Parsely

Parsely is a framework for enqueuing messages via HTTP request for celery.

CHANGELOG

Example

tasks.py:

fromtypingimportDictimportparselyapp=parsely.Parsely('example', 'redis://localhost:6379/0')
deftwo_times(text: str) ->Dict[str, str]:
return {
'text': text*2
}
@app.task(two_times)defecho(text: str) ->None:
print(text)
worker=p.celeryproducer=p.make_producer()

two_times works as pre processor. It works before enqueing. It means it can return BadRequest to your client. Parsely validate message with typehint. Also you can have extra validation and any other process here.

You can run tasks.py as Celery worker: celery -A tasks worker --loglevel=info

Also it runs as WSGI application. This is an example run it with uWSGI uwsgi --http :8080 --wsgi-file tasks.py --callable producer --enable-threads --thunder-lock --master

Send Test Message!

curl -X POST -d '{"message":{"text": "Hello"}}' http://localhost:8080/example/echo

1. producer receives your request
2. producer validates your message having text and the type is str or not. text`(str) is from typehint of two_times
3. producer validates two_times returned value having text and the type is str or not. text`(str) is from typehint of echo
4. producer put message {"message":{"text":"HelloHello"}} into queue.
5. curl receives response.
6. Celery calls echo

Extra

Preprocessor

preprocessor is optional. if you don't need it, you can:

@app.task()defecho(text: str) ->None:
print(text)

Also you can give multiple preprocessor:

@app.task(two_times, two_times)defecho(text: str) ->None:
print(text)

logging

Parsely uses logger witch named "parsely", so you can set log level like:

importloggingparsely_logger=logging.getLogger('parsely')
parsely_logger.setLevel(logging.DEBUG)