Django includes a “signal dispatcher” which helps allow decoupled applications get notified when actions occur elsewhere in the framework. In a nutshell, signals allow certain senders to notify a set of receivers that some action has taken place. They’re especially useful when many pieces of code may be interested in the same events.
Django provides a set of built-in signals that let user code get notified by Django itself of certain actions. These include some useful notifications:
Sent before or after a model’s save() method is called.
Sent before or after a model’s delete() method is called.
Sent when a ManyToManyField on a model is changed.
Sent when Django starts or finishes an HTTP request.
See the built-in signal documentation for a complete list, and a complete explanation of each signal.
You can also define and send your own custom signals; see below.
To receive a signal, you need to register a receiver function that gets called when the signal is sent by using the Signal.connect() method:
Let’s see how this works by registering a signal that gets called after each HTTP request is finished. We’ll be connecting to the request_finished signal.
First, we need to define a receiver function. A receiver can be any Python function or method:
def my_callback(sender, **kwargs): print "Request finished!"
Notice that the function takes a sender argument, along with wildcard keyword arguments (**kwargs); all signal handlers must take these arguments.
We'll look at senders a bit later, but right now look at the **kwargs argument. All signals send keyword arguments, and may change those keyword arguments at any time. In the case of request_finished, it's documented as sending no arguments, which means we might be tempted to write our signal handling as my_callback(sender).
This would be wrong -- in fact, Django will throw an error if you do so. That's because at any point arguments could get added to the signal and your receiver must be able to handle those new arguments.
There are two ways you can connect a receiever to a signal. You can take the manual connect route:
from django.core.signals import request_finished request_finished.connect(my_callback)
Alternatively, you can use a receiver decorator when you define your receiver:
from django.core.signals import request_finished from django.dispatch import receiver @receiver(request_finished) def my_callback(sender, **kwargs): print "Request finished!"
Now, our my_callback function will be called each time a request finishes.
The receiver decorator was added in Django 1.3.
Where should this code live?
You can put signal handling and registration code anywhere you like. However, you'll need to make sure that the module it's in gets imported early on so that the signal handling gets registered before any signals need to be sent. This makes your app's models.py a good place to put registration of signal handlers.
Some signals get sent many times, but you'll only be interested in receiving a certain subset of those signals. For example, consider the django.db.models.signals.pre_save signal sent before a model gets saved. Most of the time, you don't need to know when any model gets saved -- just when one specific model is saved.
In these cases, you can register to receive signals sent only by particular senders. In the case of django.db.models.signals.pre_save, the sender will be the model class being saved, so you can indicate that you only want signals sent by some model:
from django.db.models.signals import pre_save from django.dispatch import receiver from myapp.models import MyModel @receiver(pre_save, sender=MyModel) def my_handler(sender, **kwargs): ...
The my_handler function will only be called when an instance of MyModel is saved.
Different signals use different objects as their senders; you'll need to consult the built-in signal documentation for details of each particular signal.
In some circumstances, the module in which you are connecting signals may be imported multiple times. This can cause your receiver function to be registered more than once, and thus called multiples times for a single signal event.
If this behavior is problematic (such as when using signals to send an e-mail whenever a model is saved), pass a unique identifier as the dispatch_uid argument to identify your receiver function. This identifier will usually be a string, although any hashable object will suffice. The end result is that your receiver function will only be bound to the signal once for each unique dispatch_uid value.
from django.core.signals import request_finished request_finished.connect(my_callback, dispatch_uid="my_unique_identifier")
Your applications can take advantage of the signal infrastructure and provide its own signals.
All signals are django.dispatch.Signal instances. The providing_args is a list of the names of arguments the signal will provide to listeners.
import django.dispatch pizza_done = django.dispatch.Signal(providing_args=["toppings", "size"])
This declares a pizza_done signal that will provide receivers with toppings and size arguments.
Remember that you're allowed to change this list of arguments at any time, so getting the API right on the first try isn't necessary.
There are two ways to send send signals in Django.
For example, here's how sending our pizza_done signal might look:
class PizzaStore(object): ... def send_pizza(self, toppings, size): pizza_done.send(sender=self, toppings=toppings, size=size) ...
Both send() and send_robust() return a list of tuple pairs [(receiver, response), ... ], representing the list of called receiver functions and their response values.
send() differs from send_robust() in how exceptions raised by receiver functions are handled. send() does not catch any exceptions raised by receivers; it simply allows errors to propagate. Thus not all receivers may be notified of a signal in the face of an error.
send_robust() catches all errors derived from Python's Exception class, and ensures all receivers are notified of the signal. If an error occurs, the error instance is returned in the tuple pair for the receiver that raised the error.
The receiver argument indicates the registered receiver to disconnect. It may be None if dispatch_uid is used to identify the receiver.