celery是实现一个简单，灵活可靠的分布式任务队列系统的好选择

tornado则不用过多介绍

在开发机上安装rabbitmq这里就不介绍了

首先是task文件的编写

task.py

#coding=utf-8
from celery import Celery
from celery.bin import worker as celery_worker
import celeryconfig

broker = 'amqp://'
backend = 'amqp'
app = Celery('celery_test', backend=backend, broker=broker)
app.config_from_object(celeryconfig)

@app.task
def mytask0(task_name):
  print "task0:%s" %task_name
  return task_name 

@app.task
def mytask1(task_name):
  print "task1:%s" %task_name
  return task_name 

def worker_start():
  worker = celery_worker.worker(app=app)
  worker.run(broker=broker, concurrency=4,
        traceback=False, loglevel='INFO')

if __name__ == "__main__":
  worker_start()

celeryconfig.py文件中包含对celery的配置

#coding=utf-8
from kombu import Queue
CELERY_DEFAULT_QUEUE = 'mytask0'
CELERY_QUEUES = (
  Queue('mytask0',  routing_key='task.mytask0'),
  Queue('mytask1',  routing_key='task.mytask1'),
)
CELERY_DEFAULT_EXCHANGE_TYPE = 'direct'
CELERY_DEFAULT_ROUTING_KEY = 'task.mytask0'
CELERY_TIMEZONE = 'Asia/Shanghai'
CELERY_ROUTES = {
  'task.mytask0': {
    'queue': 'mytask0',
    'routing_key': 'task.mytask0',
  },
  'task.mytask1': {
    'queue': 'mytask1',
    'routing_key': 'task.mytask1',
  },
}

执行python task.py将会启动worker

tornado调用celery将阻塞任务变为非阻塞

这会使用到tcelery模块，即tornado下的一个非阻塞的broker实现

app.py

#coding=utf-8
from tornado import web
import task

class TestHandler(tornado.web.RequestHandler):

  @web.asynchronous
  def get(self):
    task.mytask0.apply_async(
      args=['task0'],
         queue='mytask0',
         routing_key='task.mytask0',
         callback=self.on_success)
  def on_success(self, result):
    self.finish({'task':result.result})

start.py

用于实现tornado服务的启动

#coding=utf-8
import tornado
from tornado.options import define, options, parse_command_line
from tornado.log import enable_pretty_logging
import tcelery
from app import TestHandler
import tornado.httpserver

define("port", default=8000, help="run on the given port", type=int)
define("debug", default=False, help="run in debug mode")

urls = [(r"/api/task/test", TestHandler)]

def server_start():
  app = tornado.web.Application(urls, debug=options.debug)
  enable_pretty_logging()
  parse_command_line()
  server = tornado.httpserver.HTTPServer(app)
  server.bind(options.port)
  server.start(2)
  tcelery.setup_nonblocking_producer(limit=2)
  tornado.ioloop.IOLoop.current().start()

if __name__ == "__main__":
  server_start()

执行python start.py即可启动服务

以上这篇tornado+celery的简单使用详解就是小编分享给大家的全部内容了，希望能给大家一个参考，也希望大家多多支持python博客。