Question

如何使用Jersey序列化Java 8 java.util.Stream<T>。我尝试编写MessageBodyWriter，但我需要知道如何使用MessageBodyWriters的新MessageBodyWriter为现有Stream撰写（装饰）。

Stream<String> get(){
  return some stream of strings  
}

public <T> class StreamMessageBodyWriter<Stream<T>> 
           implements MessageBodyWriter<Stream<T>> {

  public void writeTo(.......){
    //How can I get the handle to MessageBodyWriter that will write for type T, 
    //so that I can 'collect' the 'java.util.Stream<T>' and write it to 
    //OutputStream
  }
}

Answer 1

但是我需要知道如何使用MessageBodyWriters的新MessageBodyWriter撰写（装饰）现有Stream

您可以注入Providers并使用getMessagBodyWriter(...)，传递所需的详细信息以查找该类型的特定编写者。例如

@Provider
public class StreamBodyWriter implements MessageBodyWriter<Stream> {

    @Context
    private Providers providers;

    @Override
    public boolean isWriteable(Class<?> type, Type genericType, 
            Annotation[] annotations, MediaType mediaType) {
        return Stream.class.isAssignableFrom(type);
    }

    @Override
    public long getSize(Stream stream, Class<?> type, Type genericType, 
            Annotation[] annotations, MediaType mediaType) { return -1; }

    @Override
    public void writeTo(Stream stream, Class<?> type, Type genericType, 
            Annotation[] annotations, MediaType mediaType, 
            MultivaluedMap<String, Object> httpHeaders, OutputStream entityStream) 
            throws IOException, WebApplicationException {

        Object obj = stream.collect(Collectors.toList());
        Class<?> objType = obj.getClass();

        MessageBodyWriter writer = providers.getMessageBodyWriter(objType, 
                null, annotations, mediaType);

        writer.writeTo(obj, objType, null, annotations, 
                mediaType, httpHeaders, entityStream);

    }
}

如果查看writeTo，请先致电collect，然后获取返回的类型。然后查找该类型的编写器，然后简单地委托给编写者。

这是一个测试

@Path("stream")
public class StreamResource {

    @GET
    @Produces(MediaType.APPLICATION_JSON)
    public Response getStream() {
        List<Person> myList = Arrays.asList(
                              new Person("Stack"), 
                              new Person("Overflow"),
                              new Person("Sam"));
        Stream<Person> stream = myList.stream()
                                      .filter(p -> p.name.startsWith("S"));
        return Response.ok(stream).build();
    }

    public static class Person {
        public String name;
        public Person(String name) { this.name = name; }
        public Person() {}
    }
}

C:\>curl -v http://localhost:8080/api/stream
  的结果：
  [{"name":"Stack"},{"name":"Sam"}]

顺便说一句，如果您打算在编写器中操作Stream，可以考虑使用Interceptor。真的不会有所作为，但是如果你想坚持单一责任原则，这就是Interceptor的用途，操纵请求体。

注意：以上是标准的JAX-RS

另外...

特别是对于Jersey，您还可以注入MessageBodyWorkers，以进行更具体的查找，甚至可以调用writeTo，如果有的话，将会委托给所需的编写者。

Answer 2

根据使用流的目的（不使用stream.collect(Collectors.toList())），这篇有趣的文章显示how to serialize large data from a database。

它是这样的......

@GET
@Produces( "application/json" )
public Response streamGeneratedUuids() {

    return getNoCacheResponseBuilder( Response.Status.OK ).entity( new StreamingOutput() {

        @Override
        public void write( OutputStream os ) throws IOException, WebApplicationException {
            try (PrintWriter writer = new PrintWriter(new BufferedWriter(new OutputStreamWriter(os))) ) {
                //iterate the java.util.stream and write to the OutputStream
                writer.print("....");       
            }
        }
    }).build();
}

它没有用MessageBodyWriter实现，但在我看来可以改编。

使用Jersey序列化Java 8流

2 个答案:

另外...