【从0开始Python开发实战】Django集成Celery

目录：

1. Django集成Celery

2. 声明异步任务

3. 封装工具task_util.py

4. 单元测试test_task_util.py

5. 创建异步任务

6. 常见问题和解决方法

Celery是一个灵活可靠的分布式系统，用于异步任务调度，主要有3部分组成：消息中间件broker，任务执行单元worker，执行结果存储task result store。Celery使用第三方消息中间件Redis，RabbitMQ等。

系统通常将一些耗时的操作任务提交给Celery去异步执行，典型架构示意图如下。本文详细介绍Django集成Celery配置方法和功能测试。

时序图如下：

示例代码：https://github.com/rickding/HelloPython/tree/master/hello_celery

├── __init__.py

├── settings.py

├── celery.py

├── tasks.py

├── util

│   └── task_util.py

├── tests

│   └── test_task_util.py

一，Django集成Celery

1. 新建Django项目，运行：django-admin startproject hello_celery

2. 进到目录hello_celery，增加应用：python manage.py startapp app

项目的目录文件结构如下：

3. 安装Celery和依赖包，pip install celery >= 4.2.1，如果不是新建项目，注意版本兼容问题。

celery == 4.2.1
flower == 0.9.2
redis == 3.2.0
eventlet == 0.24.1

4. 增加celery.py，配置信息：

from __future__ import absolute_import, unicode_literals
import os
from celery import Celery, platforms
from django.conf import settings

# set the default Django settings module for the 'celery' program.
os.environ.setdefault('DJANGO_SETTINGS_MODULE', 'hello_celery.settings')

app = Celery(
    'hello_celery',
    include=['hello_celery.tasks'],
    broker=settings.CELERY_BROKER,
    backend=settings.CELERY_BACKEND
)

# Using a string here means the worker doesn't have to serialize
# the configuration object to child processes.
# - namespace='CELERY' means all celery-related configuration keys
#   should have a `CELERY_` prefix.
app.config_from_object('django.conf:settings', namespace='CELERY')

app.conf.update(
    CELERY_ACKS_LATE=True,
    CELERY_ACCEPT_CONTENT=['pickle', 'json'],
    CELERYD_FORCE_EXECV=True,
    CELERYD_MAX_TASKS_PER_CHILD=500,
    BROKER_HEARTBEAT=0,
)

# Optional configuration, see the application user guide.
app.conf.update(
    CELERY_TASK_RESULT_EXPIRES=3600,  # celery任务执行结果的超时时间，即结果在backend里的保存时间，单位s
)

# Load task modules from all registered Django app configs.
app.autodiscover_tasks()

platforms.C_FORCE_ROOT = True

5. 打开settings.py，配置BROKER和BACKEND地址：

CELERY_BROKER = 'redis://127.0.0.1:6379/2'
CELERY_BACKEND = 'redis://127.0.0.1:6379/3'

6. 打开__init__.py，增加代码：

from __future__ import absolute_import, unicode_literals
from .celery import app as celery_app

__all__ = ['celery_app']

二，增加tasks.py，声明异步任务

from __future__ import absolute_import, unicode_literals
import logging
import json

from celery import shared_task
from hello_celery.util.task_util import dispatch_task

log = logging.getLogger(__name__)


@shared_task
def task(param_str):
    log.info('task starts: %s, %s' % (type(param_str), param_str))

    param_dict = None
    try:
        param_dict = json.loads(param_str)
    except Exception as e:
        log.warning('Exception when parse param: %s' % str(e))

    log.info('parsed param: {}, {}'.format(type(param_dict), param_dict))
    return 'finished'

正确配置后，运行命令：celery -A hello_celery worker -l info -P eventlet，注意Win10环境中需要增加eventlet，Celery成功启动信息：

三，封装工具task_util.py

封装两个有用的工具函数，分别用于分发（创建）异步任务和获取任务信息：

import logging
import json

log = logging.getLogger(__name__)

def dispatch_task(task_func, param_dict):
    param_json = json.dumps(param_dict)

    try:
        return task_func.apply_async(
            [param_json],
            retry=True,
            retry_policy={
                'max_retries': 1,
                'interval_start': 0,
                'interval_step': 0.2,
                'interval_max': 0.2,
            },
        )
    except Exception as ex:
        log.info(ex)
        raise


def get_task_status(task_func, task_id):
    t = task_func.AsyncResult(task_id)
    status = t.state
    progress = 0

    if status == u'SUCCESS':
        progress = 100
    elif status == u'FAILURE':
        progress = 0
    elif status == 'PROGRESS':
        progress = t.info['progress']

    return {'status': status, 'progress': progress}

四，单元测试test_task_util.py

增加测试函数，创建一个任务任务并获取信息：

import logging

from django.test import TestCase

from hello_celery.tasks import task
from hello_celery.util.task_util import dispatch_task, get_task_status

log = logging.getLogger(__name__)


class TasksTest(TestCase):
    def test_get_task_status(self):
        t = dispatch_task(task, {'msg': 'test_task'})
        self.assertIsNotNone(t)

        ret = get_task_status(task, t.id)
        log.info('task status: %s,%s, %s' % (ret, t.id, str(task)))
        self.assertIsNotNone(ret.get('status'))

运行python manage.py test，同时Celery将执行测试函数创建的任务：

五，创建异步任务

1. 在views.py中增加请求处理函数，创建一个异步执行的任务：

from django.http import JsonResponse
from hello_celery.tasks import do_task


def chk_job(req):
    param_dict = {
        'url': req.get_raw_uri(),
        'path': req.get_full_path(),
    }
    job = do_task(param_dict)
    return JsonResponse({'code': 0, 'msg': 'success', 'job': job.task_id})

2. 在urls.py中配置路由

from django.urls import path
from app.views import chk_job

urlpatterns = [
    path('', chk_job, name='chk'),
]

3. 运行命令启动服务：python manage.py runserver 0.0.0.0:8001

4. REST接口创建异步任务示例

六，常见问题和解决方法

1. 启动Celery: celery -A hello_celery worker -l info，运行出错：

Unrecoverable error: VersionMismatch('Redis transport requires redis-py versions 3.2.0 or later. You have 2.10.6',)

解决：指定Redis使用3.2.0或更高pip install redis>=3.2.0