JarCastingbusiness-validation-common
Define base interface and common logic for API business validation
Summary
In microservice API, normally it has validation in two levels:
-
Openapi schema validation
Field level validation for request/response in the openapi specification.
For example: required fields, data type, data format...
-
Business validation
Certain business rule validation for which cannot be defined the detail in the openapi specification
For example, validate account number for different account type; validate driving license number issued from different country/place
For the openapi schema validation, schema validation can be applied directly. Some microservice framwork embedded it with the framework; for example, sprint-boot integrate with swagger codegen and bean validation, light-4j integration with request/response handler chain cross-cutting concerns.
And in the application, schema validation can be applied by applying library directly:
https://github.com/mservicetech/openapi-schema-validation
For business validation, some monolithic application use rule engineer for it. But rule engineer is too heavy for microservice API. And developers more like to build the business validation in pure java code.
Ideally, we would like to build the complex business validation to small and light validation validators as validation chain. The validator chain can be plugin to the service class to controller/handler which will not impact the business logic implmentation in the service/controller/handler classes.
Implementation detail:
For validator implementation, please refer the example in the /sample package
To invoke the validator class list stream
- Initial validators (get the list validators which implements BaseValidator interface )
For example, on Sprint-boot:
@Autowired
List<BaseValidator> validators;
- Invoke the validators
List<ValidationCode> codes = new ArrayList<>();
Stream<ValidationResult> validationResultStream = validators.stream().filter(validator->validator.support(FILTER_NAME)).flatMap(v->v.validate(null,drivingLicense));
validationResultStream.filter(v->v.isError()).forEach(v->codes.addAll(v.getValidationCodes()));
if (codes.size()>0) {
//TODO throw exception or return error codes
}
- Invoke the validators with priority order (from small number to large number) defined in the interface method
Stream<ValidationResult> validationResultStream = validators.stream().
filter(validator->validator.support(drivingLicense.getPlaceOfIssue())).sorted(Comparator.comparingInt(BaseValidator::priority)).
flatMap(v->v.validate(null,drivingLicense));
validationResultStream.filter(v->v.isError()).forEach(v->codes.addAll(v.getValidationCodes()));
rule annotation support
Provide rule annotation which user can define basic rule information there, for example:
@Rule(id="0002", description = "account balance validation", errorMessage = "Balance for account cannot less than 0.",
effectiveDate= "2020-02-20", endDate="2021-12-32")
public class AccountBalanceValidator implements AnnotationBaseValidator<String, Account> {
The validation class need implement AnnotationBaseValidator interface instead of BaseValidator; Library provide set of methods to use the annotation fields
For implementation detail, please refer to the sample in the test package
The implementation validation class need implement two interface methods:
- annotationSupport (base support and isValidTimeRange or others)
- validate (base validation rule logic)
