使用Jackson PTH和Spring Data MongoDB DBRef的Java到JSON序列化生成额外的目标属性

从Java序列化为JSON时，target当使用@DBRef带有延迟加载和Jackson的多态类型处理的Spring Data MongoDB 批注时，Jackson会为引用的实体生成一个额外的属性。为什么会发生这种情况，并且可以省略多余的target属性？

代码示例

@Document(collection = "cdBox")
public class CDBox {
  @Id
  public String id;

  @DBRef(lazy = true)
  public List<Product> products;
}

@Document(collection = "album")
public class Album extends Product {
  @DBRef(lazy = true)
  public List<Song> songs;
}

@Document(collection = "single")
public class Single extends Product {
  @DBRef(lazy = true)
  public List<Song> songs;
}

@Document(collection = "song")
public class Song {
  @Id
  public String id;

  public String title;
}

@JsonTypeInfo(use = JsonTypeInfo.Id.NAME,
                    property = "productType",
                    include = JsonTypeInfo.As.EXTERNAL_PROPERTY)
@JsonSubTypes(value = {
    @JsonSubTypes.Type(value = Single.class),
    @JsonSubTypes.Type(value = Album.class)
})
public abstract class Product {
  @Id
  public String id;
}

生成的JSON

{
  "id": "someId1",
  "products": [
    {
      "id": "someId2",
      "songs": [
        {
        "id": "someId3",
        "title": "Some title",
        "target": {
          "id": "someId3",
          "title": "Some title"
          }
        }
      ]
    }
  ]
}