RestTemplate下载文件的另一种方式

使用RestTemplate下载文件最长用的方式是：

ResponseEntity<byte[]> rsp = restTemplate.getForEntity(url, byte[].class);

除了这种方式，还有另一种方式可以实现下载；

使用ResponseExtractor接口

ResponseExtractorj响应提取器：从Response中提取数据。RestTemplate请求完成后，都是通过它来从ClientHttpResponse提取出指定内容（比如请求头、请求Body体等）

ResponseExtractorj接口只有一个方法，当客户端和服务器端连接建立之后，会调用这个方法；

使用ResponseExtractor下载文件也有多种方式：

将文件下载到内存

继承接口ResponseExtractor,重写extractData方法

 @Override    public  byte[] extractData(ClientHttpResponse response) throws IOException {        long length = response.getHeaders().getContentLength();        InputStream in = response.getBody();        int byteRead;        for( byte[] b = new byte[4096]; (byteRead = in.read(b)) != -1){            byteArrayOutputStream.write(b,0,byteRead);        }        byteArrayOutputStream.flush();        return byteArrayOutputStream.toByteArray();    }}

下载到本地

public class SmallFileResponseExtractor        extends ResponseExtractor<File>{    private long downloadLength;    private String filePath;    public SmallFileResponseExtractor(String filePath){        this.filePath = filePath;    }    @Override    File extractData(ClientHttpResponse response) throws IOException {        InputStream in = response.getBody();        File file = new File(filePath);        FileOutputStream out = new FileOutputStream(file);        int byteRead;        for( byte[] b = new byte[4096]; (byteRead = in.read(b)) != -1;downloadLength += byteRead){            out.write(b,0,byteRead);        }        out.flush();        out.close();        return file;    }}

调用方式：

使用StopWatch计算耗时时间

import java.io.File;import java.io.IOException;import java.nio.file.Files;import java.nio.file.Paths;import java.util.Objects;public class FileDownloader {    RestTemplate restTemplate = new RestTemplateBuilder().build();    public  void downloadFileToMem(String file_url,String file_path) throws IOException {        StopWatch stopWatch = new StopWatch("下载");        stopWatch.start();        byte[] body = restTemplate.execute(file_url, HttpMethod.GET,                null, new ByteArrayResponseExtractor());        Files.write(Paths.get(file_path), Objects.requireNonNull(body));        stopWatch.stop();        double total = stopWatch.getTotalTimeSeconds();        System.out.println("下载总耗时：" + total );    }    public  void downloadFile(String file_url,String file_path) throws IOException {        StopWatch stopWatch = new StopWatch("下载");        stopWatch.start();        File file = restTemplate.execute(file_url, HttpMethod.GET,                null, new SmallFileResponseExtractor(file_path));        stopWatch.stop();        double total = stopWatch.getTotalTimeSeconds();        System.out.println(file.getAbsolutePath() +"下载总耗时：" + total );    }}

以上方法，适合于小文件下载，对应大文件，还需一个多线程的方法下载，如果要考虑下载速度的话；

多线程下载

使用CompletableFuture

        ExecutorService executorService = Executors.newFixedThreadPool(threadNum);       long step = contentLength / threadNum;        List<CompletableFuture<File>> futures = new ArrayList<>();        for (int index = 0; index < threadNum; index++) {            String start = step * index + "";            String end = index == threadNum - 1 ? "" : (step * (index + 1) - 1) + "";            String tempFilePath = dir + File.separator + "." + fileName + "." + index;            SmallFileResponseExtractorextractor = new SmallFileResponseExtractor(tempFilePath);            CompletableFuture<File> future = CompletableFuture.supplyAsync(() -> {                RequestCallback callback = request -> {                    //设置HTTP请求头Range信息，开始下载到临时文件                    request.getHeaders().add(HttpHeaders.RANGE, "bytes=" + start + "-" + end);                };                return restTemplate.execute(fileURL, HttpMethod.GET, callback, extractor);            }, executorService).exceptionally(e -> {                e.printStackTrace();                return null;            });            futures.add(future);        }        //创建最终文件        String tmpFilePath = dir + File.separator + fileName + ".temp";        File file = new File(tmpFilePath);        FileChannel outChannel = new FileOutputStream(file).getChannel();        futures.forEach(future -> {            try {                File tmpFile = future.get();                FileChannel tmpIn = new FileInputStream(tmpFile).getChannel();                //合并每个临时文件                outChannel.transferFrom(tmpIn, outChannel.size(), tmpIn.size());                tmpIn.close();                tmpFile.delete(); //合并完成后删除临时文件            } catch (Exception e) {                e.printStackTrace();            }        });        outChannel.close();        executorService.shutdown();        file.renameTo(new File(dir + File.separator + fileName));