Async Signals provides an implementation of Django signals that are fired asynchronously. When you send a Django signal, all receivers are called synchronously. Sometimes you want to just send the signal, and move on. Async Signals let you do just that, in a way that can be easily mixed and matched with normal, synchronous signals. Async Signals uses Celery for signal routing.

You can send an asynchronous signal just like you would a normal one:

>>> from async_signals.signals import request_started
>>> request_started.send(None)

When you fire an asynchronous signal, a task is sent to Celery, which will propagate the signal to any listeners. If Celery is not configured, Async Signals will emit a Warning and trigger the receivers synchronously.

Async Signals are registered exactly like "normal" Django signals. You can either register them manually:

from async_signals.signals import request_finished

request_finished.connect(my_callback)

You can also use the decorator included with Django:

from async_signals.signals import request_finished
from django.dispatch import receiver

@receiver(request_finished)
def my_callback(sender, **kwargs):
    print "Request finished!"

You can define your own signals by calling the AsyncSignal class. :

from async_signals import AsyncSignal

my_signal = AsyncSignal()

You can also specify arguments that will be passed along to your signal's receivers:

from async_signals import AsyncSignal

request_started = AsyncSignal(providing_args=('host', 'path',))

You can route signals to different Celery queues by providing an explicit queue when creating your Signal:

routed_signal = AsyncSignal(queue='high_priority')

The queue name will be used for the signal notification. Note that only one Celery task is created each time a signal is emitted; once picked up by the worker, the receivers are still processed synchronously.

If no queue is specified, the default Celery queue will be used.

Async Signals is made available under a BSD license; see LICENSE for details.