How to use Python metaclasses

This time I will show you how to use Python's metaclasses, and what are the precautions for using Python metaclasses. The following is a practical case, let's take a look.

My task today is to thoroughly understand what a metaclass is, let’s take a look.

To understand metaclasses, let’s start with objects.

Object (Object)

Python Everything is an object, you must have heard this sentence (now You have heard of it), a number is an object, a string is an object, a list is an object, and a dictionary is an object, for example:

>>> i = 10
>>> s = "abc"
>>> nums = [1,2,3]
>>> dicts = {"name":"zhang"}

The right side of the equal sign is an object, the left side is the name given to these objects. Any object has three key attributes: identity, value, and type.

Identification

Identification is the same as a person’s ID card ID. Each object has a unique ID identification throughout the life cycle will not change. You can think of the identifier as the address of this object in computer memory. The ID of an object can be viewed through the function id().

>>> id(i)
40592592
>>> id(s)
44980584

Object value

The second attribute of the object is the value. The value is easy to understand. For example, the value of i is 10, and the value of s is The value is abc, and the value of nums is 1,2,3.

Type

Another very important attribute of an object is its type. Any object has its own type, and the object is composed of its It is constructed from a type. For example, the type of i above is int type, and this object is constructed from int. The s type is a string type, the nums type is a list type, and the dicts type is a dictionary type. They are all constructed from the corresponding types.

You can check the type of object through type().

>>> type(i)
<class &#39;int&#39;>
>>> type(s)
<class &#39;str&#39;>
>>> type(nums)
<class &#39;list&#39;>
>>> type(dicts)
<class &#39;dict&#39;>

The type of the object is also the same as the ID identifier, and it will not change after it is determined.

Classes and (instance) objects

In addition to the integer types and string types that have been defined by the system, In addition to types such as lists, we can also create our own types and define them with the keyword class. For example:

>>> class Person:
  # 这里的 self 指某个实例对象自己
...  def init(self, name):
   # name 是实例的属性
...   self.name = name
  # live 是类的属性
  live = True

The Person here is a custom class. The class is an abstract template, which does not refer to Zhang San or Li Si. Now we can construct (instantiate) a specific one by calling this class. A real, named object comes out, and this object is called an instance object.

>>> p1 = Person("zhangsan")
>>> p1.name
'zhangsan'
>>>
>>> p2 = Person("lisi")
>>> p2.name
'lisi'

Here p1 and p2 are the (instance) objects after instantiation. The types of these two objects are Person class. The relationship between the class and the (instance) object is like a vehicle mold and a quilt. The relationship is the same as the real car being built.

>>> p1
<main.Person object at 0x0195AA30>
>>> type(p1)
<class &#39;main.Person&#39;> # 这里的main是模块名称

Classes are also objects (also called class objects)

We just said that everything is an object, and instances (real cars) are objects Of course, a class (mold) is also an object, because it is also a real thing.

When the Python interpreter executes the keyword class instruction, it will create a file named "Person" internally. " class, this class is also an object, we call it a class object (note the difference between instance objects), it also has an ID, a type, and a value. For example:

>>> id(Person)
26564024
>>> type(Person)
<class &#39;type&#39;>
>>> Person
<class &#39;main.Person&#39;>

We noticed that the type of this Person class object is called "type", which means that the Person class is created by type. Now you have to remember that p1 and p2 are instance objects, and Person is a class object. Also, what the hell is this type?

Let’s review, the type of instance object p1 is the class object Person, the type of Person is type

>>> nums = [1,2,3]
>>> type(nums)
<class &#39;list&#39;>
>>> type(list)
<class &#39;type&#39;>

nums is list, the type of list is also type, dictionary class (dict) Type is also type, and the types of all classes are type, which means that all classes are created by type. This type is the metaclass. Tao gives birth to one, one gives birth to two, and three gives rise to all things. The metaclass is the creator in Python. (The metaclass itself is also an object)

现在我们都知道类（对象）可以使用 class 关键字创建，我们还知道类（对象）的类型是 type，既然知道了它的类型是　type，那么肯定可以通过 type（元类）来创建。

用元类创建类

前面讲到过，type 有一个作用是用于检查对象的类型，其实它还有另外一个作用就是作为元类动态地创建类（对象）。

>>> Person = type("Person", (), {"live":True})
>>> Person
<class &#39;main.Person&#39;>

Person 就是一个类，它等价于：

>>> class Person:
...  live = True
...
>>> Person
<class &#39;main.Person&#39;>

用元类 type 创建类的语法是：

type(类名,基类元组(可以为空), 属性字典)

相信看了本文案例你已经掌握了方法，更多精彩请关注php中文网其它相关文章！

推荐阅读：

Python实现求解最大公约数的方法

在python中numpy与array应该怎样转换

The above is the detailed content of How to use Python metaclasses. For more information, please follow other related articles on the PHP Chinese website!