Question

目前，我的Jersey Implemented Rest Web Service方法能够根据userId从数据库表中获取String列表。服务将一次只向用户客户端调用发送一个字符串。剩余应存储在缓存中一段时间和该userId。因此，当下一次调用Web服务时，它会检查userId和时间限制是否进行新的数据库查询或使用缓存的结果来返回结果。

请帮助我在这里使用缓存方法。我完成了我的服务方法。

我的示例Web服务方法。如何在这里放置缓存技术：

@GET
@Path("/CacheService/{userId}")
public Response invokeCacheService(@PathParam("id") String userId, @Context Request req){

    Response.ResponseBuilder rb = null;
        ArrayList songList = new ArrayList(); 
        songList = UserService.getPlaylistSongs(userId);
        rb = Response.ok(songList);

    return rb.build();

}

Answer 1

这是一个util类，以便在服务器端进行HTTP缓存控制：

    public class HttpCacheRizze extends CacheControl {


        public static CacheControl minutesSecondesMilliseconds(int min, int sec, int milli){
            HttpCacheRizze cc=new HttpCacheRizze();     
            cc.setMaxAge(min*60+sec+milli/1000);
            cc.setPrivate(true);
            return cc;
        }

        public static EntityTag etag(String tag) {
            EntityTag etag = new EntityTag(DigestUtils.sha256Hex(tag));
            return etag;
        }

        /**
         * 
         * @param req
         * @param tag
         * @param timeout
         * @return response if isUnderCache or null if not
         */
        public static Response getCachedResponseMilliseconds(Request req, String tag,int timeoutMs){
            Response.ResponseBuilder rb=null;
            EntityTag etag = etag(tag);     
            if(req!=null){
                rb = req.evaluatePreconditions(new Date(),etag);
                if (rb != null){

                    return rb.cacheControl(HttpCacheRizze.minutesSecondesMilliseconds(0,0,timeout)).tag(etag).build();
                }
            }
            return null;
        }



        /**
         * 
         * @param status
         * @param entity
         * @param tag
         * @param timeout
         * @return response will be cached
         */
        public static Response getCacheInvalidatedResponse(int status, String entity, String tag, int timeoutMs){

            //if entity is null, force status to 204 (empty)
            if(status==204 || entity==null || entity.compareTo("null")==0 || entity.compareTo("{\"null\"}")==0) {
                status=204;
                entity=null;
            }           

            return Response.status(status).entity(entity)
                    .cacheControl(HttpCacheRizze.minutesSecondesMilliseconds(0, 0, timeoutMs))
                    .tag(etag(tag))
                    .build();
        }

    }

您的代码如下：

    @GET
    @Path("/CacheService/{userId}")
    public Response invokeCacheService(@PathParam("id") String userId, @Context Request req){

            //the key of a good cache control technique, is to : be quick in order to determine if present or not in cache, 
            //and to try to avoid the maximum data processing in order to retrieve fromthe cache (example avoid performing getPlaylistSong  under cache
           int TTL_CACHE_SONGS=10000; //in ms
           String tag =  "CacheService"+userid";

           //is under cache ?
           Response r = HttpCacheRizze.getCachedResponseMilliseconds(req, tag, TTL_CACHE_SONGS);
           if(r!=null){
                // under cache
                return r;
           }

           // cache is not present or need to be refreshed

            ArrayList songList = new ArrayList(); 
            songList = UserService.getPlaylistSongs(userId);
            int status = 200;

            //catch here errors .... empty....
            if(songList==null || songList.size()==0 )
                status = 204;

            r = HttpCacheRizze.getCacheInvalidatedResponse(status, new Gson().toJson(songList), tag, TTL_CACHE_SONGS);

            return r;

    }

在Jersey中缓存实现了Web服务

1 个答案: