Celery is a distributed task queue library for Python. It accepts some of its configuration via environment variables--but some configuration needs to be specified as Python code.
new-celery-config is a Python package that lets you set any top-level Celery key using an environment variable containing YAML.
The latest stable can be installed via pip:
python3 -m pip install new-celery-config
To set configuration values, you must set an environment variables for each top-level key (as documented in the Celery documentation).
Each environment variable is prefixed with NEW_CELERY_, followed by the config key name in lowercase. The value for each environment variable must be valid YAML (or JSON--remember that JSON is a subset of YAML).
You must also set the environment variable CELERY_CONFIG_MODULE to new_celery_config.as_module to enable Celery to read all of the other environment variables that you have set.
For example, setting these environment variables in the shell looks like:
export CELERY_CONFIG_MODULE=new_celery_config.as_module
export NEW_CELERY_broker_url='transport://userid:password@hostname:port/virtual_host'
export NEW_CELERY_broker_transport_options='{"visibility_timeout": 36000}'And in your Python code, initialize the Celery object as follows:
app = Celery()If you want to change the name of the CELERY_CONFIG_MODULE, you can use the config_from_envvar function. For example:
export ARBITRARY_CELERY_CONFIG_MODULE=new_celery_config.as_moduleapp.config_from_envvar("ARBITRARY_CELERY_CONFIG_MODULE")You can test that the configuration works by examining the app.conf object:
print(app.conf.broker_transport_options)
# prints out {'visibility_timeout': 36000}Celery also accepts configuration in the form of a Python object. If you prefer this way, you can give Celery a new_celery_config.Config object. For example:
from celery import Celery
import new_celery_config
app = Celery()
app.config_from_object(new_celery_config.Config())If you want to make changes to new-celery-config, you can clone this repository. You can run make in the root directory to show commands relevant to development.
- For example:
make fmtautomatically formats Python code.make lintruns pylint and mypy to catch errors.make testruns unit tests.