使用 python tornado 框架异步调用 celery 队列任务的实现，很简单的小函数，但是很好用

tasks.py

from celery import Celery
import time 

celery = Celery('tasks', backend='redis://localhost', broker='amqp://')


@celery.task
def test(strs):
    return strs

torncelery.py

# -*- coding: utf-8 -*-
"""
    torncelery
    ~~~~~~~~~~
    assign celery task to Future and get task result async.
"""


from tornado.concurrent import TracebackFuture
from tornado.ioloop import IOLoop

def async(task, *args, **kwargs):
    future = TracebackFuture()
    callback = kwargs.pop("callback", None)
    if callback:
        IOLoop.instance().add_future(future,
                                     lambda future: callback(future.result()))
    result = task.delay(*args, **kwargs)
    IOLoop.instance().add_callback(_on_result, result, future)
    return future

def _on_result(result, future):
    # if result is not ready, add callback function to next loop,
    if result.ready():
        future.set_result(result.result)
    else:
        IOLoop.instance().add_callback(_on_result, result, future)

usage.py

import tornado.ioloop
import tornado.web
from tornado.gen import coroutine

from tasks import test
import torncelery


class MainHandler(tornado.web.RequestHandler):
    @coroutine
    def get(self):
        result = yield torncelery.async(test, "hello world")
        self.write("%s" % result )

application = tornado.web.Application([
    (r"/", MainHandler),
])

if __name__ == "__main__":
    application.listen(8888)
    tornado.ioloop.IOLoop.instance().start()