Celery
======

Running Celery
^^^^^^^^^^^^^^
Celery is a message queue to allow methods to run asynchronously in a
different thread. For it to work, rabbitmq has to be installed and running.
In Enterobase, Celery is run with the following command in order to be able
to access all of Enterobase's resources. You can specify the number of
threads (the default is 1 and at the time of writing, using multiple threads
caused SQLAlchemy problems)

.. code-block:: bash

    python manage.py runcelery -t 1

This has to be run as a deamon, if using supervisord, the following can be included in the supervisord.conf

.. code-block:: bash

    [program:celery]
    ; path should be the python exe in the enterobase virtual environment
    command=/var/www/venv/bin/python manage.py runcelery
    ;the directory is where manage.py is located in the local installation
    directory=/var/www/entero/
    ;user must have correct permissions
    user=enterobase
    autorestart=true


From time to time rabbitmq will stop working. You can check is status with

.. code-block:: bash

    service rabbitmq-server status

If it is down, you can restart it with

.. code-block:: bash

    service rabbitmq-server start

Celery Setup In Enterobase
^^^^^^^^^^^^^^^^^^^^^^^^^^
The way it is actually setup in Enterobase is very simple - in entero \_\_init\_\_ is the following code:-

.. code-block:: python

    from celery import Celery
    celery = Celery(__name__)
    celery.conf.update(app.config)


The config for celery is in the main enterobase config and assumes that
rabbitmq is running locally

.. code-block:: python

    BROKER_URL = 'amqp://guest:guest@localhost:5672//'
    #CELERY_IMPORTS = ('entero.jobs.jobs', )
    CELERY_RESULT_BACKEND = 'amqp'
    CELERY_ACCEPT_CONTENT = ['pickle']
    USE_CELERY=False

To use celery just import entero.celery and wrap the method in celery.task.

.. code-block:: python

    from entero import celery
    @celery.task
    def my_method():
         #do stuff

**N.B.** If the module where the method resides is not imported when the app
is loaded, it has to be imported in the runcelery method of manage.py in
order for it to be registered.

When calling the a celery method is best to give an alternative if celery is
not running e.g.

.. code-block:: python

    if app.config['USE_CELERY']:
          my_method.apply_async(args=[arg1,arg2],kwargs={"arg3":"value"},queue="entero")
    else:
          my_method(arg1,arg2,arg3="value")