Using @property
When binding properties, if we directly expose the properties, although it is very simple to write, but there is no way to check the parameters, resulting in the result can be changed casually:
s == 9999
This is obviously illogical. To limit the scope of the score, you can set the score by one method, and set_score()
then get the score by one, get_score()
so that in the set_score()
method, you can check the parameters:
classStudent (object):defGet_score (self):returnSelf._scoredefSet_score (self, value):if notisinstance (value, int):RaiseValueError ('score must is an integer!') ifValue < 0orValue > 100: RaiseValueError ('score must between 0 ~ 100!') Self._score= value
Now, to operate on any of the student instances, you cannot set score as you wish:
>>> s = Student ()# ok! >>> s.get_score ()60>>> S.set_score (9999) Traceback (most recent call last): ~ 100!
However, the above method of invocation is slightly more complex, and it is not straightforward to use attributes directly.
Is there any way to access the variables of a class in such an easy manner as to check parameters and use similar properties? For the perfect Python programmer, this must be done!
Remember the adorner (decorator) can add functionality to a function dynamically? For class methods, adorners work as well. Python's built-in @property
decorator is responsible for turning a method into a property invocation :
classStudent (object): @propertydefscore (self):returnself._score @score. Setterdefscore (self, value):if notisinstance (value, int):RaiseValueError ('score must is an integer!') ifValue < 0orValue > 100: RaiseValueError ('score must between 0 ~ 100!') Self._score= value
@property
Implementation is more complex, we first examine how to use. To turn a getter method into a property, just add it, and @property
at this point, it @property
creates another adorner @score.setter
, which is responsible for turning a setter method into a property assignment, so we have a controllable property operation:
>>> s = Student ()# OK, actual conversion to S.set_score# OK, actual conversion to S.get_score () 60>>> s.score = 9999Traceback (most recent call last): ~ 100!
Notice that this magical @property
, when we operate on instance properties, we know that this property is probably not directly exposed, but is implemented through getter and setter methods.
You can also define read-only properties, define getter methods only, and do not define setter methods as a read-only property :
class Student (object): @property def Birth (self): return self._ Birth @birth. Setter def Birth (self, value): = value @property def Age (self): return 2015-self._birth
The above birth
is a read- write property , which age
is a readonly property because age
it can be calculated based on the birth
current time.
Summary
@property
Widely used in the definition of a class, you can let the caller write short code, while ensuring that the parameters are necessary to check, so that the program runs to reduce the likelihood of errors.
Python Multiple Inheritance (mixin mode)
Languages such as C or C + + support multiple inheritance, and a subclass can have more than one parent, a design that is often criticized. Because inheritance should be a "is-a" relationship. Cars, for example, inherit the vehicle class, because the sedan is a ("is-a") vehicle. An item cannot be many different things, so there should be no multiple inheritance. But is there a case where a class does need to inherit multiple classes?
The answer is yes, we still take the transportation to give example, the civil aircraft is a kind of transportation, for local tyrants The helicopter is also a kind of transportation. For both modes of transport, they all have a function of flying, but the sedan does not. Therefore, we can not write the flight function in the vehicle of the parent class. But if both civil aircraft and helicopters write their own flight methods, they violate the principle of reusing the code as much as possible (if there are more flying tools later, there will be a lot of duplicate code). What to do, then we have to let the two planes inherit the vehicle and the aircraft two of the parent class, so there are multiple inheritance. At this time also violated the inheritance must be "is-a" relationship. How can this puzzle be broken?
Different languages give a different approach, let's look at Java first. Java provides interface interface functionality to implement multiple inheritance :
class Vehicle {} public interface flyable { class Flyableimpl implements flyable {public void FL Y () { System.out.println ("I am Flying"); class airplane extends Vehicle implements flyable { private flyable; Public airplane () { = new Flyableimpl (); } public void Fly () { flyable.fly (); }}
Now our aircraft has both the vehicle and the vehicle properties, and we do not need to rewrite the flight methods in the aircraft, and we do not undermine the principle of single inheritance. Aircraft is a means of transport, the ability to fly is the property of the aircraft, through the inheritance interface to obtain.
Back to the topic, the Python language can have no interface functionality, but it can inherit multiple. Is Python supposed to be implemented with multiple inheritance? Yes, it's not. That is, because syntactically, it is true through multiple inheritance. Say no, because its inheritance still abide by "is-a" relationship, from the meaning of still follow the principle of single inheritance. How does that make sense? Let's look at an example.
class Vehicle (object): Pass class Planemixin (object): def Fly (self): Print ' I am Flying ' class airplane (Vehicle, planemixin): Pass
As you can see, the airplane class above implements multiple inheritance, but the second class it inherits is named Planemixin, not plane, which does not affect functionality, but tells the person who later reads the code, which is a mixin class. So from the meaning of understanding, airplane is just a vehicle, not a plane. this mixin, which means mixed (mix-in), tells others that this class is added as a function to a subclass, not as a parent class, and it functions as an interface in Java.
Use the Mixin class to implement multiple inheritance with great care
- First it must represent a function, not an item, as Runnable,callable in Java.
- Second it must be a single responsibility, if there are multiple functions, then write multiple Mixin class
- Then, it does not depend on the implementation of the subclass
- Finally, subclasses can still work even if they do not inherit the Mixin class, which is missing a function. (such as the aircraft can still carry passengers, just can't fly ^_^)
In addition, Reactjs also has a mixin function, and the syntax is simple:
var planemixin = function () { return { fly:function () { console.log (' I am Flying'); } = React.createclass ({ mixins: [Planemixin ()], render:function () { return ' ' ; }});
Python Learning Day13 Python object-oriented Learning 2: @property, multiple inheritance