Python多继承与包装类相关技术疑问咨询

第一种继承模式与疑问 #

先看你给出的第一种模式代码：

from api.client import Client
from api.wrapper import Wrapper

class TestClient(Client):
    def __init__(self, wrapper):
        Client.__init__(self, wrapper)
 
class TestWrapper(Wrapper):
    def __init__(self):
        Wrapper.__init__(self)

    def updateMktDepth(self, reqId: TickerId, position: int, operation: int,
                side: int, price: float, size: int):
            super().updateMktDepth(reqId, position, operation, side, price, size)
            print("UpdateMarketDepth", "ReqId:", reqId, "Position:", position, "Operation:",
                  operation, "Side:", side, "Price:", price, "Size:", size)
 
    def updateMktDepthL2(self, reqId: TickerId, position: int, marketMaker: str,
            operation: int, side: int, price: float, size: int, isSmartDepth: bool):
        super().updateMktDepthL2(reqId, position, marketMaker, operation, side,
            price, size, isSmartDepth)
        print("UpdateMarketDepthL2", "ReqId:", reqId, "Position:", position, "MarketMaker:", marketMaker, "Operation:",
            operation, "Side:", side, "Price:", price, "Size:", size, "isSmartDepth:", isSmartDepth)
 
def main():
    wrapper = TestWrapper()
    client = TestClient(wrapper)
    client.connect('127.0.0.1', 3847, 100)
    print('Connection successful')
    client.disconnect()

疑问解答：实例化TestClient传wrapper参数是否和*args/**kwargs相关？ #

完全无关，你可能看错了代码——TestClient的__init__方法是显式声明了wrapper参数的：def __init__(self, wrapper):。其中self是Python类方法的默认第一个参数（代表当前实例），第二个参数就是wrapper，所以实例化TestClient(wrapper)是完全匹配参数定义的正常传参操作。

第二种多继承模式与疑问 #

再看第二种模式的代码（包含Client类定义）：

from api.client import Client
from api.wrapper import Wrapper
    
class TestApp(Client, Wrapper):
    def __init__(self):
        Client.__init__(self, self)

    def updateMktDepth(self, reqId: TickerId, position: int, operation: int, side: int, price: float, size: int):
        super.updateMktDepth(reqId, position, operation, side, price, size)
        print("UpdateMarketDepth", "ReqId:", reqId, "Position:", position, "Operation:", operation, "Side:", side, "Price:", price, "Size:", size)

    def updateMktDepthL2(self, reqId: TickerId, position: int, marketMaker: str, operation: int, side: int, price: float, size: int, isSmartDepth: bool):
        super.updateMktDepthL2(reqId, position, operation, side, price, size)
        print("UpdateMarketDepthL2", "ReqId:", reqId, "Position:", position, "MarketMaker:", marketMaker, "Operation:", 
              operation, "Side:", side, "Price:", price, "Size:", size, "isSmartDepth:", isSmartDepth)

class Client(object):
    (DISCONNECTED, CONNECTING, CONNECTED, REDIRECT) = range(4)

    def __init__(self, wrapper):
        self.msg_queue = queue.Queue()
        self.wrapper = wrapper
        self.decoder = None
        self.reset()

疑问1：多继承下super如何确定方法位置？ #

Python多继承靠**方法解析顺序（MRO）**来决定查找方法的顺序，你可以通过TestApp.__mro__或者TestApp.mro()打印出具体的顺序列表。当调用super().xxx()时，会严格按照MRO列表的顺序，从当前类的下一个类开始依次查找，找到第一个包含目标方法的类就调用它。

比如TestApp继承了Client和Wrapper，假设MRO是[TestApp, Client, Wrapper, object]，如果updateMktDepth只在Wrapper中存在，那super就会跳过Client，直接找到Wrapper的该方法执行。

疑问2：Client.__init__(self, self)两次传self的原理？ #

首先，Client.__init__的第一个self是Python构造方法的常规要求——调用父类构造时必须传入当前子类的实例（也就是TestApp的实例），这是为了让父类的初始化逻辑作用在子类实例上。

而第二个self，是因为Client的构造需要一个wrapper对象（用来处理回调，比如updateMktDepth这类方法）。由于TestApp同时继承了Wrapper，它的实例本身就具备Wrapper的所有能力（包括你重写的那些回调方法），所以直接把当前实例传给Client的wrapper参数，就能让Client在触发回调时，直接调用TestApp自身实现的那些方法，不用再单独创建一个Wrapper实例，相当于把Client和Wrapper的功能合并到了同一个实例里。

内容的提问来源于stack exchange，提问作者Chris