Flask并发之gevent模块（九）

并发

• 同时出发，不一定同时执行。
• 多进程 multiprocessing

异步并发

• 执行时间片，在效果上神似并发的效果，
  Threading
  Threading.Thread
  _thread 是threading模块的父类
  Twisted(扭曲)框架，tornado框架的核心框架

并行

• 同时执行，不强调同时出发

多线程：讨论的是异步问题

协程：单线程，在单个线程内，自由的在两个不同函数，执行到一部分的时候， 切换。协程解决了IO等待问题，在python当中协程最原始是由yeild生成器实现的

gevent 协程框架。解决多IO问题

yield生成器案例:

"""
协程回顾
"""

# 生成器
def hello():
   for i in (1,3,5):
       key = yield i
       print(key)
       print("hello world")

# h = hello()
# print("+++++++++++++++++++++++++++++++++++++++++++++++")
# print(next(h))
# print("+++++++++++++++++++++++++++++++++++++++++++++++")
# print(next(h))
# print("+++++++++++++++++++++++++++++++++++++++++++++++")
# print(next(h))
# print("+++++++++++++++++++++++++++++++++++++++++++++++")

# print("+++++++++++++++++++++++++++++++++++++++++++++++")
# print(next(h))
# print("+++++++++++++++++++++++++++++++++++++++++++++++")
# print(h.send(10))
# print("+++++++++++++++++++++++++++++++++++++++++++++++")
# print(h.send(20))


# 实际工作中，协程至少需要两个函数
def getContent():
   """
   获取内容的方法
   """
   while True:
       url = yield  "I have content"
       print("get content from url : %s"%url)

def getUrl(g):
   url_list = ["url1","url2","url3","url4","url5"]
   for i in url_list:
       print("+++++++++++++++++++++++++++++++++++++++++++++++")
       g.send(i)
       print("+++++++++++++++++++++++++++++++++++++++++++++++")

if __name__ == "__main__":
   g = getContent()
   print(next(g))
   getUrl(g)

安装gevent模块

pip install gevent

gevent官方案例

"""
gevent官方案例
"""
import gevent

def fun1():
    for i in range(5):
        print("I am fun 1 this is %s"%i)
        gevent.sleep(0)

def fun2():
    for i in range(5):
        print("I am fun 2 this is %s"%i)
        gevent.sleep(0)


# fun1()
# fun2()

# 交替执行
t1 = gevent.spawn(fun1)
t2 = gevent.spawn(fun2)
gevent.joinall([t1,t2])

协程锁案例：

import gevent
from gevent.lock import Semaphore

# 实例信号量
sem = Semaphore(1)

def fun1():
    for i in range(5):
        sem.acquire()
        print("I am fun 1 this is %s"%i)
        sem.release()
def fun2():
    for i in range(5):
        sem.acquire()
        print("I am fun 2 this is %s"%i)
        sem.release()

# fun1()
# fun2()

t1 = gevent.spawn(fun1)
t2 = gevent.spawn(fun2)
gevent.joinall([t1,t2])

flask基于gevent模块实现并发

Flask+gevent高效部署：当前优化适用于io访问频繁的项目，算法类型不适用。

如果Flask项目只是小规模的优化只需要将命令配置(app)当中的threaded改为True就可以了

manage.py

from app import create_app,db
from flask_script import Manager
from flask_migrate import Migrate # 用来同步数据库
from flask_migrate import MigrateCommand # 用来同步数据库的命令

from gevent import monkey

# 猴子补丁，将之前代码当中所有不契合协程的代码修改为契合
monkey.patch_all()

# 实例化app
app = create_app("running")

# 命令行封装app
manager = Manager(app)

# 绑定可以管理的数据库模型
migrate = Migrate(app,db)

# 加载数据库管理命令
manager.add_command("db",MigrateCommand)

@manager.command
def runserver_gevent():
    from gevent import pywsgi
    server = pywsgi.WSGIServer(("127.0.0.1",5000),app)
    server.serve_forever()

if __name__ == "__main__":
    manager.run()

以gevent方式运行项目：

python manage.py runserver_gevent