This document is for Celery's development version, which can be significantly different from previous releases. Get old docs here: 3.0.
You need four simple steps to use celery with your Django project.
Install the django-celery library:$ pip install django-celery
Add the following lines to settings.py:import djcelery djcelery.setup_loader()
Add djcelery to INSTALLED_APPS.
Create the celery database tables.
If you are using south for schema migrations, you’ll want to:$ python manage.py migrate djcelery
For those who are not using south, a normal syncdb will work:$ python manage.py syncdb
All settings mentioned in the Celery documentation should be added to your Django project’s settings.py module. For example you can configure the BROKER_URL setting to specify what broker to use:
BROKER_URL = 'amqp://guest:guest@localhost:5672/'
If you’re using mod_wsgi to deploy your Django application you need to include the following in your .wsgi module:
import djcelery djcelery.setup_loader()
Tasks are defined by wrapping functions in the @task decorator. It is a common practice to put these in their own module named tasks.py, and the worker will automatically go through the apps in INSTALLED_APPS to import these modules.
For a simple demonstration create a new Django app called celerytest. To create this app you need to be in the directory of your Django project where manage.py is located and execute:
$ python manage.py startapp celerytest
Next you have to add the new app to INSTALLED_APPS so that your Django project recognizes it. This setting is a tuple/list so just append celerytest as a new element at the end
INSTALLED_APPS = ( ..., 'djcelery', 'celerytest', )
After the new app has been created and added to INSTALLED_APPS, you can define your tasks by creating a new file called celerytest/tasks.py:
from celery import task @task() def add(x, y): return x + y
Our example task is pretty pointless, it just returns the sum of two arguments, but it will do for demonstration, and it is referred to in many parts of the Celery documentation.
You have to be consistent in how you import the task module, e.g. if you have project.app in INSTALLED_APPS then you also need to import the tasks from project.app or else the names of the tasks will be different.
In a production environment you will want to run the worker in the background as a daemon - see Running the worker as a daemon - but for testing and development it is useful to be able to start a worker instance by using the celery worker manage command, much as you would use Django’s runserver:
$ python manage.py celery worker --loglevel=info
For a complete listing of the command-line options available, use the help command:
$ python manage.py celery help
Now that the worker is running, open up a new python interactive shell with python manage.py shell to actually call the task you defined:
>>> from celerytest.tasks import add >>> add.delay(2, 2)
Note that if you open a regular python shell by simply running python you will need to import your Django application’s settings by running:
# Replace 'myproject' with your project's name >>> from myproject import settings
The delay method used above is a handy shortcut to the apply_async method which enables you to have greater control of the task execution. To read more about calling tasks, including specifying the time at which the task should be processed see Calling Tasks.
Tasks need to be stored in a real module, they can’t be defined in the python shell or IPython/bpython. This is because the worker server must be able to import the task function.
The task should now be processed by the worker you started earlier, and you can verify that by looking at the worker’s console output.
Calling a task returns an AsyncResult instance, which can be used to check the state of the task, wait for the task to finish or get its return value (or if the task failed, the exception and traceback).
By default django-celery stores this state in the Django database. You may consider choosing an alternate result backend or disabling states alltogether (see Result Backends).
To demonstrate how the results work call the task again, but this time keep the result instance returned:
>>> result = add.delay(4, 4) >>> result.ready() # returns True if the task has finished processing. False >>> result.result # task is not ready, so no return value yet. None >>> result.get() # Waits until the task is done and returns the retval. 8 >>> result.result # direct access to result, doesn't re-raise errors. 8 >>> result.successful() # returns True if the task didn't end in failure. True
If the task raises an exception, the return value of result.successful() will be False, and result.result will contain the exception instance raised by the task.