设计模式在美团外卖营销业务中的实践

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
//抽象的产品
public abstract class Product {
    public abstract void method();
}
//定义一个具体的产品 (可以定义多个具体的产品)
class ProductA extends Product {
    @Override
    public void method() {}  //具体的执行逻辑
}
//抽象的工厂
abstract class Factory<T> {
    abstract Product createProduct(Class<T> c);
}
//具体的工厂可以生产出相应的产品
class FactoryA extends Factory{
    @Override
    Product createProduct(Class c) {
        Product product = (Product) Class.forName(c.getName()).newInstance();
        return product;
    }
}

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
//定义一个策略接口
public interface Strategy {
    void strategyImplementation();
}

//具体的策略实现(可以定义多个具体的策略实现)
public class StrategyA implements Strategy{
    @Override
    public void strategyImplementation() {
        System.out.println("正在执行策略A");
    }
}

//封装策略，屏蔽高层模块对策略、算法的直接访问，屏蔽可能存在的策略变化
public class Context {
    private Strategy strategy = null;

    public Context(Strategy strategy) {
        this.strategy = strategy;
    }

    public void doStrategy() {
        strategy.strategyImplementation();
    }
}

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
//抽象策略
public abstract class RewardStrategy {
    public abstract void reward(long userId);

    public void insertRewardAndSettlement(long userId, int reward) {} ; //更新用户信息以及结算
}
//新用户返奖具体策略A
public class newUserRewardStrategyA extends RewardStrategy {
    @Override
    public void reward(long userId) {}  //具体的计算逻辑，...
}

//老用户返奖具体策略A
public class OldUserRewardStrategyA extends RewardStrategy {
    @Override
    public void reward(long userId) {}  //具体的计算逻辑，...
}

//抽象工厂
public abstract class StrategyFactory<T> {
    abstract RewardStrategy createStrategy(Class<T> c);
}

//具体工厂创建具体的策略
public class FactorRewardStrategyFactory extends StrategyFactory {
    @Override
    RewardStrategy createStrategy(Class c) {
        RewardStrategy product = null;
        try {
            product = (RewardStrategy) Class.forName(c.getName()).newInstance();
        } catch (Exception e) {}
        return product;
    }
}

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
public class RewardContext {
    private RewardStrategy strategy;

    public RewardContext(RewardStrategy strategy) {
        this.strategy = strategy;
    }

    public void doStrategy(long userId) { 
        int rewardMoney = strategy.reward(userId);
        insertRewardAndSettlement(long userId, int reward) {
          insertReward(userId, rewardMoney);
          settlement(userId);
       }  
    }
}

1
2
3
4
5
6
7
8
9
10
11
12
13
public class InviteRewardImpl {
    //返奖主流程
    public void sendReward(long userId) {
        FactorRewardStrategyFactory strategyFactory = new FactorRewardStrategyFactory();  //创建工厂
        Invitee invitee = getInviteeByUserId(userId);  //根据用户id查询用户信息
        if (invitee.userType == UserTypeEnum.NEW_USER) {  //新用户返奖策略
            NewUserBasicReward newUserBasicReward = (NewUserBasicReward) strategyFactory.createStrategy(NewUserBasicReward.class);
            RewardContext rewardContext = new RewardContext(newUserBasicReward);
            rewardContext.doStrategy(userId); //执行返奖策略
        }if(invitee.userType == UserTypeEnum.OLD_USER){}  //老用户返奖策略，... 
    }
}

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
//定义一个抽象的状态类
public abstract class State {
    Context context;
    public void setContext(Context context) {
        this.context = context;
    }
    public abstract void handle1();
    public abstract void handle2();
}
//定义状态A
public class ConcreteStateA extends State {
    @Override
    public void handle1() {}  //本状态下必须要处理的事情

    @Override
    public void handle2() {
        super.context.setCurrentState(Context.contreteStateB);  //切换到状态B        
        super.context.handle2();  //执行状态B的任务
    }
}
//定义状态B
public class ConcreteStateB extends State {
    @Override
    public void handle2() {}  //本状态下必须要处理的事情，...

