spakky-celery 6.1.0

Celery DI/AOP integration plugin for Spakky Framework

spakky-celery

Celery integration plugin for Spakky Framework.

Provides AOP-based automatic task dispatch and periodic schedule registration — methods decorated with @task or @schedule are intercepted and routed to Celery without explicit dispatcher calls.

Installation

pip install spakky-celery

Requires: spakky-task (installed automatically as a dependency)

Features

• AOP-based dispatch: @task methods are intercepted by aspects — no manual dispatch() calls
• Broker dispatch: All @task calls are sent to the Celery broker via send_task()
• Schedule registration: @schedule methods are registered to Celery Beat automatically
• Worker-context detection: Uses a context key to prevent re-dispatch inside workers
• Auto-registration: @TaskHandler pods are scanned and registered as Celery tasks automatically
• Full configuration: Broker URL, serializer, timezone, and more via environment variables

Configuration

Set environment variables with the SPAKKY_CELERY__ prefix:

Variable	Default	Description
SPAKKY_CELERY__BROKER_URL	(required)	Celery broker URL (e.g., amqp://user:pass@host:5672//)
SPAKKY_CELERY__RESULT_BACKEND	None	Result backend URL. None disables result storage
SPAKKY_CELERY__APP_NAME	spakky-celery	Celery application name
SPAKKY_CELERY__TASK_SERIALIZER	json	Task message serializer (json, pickle, yaml, msgpack)
SPAKKY_CELERY__RESULT_SERIALIZER	json	Result serializer
SPAKKY_CELERY__ACCEPT_CONTENT	["json"]	Accepted content types
SPAKKY_CELERY__TIMEZONE	UTC	IANA timezone (e.g., Asia/Seoul)
SPAKKY_CELERY__ENABLE_UTC	true	Use UTC for internal datetime handling

Usage

1. Define task handlers

from datetime import time, timedelta

from spakky.task import TaskHandler, Crontab, Weekday, task, schedule


@TaskHandler()
class EmailTaskHandler:
    @task
    def send_email(self, to: str, subject: str, body: str) -> None:
        """Dispatched to Celery broker via send_task()."""
        send_smtp(to, subject, body)


@TaskHandler()
class MaintenanceHandler:
    @schedule(interval=timedelta(minutes=30))
    def health_check(self) -> None:
        """Registered as Celery Beat periodic task — runs every 30 minutes."""
        ...

    @schedule(at=time(3, 0))
    def daily_cleanup(self) -> None:
        """Runs daily at 03:00."""
        ...

    @schedule(crontab=Crontab(weekday=Weekday.MONDAY, hour=9))
    def weekly_report(self) -> None:
        """Runs every Monday at 09:00."""
        ...

2. Bootstrap the application

from spakky.core.application.application import SpakkyApplication
from spakky.core.application.application_context import ApplicationContext

import spakky.plugins.celery

app = (
    SpakkyApplication(ApplicationContext())
    .load_plugins(include={spakky.plugins.celery.PLUGIN_NAME})
    .scan()
    .start()
)

3. Call task methods normally

handler = app.container.get(EmailTaskHandler)

# Dispatched to broker — AOP intercepts and calls send_task()
handler.send_email("[e-mail]", "Hello", "World")

The AOP aspect intercepts the calls and routes them to Celery automatically.

Dispatch Behavior

Decorator	Behavior	Celery API
@task	Dispatch to broker on each call	send_task()
@schedule	Register in Celery Beat	beat_schedule

Components

Component	Description
CeleryConfig	Configuration loaded from environment variables
CeleryPostProcessor	Scans @TaskHandler pods and registers methods as Celery tasks/schedules
CeleryTaskDispatchAspect	AOP aspect intercepting sync @task calls
AsyncCeleryTaskDispatchAspect	AOP aspect intercepting async @task calls

Errors

Error	Description
InvalidScheduleRouteError	ScheduleRoute has no valid schedule specification

Related Packages

• spakky-task: Core task abstractions (@TaskHandler, @task, @schedule, Crontab)
• spakky-rabbitmq: RabbitMQ event transport (can also be used as Celery broker)

License

MIT License