Python Flask后端异步处理（三）

　　前一篇博文我们已经将基础知识和环境配置进行了介绍：https://www.cnblogs.com/Cl0ud/p/13192925.html，本篇博文在实际应用场景中使用Celery，对Flask后端进行异步处理。

　　首先编写一个celerytask.py文件进行Celery的配置，同时耗时任务也写在该文件中

from celery import Celery
from init import app
from SZheConsole import SZheScan

app.config['CELERY_BROKER_URL'] = 'redis://localhost:6379/0'
app.config['CELERY_RESULT_BACKEND'] = 'redis://localhost:6379/0'

celery = Celery(app.name, broker=app.config['CELERY_BROKER_URL'])
celery.conf.update(app.config)


@celery.task
def SZheConsole(urls):
    try:
        for url in urls:
            print("="*20)
            print(url)
            SZheScan(url)
    except Exception as e:
        print("错误")
        print(e)
        pass
    print("allend!")

使用@celery.task装饰器修饰耗时函数SZheConsole，让Celery能够正确调用。

SZheScan函数是另外一个文件里面的函数，即对每一个URL进行单独的扫描，这部分过几天为了提高扫描速度会进行优化，这里将Celery用到项目里，暂时不改。

调用耗时任务SZheConsole的地方在index.py视图函数中，只选调用部分代码

@app.route('/console', methods=['GET', 'POST'])
@login_required
def console():
    bugbit, bugtype = core.GetBit()
    counts = core.GetCounts()
    ports = core.GetPort()
    services = core.GetServices()
    target = core.GetTargetCount()
    try:
        lastscantime = BaseInfo.query.order_by(BaseInfo.id.desc()).first().date
    except:
        lastscantime = "暂无扫描"
        pass
    if request.method == 'GET':
        return render_template('console.html', bugbit=bugbit, bugtype=bugtype, counts=counts, lastscantime=lastscantime,
                               ports=ports, services=services, target=target)
    else:
        urls = request.form.get('urls')
        urls = urls.split()
        print(urls)
        for url in urls:
            redispool.hincrby('targetscan', 'waitcount', 1)
        # executor.submit(SZheConsole, urls)
        SZheConsole.delay(urls)
        target = core.GetTargetCount()
        return render_template('console.html', bugbit=bugbit, bugtype=bugtype, counts=counts, lastscantime=lastscantime,
                               ports=ports, services=services, target=target)

可以看到原来的处理方式是多进程处理

executor.submit(SZheConsole, urls)

将原来的处理方式代码注释掉后，关键代码修改为

SZheConsole.delay(urls)

这样就可以将耗时任务丢给Celery进行处理，页面立即返回

return render_template('console.html', bugbit=bugbit, bugtype=bugtype, counts=counts, lastscantime=lastscantime,ports=ports, services=services, target=target)

接着启动redis和Celery服务，启动redis自不用说，Celery启动命令为：