit-config.yml for details.

Run pre-commit install to register the configuration with your local repository.

We always apply at least

• ruff to lint and autofix our code
• mypy to statically check type-hints