Python中关于类和继承以及多态的示例详解


本文摘自php中文网,作者黄舟,侵删。

本文通过实例给大家详细解释了Python 的类、继承和多态的定义和用法,非常实用,有需要的小伙伴可以参考下

类的定义

假如要定义一个类 Point,表示二维的坐标点:

1

2

3

4

# point.py

class Point:

  def init(self, x=0, y=0):

    self.x, self.y = x, y

最最基本的就是 init 方法,相当于 C++ / Java 的构造函数。带双下划线 的方法都是特殊方法,除了 init 还有很多,后面会有介绍。

参数 self 相当于 C++ 的 this,表示当前实例,所有方法都有这个参数,但是调用时并不需要指定。

1

2

3

4

5

6

>>> from point import *

>>> p = Point(10, 10) # init 被调用

>>> type(p)

<class 'point.Point'>

>>> p.x, p.y

(10, 10)

几乎所有的特殊方法(包括 init)都是隐式调用的(不直接调用)。

对一切皆对象的 Python 来说,类自己当然也是对象:

1

2

3

4

5

6

>>> type(Point)

<class 'type'>

>>> dir(Point)

['class', 'delattr', 'dict', ..., 'init', ...]

>>> Point.class

<class 'type'>

Point 是 type 的一个实例,这和 p 是 Point 的一个实例是一回事。

现添加方法 set:

1

2

3

4

class Point:

  ...

  def set(self, x, y):

    self.x, self.y = x, y

1

2

3

4

>>> p = Point(10, 10)

>>> p.set(0, 0)

>>> p.x, p.y

(0, 0)

p.set(...) 其实只是一个语法糖,你也可以写成 Point.set(p, ...),这样就能明显看出 p 就是 self 参数了:

1

2

3

>>> Point.set(p, 0, 0)

>>> p.x, p.y

(0, 0)

值得注意的是,self 并不是关键字,甚至可以用其它名字替代,比如 this:

1

2

3

4

class Point:

  ...

  def set(this, x, y):

    this.x, this.y = x, y

与 C++ 不同的是,“成员变量”必须要加 self. 前缀,否则就变成类的属性(相当于 C++ 静态成员),而不是对象的属性了。

访问控制

Python 没有 public / protected / private 这样的访问控制,如果你非要表示“私有”,习惯是加双下划线前缀。

1

2

3

4

5

6

7

8

9

class Point:

  def init(self, x=0, y=0):

    self.x, self.y = x, y

 

  def set(self, x, y):

    self.x, self.y = x, y

 

  def f(self):

    pass

x、y 和 f 就相当于私有了:

1

2

3

4

5

6

7

>>> p = Point(10, 10)

>>> p.x

...

AttributeError: 'Point' object has no attribute 'x'

>>> p.f()

...

AttributeError: 'Point' object has no attribute 'f'

_repr_

尝试打印 Point 实例:

1

2

3

>>> p = Point(10, 10)

>>> p

<point.Point object at 0x000000000272AA20>

通常,这并不是我们想要的输出,我们想要的是:

1

2

>>> p

Point(10, 10)

添加特殊方法 repr 即可实现:

1

2

3

class Point:

  def repr(self):

    return 'Point({}, {})'.format(self.x, self.y)

不难看出,交互模式在打印 p 时其实是调用了 repr(p):

>>> repr(p)
'Point(10, 10)'

_str_

如果没有提供 str,str() 缺省使用 repr() 的结果。
这两者都是对象的字符串形式的表示,但还是有点差别的。简单来说,repr() 的结果面向的是解释器,通常都是合法的 Python 代码,比如 Point(10, 10);而 str() 的结果面向用户,更简洁,比如 (10, 10)。

按照这个原则,我们为 Point 提供 str 的定义如下:

1

2

3

class Point:

  def str(self):

    return '({}, {})'.format(self.x, self.y)

_add_

两个坐标点相加是个很合理的需求。

1

2

3

4

5

6

>>> p1 = Point(10, 10)

>>> p2 = Point(10, 10)

>>> p3 = p1 + p2

Traceback (most recent call last):

 File "<stdin>", line 1, in <module>

TypeError: unsupported operand type(s) for +: 'Point' and 'Point'

添加特殊方法 add 即可做到:

1

2

3

class Point:

  def add(self, other):

    return Point(self.x + other.x, self.y + other.y)

1

2

3

>>> p3 = p1 + p2

>>> p3

Point(20, 20)

这就像 C++ 里的操作符重载一样。
Python 的内建类型,比如字符串、列表,都“重载”了 + 操作符。

特殊方法还有很多,这里就不逐一介绍了。

继承

举一个教科书中最常见的例子。Circle 和 Rectangle 继承自 Shape,不同的图形,面积(area)计算方式不同。

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

# shape.py

 

class Shape:

  def area(self):

    return 0.0

     

class Circle(Shape):

  def init(self, r=0.0):

    self.r = r

 

  def area(self):

    return math.pi * self.r * self.r

 

class Rectangle(Shape):

  def init(self, a, b):

    self.a, self.b = a, b

 

  def area(self):

    return self.a * self.b

用法比较直接:

1

2

3

4

5

6

7

>>> from shape import *

>>> circle = Circle(3.0)

>>> circle.area()

28.274333882308138

>>> rectangle = Rectangle(2.0, 3.0)

>>> rectangle.area()

6.0

如果 Circle 没有定义自己的 area:

1

2

class Circle(Shape):

  pass

那么它将继承父类 Shape 的 area:

1

2

>>> Shape.area is Circle.area

True

一旦 Circle 定义了自己的 area,从 Shape 继承而来的那个 area 就被重写(overwrite)了:

1

2

3

>>> from shape import *

>>> Shape.area is Circle.area

False

通过类的字典更能明显地看清这一点:

1

2

3

4

>>> Shape.dict['area']

<function Shape.area at 0x0000000001FDB9D8>

>>> Circle.dict['area']

<function Circle.area at 0x0000000001FDBB70>

所以,子类重写父类的方法,其实只是把相同的属性名绑定到了不同的函数对象。可见 Python 是没有覆写(override)的概念的。

同理,即使 Shape 没有定义 area 也是可以的,Shape 作为“接口”,并不能得到语法的保证。

甚至可以动态的添加方法:

1

2

3

4

5

6

7

class Circle(Shape):

  ...

  # def area(self):

    # return math.pi * self.r * self.r

 

# 为 Circle 添加 area 方法。

Circle.area = lambda self: math.pi * self.r * self.r

动态语言一般都是这么灵活,Python 也不例外。

Python 官方教程「9. Classes」第一句就是:

相关阅读 >>

配置Python连接oracle,读取excel数据写入数据库

Python中一些常用模块的介绍

Python必学知识点总汇

Python编程时如何添加中文注释

Python中类的创建与使用详解

Python中如何合并两个字典教程

Python怎么输出分数形式

Python是什么意思?怎么读?

Python虚拟茶话会效果

Python中all()函数和any()函数的便捷用法

更多相关阅读请进入《Python》频道 >>




打赏

取消

感谢您的支持,我会继续努力的!

扫码支持
扫码打赏,您说多少就多少

打开支付宝扫一扫,即可进行扫码打赏哦

分享从这里开始,精彩与您同在

评论

管理员已关闭评论功能...