qingkuangs 发表于 2015-12-3 10:22:36

Python闭包的高级应用-装饰器的实现

  我们先看一个闭包的例子:

from time import ctime
def before_call(f):
def wrapped(*args, **kargs):
print 'before calling, now is %s' % ctime()
return f(*args, **kargs)
return wrapped
def test(name):
print 'hello, %s' % (name)
if __name__ == '__main__':
before_call(test)("lucky")
  我们先看运行结果:



~/Documents/py python 2.py
before calling, now is Sat Dec 27 21:30:18 2014
hello, lucky
  上面的代码使用了闭包,因为子函数wrapped将父函数的内部变量f与之绑定。
  这样,wrapped这个闭包函数,实际上先打印时间,然后调用f,所以正如结果打印的一般,before_call起到的是一种装饰的作用。
  
  这里我扩展它的功能,增加一个调用函数后,打印时间:



from time import ctime
def before_call(f):
def wrapped(*args, **kargs):
print 'before calling, now is %s' % ctime()
return f(*args, **kargs)
return wrapped
def after_call(f):
def wrapped(*args, **kargs):
try:
return f(*args, **kargs)
finally:
print 'after calling, now is %s' % ctime()
return wrapped
def test(name):
print 'hello, %s' % (name)
if __name__ == '__main__':
before_call(test)("lucky")
after_call(test)("peter")
before_call(after_call(test))("john")
after_call(before_call(test))('marry')
  运行结果为:



~/Documents/py python 2.py
before calling, now is Sat Dec 27 21:37:24 2014
hello, lucky
hello, peter
after calling, now is Sat Dec 27 21:37:24 2014
before calling, now is Sat Dec 27 21:37:24 2014
hello, john
after calling, now is Sat Dec 27 21:37:24 2014
before calling, now is Sat Dec 27 21:37:24 2014
hello, marry
after calling, now is Sat Dec 27 21:37:24 2014
  运行结果是正确的。注意最后两个,顺序交换了,对结果无影响。
  
  下面我们再包装一层:



def after_call():
def after(f):
def wrapped(*args, **kargs):
try:
return f(*args, **kargs)
finally:
print 'after calling, now is %s' % ctime()
return wrapped
return after

def before_call():
def before(f):
def wrapped(*args, **kargs):
print 'before calling, now is %s' % ctime()
return f(*args, **kargs)
return wrapped
return before
  那么如何使用呢?这里就是python装饰器的语法,
  如果我们这样使用:



@before_call()
def test(name):
print 'hello, %s' % (name)
if __name__ == '__main__':
test("lucky")
  注意test函数前加了装饰的符号。
  还可以这样:



@after_call()
def test(name):
print 'hello, %s' % (name)
  甚至可以嵌套多层:



@before_call()
@after_call()
def test(name):
print 'hello, %s' % (name)
  
  这就是python中装饰器的原理,内部采用了闭包。
页: [1]
查看完整版本: Python闭包的高级应用-装饰器的实现