Still writing a lot of if to judge? A rule executor kills all if judgments in the project

Business scenario

Recently, I received a small demand in the company , You need to expand the existing trial user application rules . Our scenario is roughly as follows :

if ( Whether overseas users ) {
 return false;
}

if ( Billing user ) {
  return false;
}

if ( Unpaid users  &&  No longer in service ) {
  return false
}

if ( Transfer to introduce the user  ||  Paying users  ||  Push users ) {
  return true;
}

According to the above conditions, we can come to the conclusion that ：

• Our main process is mainly based on and perhaps or The relationship between .

• If there is a mismatch , In fact, our follow-up process does not need to be implemented , It needs to have a short circuit function .

• For the current situation , If I change it on the original basis , Just pay a little attention to solving the problem that the demand is not very big , But the maintainability is very poor .

After weighing in the back , I decided to refactor this part .

Rule executor

For this demand , I first sorted out the general design of our rule actuator , Then I designed a V1 Version to share with you , If everyone has such a case You can share messages with me , The following part is mainly the design and implementation process code.

Design of rule executor

Abstract rules and implement rules

//  Business data 
@Data
public class RuleDto {
  private String address;
 private int age;
}

//  Rule abstraction 
public interface BaseRule {

    boolean execute(RuleDto dto);
}

//  Rule templates 
public abstract class AbstractRule implements BaseRule {

    protected <T> T convert(RuleDto dto) {
        return (T) dto;
    }

    @Override
    public boolean execute(RuleDto dto) {
        return executeRule(convert(dto));
    }
  
    protected <T> boolean executeRule(T t) {
        return true;
    }
}

//  Specific rules -  Example 1
public class AddressRule extends AbstractRule {

    @Override
    public boolean execute(RuleDto dto) {
        System.out.println("AddressRule invoke!");
        if (dto.getAddress().startsWith(MATCH_ADDRESS_START)) {
            return true;
        }
        return false;
    }
}

//  Specific rules -  Example 2
public class NationalityRule extends AbstractRule {

    @Override
    protected <T> T convert(RuleDto dto) {
        NationalityRuleDto nationalityRuleDto = new NationalityRuleDto();
        if (dto.getAddress().startsWith(MATCH_ADDRESS_START)) {
            nationalityRuleDto.setNationality(MATCH_NATIONALITY_START);
        }
        return (T) nationalityRuleDto;
    }


    @Override
    protected <T> boolean executeRule(T t) {
        System.out.println("NationalityRule invoke!");
        NationalityRuleDto nationalityRuleDto = (NationalityRuleDto) t;
        if (nationalityRuleDto.getNationality().startsWith(MATCH_NATIONALITY_START)) {
            return true;
        }
        return false;
    }
}

//  Constant definition 
public class RuleConstant {
    public static final String MATCH_ADDRESS_START= " Beijing ";
    public static final String MATCH_NATIONALITY_START= " China ";
}

Actuator build

public class RuleService {

    private Map<Integer, List<BaseRule>> hashMap = new HashMap<>();
    private static final int AND = 1;
    private static final int OR = 0;

    public static RuleService create() {
        return new RuleService();
    }


    public RuleService and(List<BaseRule> ruleList) {
        hashMap.put(AND, ruleList);
        return this;
    }

    public RuleService or(List<BaseRule> ruleList) {
        hashMap.put(OR, ruleList);
        return this;
    }

    public boolean execute(RuleDto dto) {
        for (Map.Entry<Integer, List<BaseRule>> item : hashMap.entrySet()) {
            List<BaseRule> ruleList = item.getValue();
            switch (item.getKey()) {
                case AND:
                    //  If it is  and  Relationship , Synchronous execution 
                    System.out.println("execute key = " + 1);
                    if (!and(dto, ruleList)) {
                        return false;
                    }
                    break;
                case OR:
                    //  If it is  or  Relationship , Parallel execution 
                    System.out.println("execute key = " + 0);
                    if (!or(dto, ruleList)) {
                        return false;
                    }
                    break;
                default:
                    break;
            }
        }
        return true;
    }

    private boolean and(RuleDto dto, List<BaseRule> ruleList) {
        for (BaseRule rule : ruleList) {
            boolean execute = rule.execute(dto);
            if (!execute) {
                // and  Relationship matching failed once , return  false
                return false;
            }
        }
        // and  All relationships match successfully , return  true
        return true;
    }

    private boolean or(RuleDto dto, List<BaseRule> ruleList) {
        for (BaseRule rule : ruleList) {
            boolean execute = rule.execute(dto);
            if (execute) {
                // or  If the relationship matches one, it returns  true
                return true;
            }
        }
        // or  If none of the relationships match, it returns  false
        return false;
    }
}

The call of the actuator

public class RuleServiceTest {

    @org.junit.Test
    public void execute() {
        // Rule executor 
        // advantage ： Relatively simple , Each rule can be independent , Put the rules , data , The actuator is disassembled , The caller is regular 
        // shortcoming ： Data depends on common transport objects  dto

        //1.  Define the rules   init rule
        AgeRule ageRule = new AgeRule();
        NameRule nameRule = new NameRule();
        NationalityRule nationalityRule = new NationalityRule();
        AddressRule addressRule = new AddressRule();
        SubjectRule subjectRule = new SubjectRule();

        //2.  Construct the required data  create dto
        RuleDto dto = new RuleDto();
        dto.setAge(5);
        dto.setName(" Zhang San ");
        dto.setAddress(" Beijing ");
        dto.setSubject(" mathematics ");;

        //3.  Build and execute by chaining calls  rule execute
        boolean ruleResult = RuleService
                .create()
                .and(Arrays.asList(nationalityRule, nameRule, addressRule))
                .or(Arrays.asList(ageRule, subjectRule))
                .execute(dto);
        System.out.println("this student rule execute result :" + ruleResult);
    }
}

summary

Advantages and disadvantages of rule executors

advantage ：

• Relatively simple , Each rule can be independent , Put the rules , data , The actuator is disassembled , The caller is regular ;

• I am here Rule Template class convert Method to convert parameters so that , For specific rule The required scene data is provided .

shortcoming ：

• Up and down rule There's data dependence , If you modify the public transport object directly dto This design is not very reasonable , It is recommended to build the data in advance .