如何设计一个 DataSource/Service 类以获取网络或本地数据？

Question

假设一个很常见的场景：
XYZMovie(一个虚构电影网站)提供了接口可以让开发者获取某部电影的基本信息（电影名、演员、放映时间、评分等）

接口类似这样：http:api.xyzmovie.com/get?nam...

我利用爬虫缓存了该网站的数据到本地数据库，现在有一个电影查询程序，它应该能设置本地或网络优先查询，如本地或网络查询无果则换另一种查询：

class XYZMovieDataSource
{
    XYZMovie GetMovie(string id or name, etc, priority)
    {
        XYZMovie m;
        if(priority == local) { dosomething }
        else { dosomething }
        return m; 
    }
    
    XYZMovie GetMovieFromLocal(string id or name, etc) { dosomething }
    XYZMovie GetMovieFromNetwork(string id or name, etc) { dosomething }
}

请问有什么更好的设计模式吗？

Answer 1

import java.util.Random;
import java.util.concurrent.CountDownLatch;

public class MovieDataSource {
    public static void main(String[] args) {
        Movie movie = new MovieDataSource().getMovie("码匠笔记");
        System.out.println("最终得到视频：" + movie);
    }

    public Movie getMovie(String id) {
        CountDownLatch countDownLatch = new CountDownLatch(1);
        MovieGetter remoteMovieGetter = new RemoteMovieGetter(id, countDownLatch);
        MovieGetter localMovieGetter = new LocalMovieGetter(id, countDownLatch);

        new Thread(remoteMovieGetter).start();
        new Thread(localMovieGetter).start();

        try {
            countDownLatch.await();
        } catch (InterruptedException e) {
            System.out.println("这都出异常了？还想获取啥");
        }

        if (remoteMovieGetter.getMovie() != null) {
            return remoteMovieGetter.getMovie();
        }
        if (localMovieGetter.getMovie() != null) {
            return localMovieGetter.getMovie();
        }

        return null;
    }

    public static abstract class MovieGetter implements Runnable {
        protected String id;
        protected CountDownLatch countDownLatch;

        public MovieGetter(String id, CountDownLatch countDownLatch) {
            this.id = id;
            this.countDownLatch = countDownLatch;
        }

        public abstract Movie getMovie();
    }

    public static class RemoteMovieGetter extends MovieGetter {
        public RemoteMovieGetter(String id, CountDownLatch countDownLatch) {
            super(id, countDownLatch);
        }

        private Movie movie;

        @Override
        public Movie getMovie() {
            return movie;
        }

        @Override
        public void run() {
            System.out.println("拉取远程视频");
            try {
                Thread.sleep(new Random().nextInt(3) * 1000);
                System.out.println("获取远程视频");
                movie = new Movie("获取远程视频");
            } catch (InterruptedException e) {
                e.printStackTrace();
            } finally {
                countDownLatch.countDown();
            }
        }
    }

    public static class LocalMovieGetter extends MovieGetter {

        public LocalMovieGetter(String id, CountDownLatch countDownLatch) {
            super(id, countDownLatch);
        }

        private Movie movie;

        @Override
        public Movie getMovie() {
            return movie;
        }

        @Override
        public void run() {
            System.out.println("拉取本地视频");
            try {
                Thread.sleep(new Random().nextInt(3) * 1000);
                System.out.println("获取本地视频");
                movie = new Movie("获取本地视频");
            } catch (InterruptedException e) {
                e.printStackTrace();
            } finally {
                countDownLatch.countDown();
            }
        }
    }

    public static class Movie {
        private String name;

        public Movie(String name) {
            this.name = name;
        }

        @Override
        public String toString() {
            return this.name;
        }
    }
}

使用 CountDownLatch ，设置为 1，同时多线程去获取远程的本地，谁先获取到，返回谁。
多次运行成功你会发现可以有不同的值
因为
Thread.sleep(new Random().nextInt(3) * 1000); 随机设置他们的快慢

正如如果你本地数据库中没有，那么自然就去等待远程了，那么你本地可以不调用
countDownLatch.countDown();
主线程就会一直等待

Answer 2

没有发现有使用设计模式的必要，如果非要说，我觉得可以通过url的路径将两种方式区分开，一个叫cache（也就是你的local），一个叫origin（也就是你的network）。