    @Override
    public void handle1() {
        super.context.setCurrentState(Context.contreteStateA);  //切换到状态A
        super.context.handle1();  //执行状态A的任务
    }
}
//定义一个上下文管理环境
public class Context {
    public final static ConcreteStateA contreteStateA = new ConcreteStateA();
    public final static ConcreteStateB contreteStateB = new ConcreteStateB();

    private State CurrentState;
    public State getCurrentState() {return CurrentState;}

    public void setCurrentState(State currentState) {
        this.CurrentState = currentState;
        this.CurrentState.setContext(this);
    }

    public void handle1() {this.CurrentState.handle1();}
    public void handle2() {this.CurrentState.handle2();}
}
//定义client执行
public class client {
    public static void main(String[] args) {
        Context context = new Context();
        context.setCurrentState(new ContreteStateA());
        context.handle1();
        context.handle2();
    }
}

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
//返奖状态执行的上下文
public class RewardStateContext {

    private RewardState rewardState;

    public void setRewardState(RewardState currentState) {this.rewardState = currentState;}
    public RewardState getRewardState() {return rewardState;}
    public void echo(RewardStateContext context, Request request) {
        rewardState.doReward(context, request);
    }
}

public abstract class RewardState {
    abstract void doReward(RewardStateContext context, Request request);
}

//待校验状态
public class OrderCheckState extends RewardState {
    @Override
    public void doReward(RewardStateContext context, Request request) {
        orderCheck(context, request);  //对进来的订单进行校验，判断是否用券，是否满足优惠条件等等
    }
}

//待补偿状态
public class CompensateRewardState extends RewardState {
    @Override
    public void doReward(RewardStateContext context, Request request) {
        compensateReward(context, request);  //返奖失败，需要对用户进行返奖补偿
    }
}

//预返奖状态，待返奖状态，成功状态，失败状态(此处逻辑省略)
//..

public class InviteRewardServiceImpl {
    public boolean sendRewardForInvtee(long userId, long orderId) {
        Request request = new Request(userId, orderId);
        RewardStateContext rewardContext = new RewardStateContext();
        rewardContext.setRewardState(new OrderCheckState());
        rewardContext.echo(rewardContext, request);  //开始返奖，订单校验
        //此处的if-else逻辑只是为了表达状态的转换过程，并非实际的业务逻辑
        if (rewardContext.isResultFlag()) {  //如果订单校验成功，进入预返奖状态
            rewardContext.setRewardState(new BeforeRewardCheckState());
            rewardContext.echo(rewardContext, request);
        } else {//如果订单校验失败，进入返奖失败流程，...
            rewardContext.setRewardState(new RewardFailedState());
            rewardContext.echo(rewardContext, request);
            return false;
        }
        if (rewardContext.isResultFlag()) {//预返奖检查成功，进入待返奖流程，...
            rewardContext.setRewardState(new SendRewardState());
            rewardContext.echo(rewardContext, request);
        } else {  //如果预返奖检查失败，进入返奖失败流程，...
            rewardContext.setRewardState(new RewardFailedState());
            rewardContext.echo(rewardContext, request);
            return false;
        }
        if (rewardContext.isResultFlag()) {  //返奖成功，进入返奖结束流程，...
            rewardContext.setRewardState(new RewardSuccessState());
            rewardContext.echo(rewardContext, request);
        } else {  //返奖失败，进入返奖补偿阶段，...
            rewardContext.setRewardState(new CompensateRewardState());
            rewardContext.echo(rewardContext, request);
        }
        if (rewardContext.isResultFlag()) {  //补偿成功，进入返奖完成阶段，...
            rewardContext.setRewardState(new RewardSuccessState());
            rewardContext.echo(rewardContext, request);
        } else {  //补偿失败，仍然停留在当前态，直至补偿成功（或多次补偿失败后人工介入处理）
            rewardContext.setRewardState(new CompensateRewardState());
            rewardContext.echo(rewardContext, request);
        }
        return true;
    }
}

