Celery

Offload slow or reliable background work to a Celery task queue.

pip install "endocore[celery]"     # plus a broker, e.g. Redis or RabbitMQ

Register the Celery app

# extensions.py
from endocore.extensions import CeleryExtension

extensions = [
    CeleryExtension(
        app_name="myapp",
        broker="redis://localhost:6379/1",
        backend="redis://localhost:6379/2",
    ),
]

This registers the Celery app as a DI provider under the name celery.

Define tasks

# Services/tasks.py
from endocore.extensions import celery_app

app = celery_app("myapp", broker="redis://localhost:6379/1")

@app.task
def send_email(to: str, subject: str, body: str):
    ...

Run a worker as usual:

celery -A Services.tasks:app worker --loglevel=info

Enqueue from a handler

from Services.tasks import send_email

async def handler(request):                # Api/v1/Signup/Post.py
    data = await request.json()
    send_email.delay(data["email"], "Welcome", "Thanks for signing up!")
    return Response.json({"queued": True}, status=202)

Background tasks vs Celery

Need	Use
Fire-and-forget work tied to one response, best-effort	Response background
Durable, retried, distributed work	Celery

Just the factory, without the extension:

from endocore.extensions import celery_app
app = celery_app("myapp", broker="redis://localhost:6379/1")