Jackson定制序列化和反序列化

Question

i'm unable to figure out the proper way to implement the custom serialization/deserialization with jackson. I have many classes (~50) with primitive fields that should be serialized/deserialized not as primitives. like:

class User {
    int height // this field should be serialized as "height": "10 m"
}

class Food {
    int temperature // this field should be serialized as "temperature": "50 C"
}

class House {
    int width // this field should be serialized as "width": "10 m"
}

all serializations and deserializations are very similar, I just need to add a suffix after the integer (C, pages, meters, etc..)

A straightforward way to do this is to put a pair of @JsonSerialize / @JsonDeserialize annotation to each such field and implement them. But i will end up with 100 very similar serializers / deserializers.

I thought about adding custom annotation to each field, say @Units("Degree") or @Units("Meters") , to such integer fields and implement a SerializationProvider that will create serializers in a generic way based on an annotation value. But I didn't find a place where the information about the property annotations is available.

Answer 1

Idea with Unit annotation is really good. We need to only add custom com.fasterxml.jackson.databind.ser.BeanSerializerModifier and com.fasterxml.jackson.databind.ser.BeanPropertyWriter implementations. Let's create first our annotation class:

@Retention(RetentionPolicy.RUNTIME)
@Target(ElementType.FIELD)
@interface Unit {
    String value();
}

POJO model could look like below:

class Pojo {

    private User user = new User();
    private Food food = new Food();
    private House house = new House();

    // getters, setters, toString
}

class User {

    @Unit("m")
    private int height = 10;

    // getters, setters, toString
}

class Food {

    @Unit("C")
    private int temperature = 50;

    // getters, setters, toString
}

class House {

    @Unit("m")
    private int width = 10;

    // getters, setters, toString
}

Having all of that we need to customise property serialisation:

class UnitBeanSerializerModifier extends BeanSerializerModifier {

    @Override
    public List<BeanPropertyWriter> changeProperties(SerializationConfig config, BeanDescription beanDesc, List<BeanPropertyWriter> beanProperties) {
        for (int i = 0; i < beanProperties.size(); ++i) {
            final BeanPropertyWriter writer = beanProperties.get(i);
            AnnotatedMember member = writer.getMember();
            Unit units = member.getAnnotation(Unit.class);
            if (units != null) {
                beanProperties.set(i, new UnitBeanPropertyWriter(writer, units.value()));
            }
        }
        return beanProperties;
    }
}

class UnitBeanPropertyWriter extends BeanPropertyWriter {

    private final String unit;

    protected UnitBeanPropertyWriter(BeanPropertyWriter base, String unit) {
        super(base);
        this.unit = unit;
    }

    @Override
    public void serializeAsField(Object bean, JsonGenerator gen, SerializerProvider prov) throws Exception {
        gen.writeFieldName(_name);
        final Object value = (_accessorMethod == null) ? _field.get(bean) : _accessorMethod.invoke(bean, (Object[]) null);
        gen.writeString(value + " " + unit);
    }
}

Using SimpleModule we can register it and use with ObjectMapper :

import com.fasterxml.jackson.core.JsonGenerator;
import com.fasterxml.jackson.databind.BeanDescription;
import com.fasterxml.jackson.databind.ObjectMapper;
import com.fasterxml.jackson.databind.SerializationConfig;
import com.fasterxml.jackson.databind.SerializerProvider;
import com.fasterxml.jackson.databind.introspect.AnnotatedMember;
import com.fasterxml.jackson.databind.module.SimpleModule;
import com.fasterxml.jackson.databind.ser.BeanPropertyWriter;
import com.fasterxml.jackson.databind.ser.BeanSerializerModifier;

import java.lang.annotation.ElementType;
import java.lang.annotation.Retention;
import java.lang.annotation.RetentionPolicy;
import java.lang.annotation.Target;
import java.util.List;

public class JsonApp {

    public static void main(String[] args) throws Exception {
        SimpleModule unitModule = new SimpleModule();
        unitModule.setSerializerModifier(new UnitBeanSerializerModifier());

        ObjectMapper mapper = new ObjectMapper();
        mapper.registerModule(unitModule);

        Pojo pojo = new Pojo();
        System.out.println(mapper.writeValueAsString(pojo));
    }
}

prints:

{
  "user" : {
    "height" : "10 m"
  },
  "food" : {
    "temperature" : "50 C"
  },
  "house" : {
    "width" : "10 m"
  }
}

Of course, you need to test it and handle all corner cases but above example shows general idea. In the similar way we can handle deserialisation. We need to implement custom BeanDeserializerModifier and one custom UnitDeserialiser :

class UnitBeanDeserializerModifier extends BeanDeserializerModifier {

    @Override
    public JsonDeserializer<?> modifyDeserializer(DeserializationConfig config, BeanDescription beanDesc, JsonDeserializer<?> deserializer) {
        JsonDeserializer<?> jsonDeserializer = super.modifyDeserializer(config, beanDesc, deserializer);
        if (jsonDeserializer instanceof StdScalarDeserializer) {
            StdScalarDeserializer scalarDeserializer = (StdScalarDeserializer) jsonDeserializer;
            Class scalarClass = scalarDeserializer.handledType();
            if (int.class == scalarClass) {
                return new UnitIntStdScalarDeserializer(scalarDeserializer);
            }
        }
        return jsonDeserializer;
    }
}

and example deserialiser for int :

class UnitIntStdScalarDeserializer extends StdScalarDeserializer<Integer> {

    private StdScalarDeserializer<Integer> src;

    public UnitIntStdScalarDeserializer(StdScalarDeserializer<Integer> src) {
        super(src);
        this.src = src;
    }

    @Override
    public Integer deserialize(JsonParser p, DeserializationContext ctxt) throws IOException {
        String value = p.getValueAsString();
        String[] parts = value.split("\\s+");
        if (parts.length == 2) {
            return Integer.valueOf(parts[0]);
        }
        return src.deserialize(p, ctxt);
    }
}

Above implementation is just an example and should be improved for other primitive types. We can register it in the same way using simple module. Reuse the same as for serialisation:

unitModule.setDeserializerModifier(new UnitBeanDeserializerModifier());