1
 

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
//定义一个抽象的handle
public abstract class Handler {
    private Handler nextHandler;  //指向下一个处理者
    private int level;  //处理者能够处理的级别

    public Handler(int level) {
        this.level = level;
    }

    public void setNextHandler(Handler handler) {
        this.nextHandler = handler;
    }

    // 处理请求传递，注意final，子类不可重写
    public final void handleMessage(Request request) {
        if (level == request.getRequstLevel()) {
            this.echo(request);
        } else {
            if (this.nextHandler != null) {
                this.nextHandler.handleMessage(request);
            } else {
                System.out.println("已经到最尽头了");
            }
        }
    }
    // 抽象方法，子类实现
    public abstract void echo(Request request);
}

// 定义一个具体的handleA
public class HandleRuleA extends Handler {
    public HandleRuleA(int level) {
        super(level);
    }
    @Override
    public void echo(Request request) {
        System.out.println("我是处理者1,我正在处理A规则");
    }
}

//定义一个具体的handleB
public class HandleRuleB extends Handler {}  //...

//客户端实现
class Client {
    public static void main(String[] args) {
        HandleRuleA handleRuleA = new HandleRuleA(1);
        HandleRuleB handleRuleB = new HandleRuleB(2);
        handleRuleA.setNextHandler(handleRuleB);  //这是重点，将handleA和handleB串起来
        handleRuleA.echo(new Request());
    }
}

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
//定义一个抽象的规则
public abstract class BasicRule<CORE_ITEM, T extends RuleContext<CORE_ITEM>>{
    //有两个方法，evaluate用于判断是否经过规则执行，execute用于执行具体的规则内容。
    public abstract boolean evaluate(T context);
    public abstract void execute(T context) {
}

//定义所有的规则具体实现
//规则1：判断服务可用性
public class ServiceAvailableRule extends BasicRule<UserPortrait, UserPortraitRuleContext> {
    @Override
    public boolean evaluate(UserPortraitRuleContext context) {
        TakeawayUserPortraitBasicInfo basicInfo = context.getBasicInfo();
        if (basicInfo.isServiceFail()) {
              return false;
        }
        return true;
    }

    @Override
    public void execute(UserPortraitRuleContext context) {}

}
//规则2：判断当前用户属性是否符合当前资源位投放的用户属性要求
public class UserGroupRule extends BasicRule<UserPortrait, UserPortraitRuleContext> {
    @Override
    public boolean evaluate(UserPortraitRuleContext context) {}

    @Override
    public void execute(UserPortraitRuleContext context) {
        UserPortrait userPortraitPO = context.getData();
        if(userPortraitPO.getUserGroup() == context.getBasicInfo().getUserGroup().code) {
          context.setValid(true);
        } else {
          context.setValid(false);
        }
    }
}

//规则3：判断当前用户是否在投放城市，具体逻辑省略
public class CityInfoRule extends BasicRule<UserPortrait, UserPortraitRuleContext> {}
//规则4：根据用户的活跃度进行资源过滤，具体逻辑省略
public class UserPortraitRule extends BasicRule<UserPortrait, UserPortraitRuleContext> {} 

//我们通过spring将这些规则串起来组成一个一个请求链
    <bean name="serviceAvailableRule" class="com.dianping.takeaway.ServiceAvailableRule"/>
    <bean name="userGroupValidRule" class="com.dianping.takeaway.UserGroupRule"/>
    <bean name="cityInfoValidRule" class="com.dianping.takeaway.CityInfoRule"/>
    <bean name="userPortraitRule" class="com.dianping.takeaway.UserPortraitRule"/>

    <util:list id="userPortraitRuleChain" value-type="com.dianping.takeaway.Rule">
        <ref bean="serviceAvailableRule"/>
        <ref bean="userGroupValidRule"/>
        <ref bean="cityInfoValidRule"/>
        <ref bean="userPortraitRule"/>
    </util:list>

//规则执行
public class DefaultRuleEngine{
    @Autowired
    List<BasicRule> userPortraitRuleChain;

    public void invokeAll(RuleContext ruleContext) {
        for(Rule rule : userPortraitRuleChain) {
            rule.evaluate(ruleContext)
        }
    }
}