先说使用Java实现断点续传的关键点：
- 本地能够判断当前请求下载的文件是否已经存在，或者部分存在；
- 如果存在且为部分，则本地可以记录下载位置，并从服务器端继续完成下载；
技术上的关键：
- 请求的需要：
URL url = new URL("urlStr"); HttpURLConnection httpConnection = (HttpURLConnection)url.openConnection //向服务器请求指定区间段的数据，这是实现断点续传的根本。 httpcon.setRequestProperty("Range", "bytes=" + startPos+ "-" + endPos);
- 保存文件的方法：采用IO包中的RandomAccessFile类：
RandomAccess downfile = new RandomAccessFile("file","rw"); Array[] startPos ; downloadfile.seek(startPos); byte[] buf = new byte[1024]; downloadfile.write(buf, 0, length);
下面是功能实现的代码以及测试类，在测试类中传入实现类中的两个参数是线程数与文件来源（链接），断点续传的体现可以在下载过程中关闭IDE，在工程文件目录下，可以看到未下载完全的部分文件以及临时文件。重新启动IDE运行程序，文件继续下载，以下程序基本行行注释，在后面我还会说明一些：
一、功能实现类MultiThreadDownLoad
package **********;
import java.io.File;
import java.io.FileNotFoundException;
import java.io.IOException;
import java.io.InputStream;
import java.io.RandomAccessFile;
import java.net.HttpURLConnection;
import java.net.MalformedURLException;
import java.net.URL;
import java.util.concurrent.CountDownLatch;
import java.util.concurrent.ExecutorService;
import java.util.concurrent.Executors;
/*
* Encode:UTF-8
*
* Author:GFC
*
* 根据输入的url和设定的线程数，来完成断点续传功能。
*
* 每个线程支负责某一小段的数据下载；再通过RandomAccessFile完成数据的整合。
*/
public class MultiTheradDownLoad {
private String urlStr = null;
private String filename = null;
private String tmpfilename = null;
private int threadNum = 0;
private CountDownLatch latch = null;//设置一个计数器，代码内主要用来完成对缓存文件的删除
private long fileLength = 0l;
private long threadLength = 0l;
private long[] startPos;//保留每个线程下载数据的起始位置。
private long[] endPos;//保留每个线程下载数据的截止位置。
private boolean bool = false;
private URL url = null;
//有参构造函数，先构造需要的数据
public MultiTheradDownLoad(String urlStr, int threadNum) {
this.urlStr = urlStr;
this.threadNum = threadNum;
startPos = new long[this.threadNum];
endPos = new long[this.threadNum];
latch = new CountDownLatch(this.threadNum);
}
/*
* 组织断点续传功能的方法
*/
public void downloadPart() {
File file = null;
File tmpfile = null;
//设置HTTP网络访问代理
System.setProperty("http.proxySet", "true");
System.setProperty("http.proxyHost", "proxy3.bj.petrochina");
System.setProperty("http.proxyPort", "8080");
//从文件链接中获取文件名，此处没考虑文件名为空的情况，此种情况可能需使用UUID来生成一个唯一数来代表文件名。
filename = urlStr.substring(urlStr.lastIndexOf('/') + 1, urlStr
.contains("?") ? urlStr.lastIndexOf('?') : urlStr.length());
tmpfilename = filename + "_tmp";
try {
//创建url
url = new URL(urlStr);
//打开下载链接，并且得到一个HttpURLConnection的一个对象httpcon
HttpURLConnection httpcon = (HttpURLConnection) url.openConnection();
httpcon.setRequestMethod("GET");
//获取请求资源的总长度。
fileLength = httpcon.getContentLengthLong();
//下载文件和临时文件
file = new File(filename);//相对目录
tmpfile = new File(tmpfilename);
//每个线程需下载的资源大小；由于文件大小不确定，为避免数据丢失
threadLength = fileLength%threadNum == 0 ? fileLength/threadNum : fileLength/threadNum+1;
//打印下载信息
System.out.println("fileName: " + filename + " ," + "fileLength= "
+ fileLength + " the threadLength= " + threadLength);
//各个线程在exec线程池中进行，起始位置--结束位置
if (file.exists() && file.length() == fileLength) {
System.out.println("文件已存在!!");
return;
} else {
setBreakPoint(startPos, endPos, tmpfile);
ExecutorService exec = Executors.newCachedThreadPool();
for (int i = 0; i < threadNum; i++) {
exec.execute(new DownLoadThread(startPos[i], endPos[i],
this, i, tmpfile, latch));
}
latch.await();//当你的计数器减为0之前，会在此处一直阻塞。
exec.shutdown();
}
} catch (MalformedURLException e) {
e.printStackTrace();
} catch (IOException e) {
e.printStackTrace();
} catch (InterruptedException e) {
e.printStackTrace();
}
//下载完成后，判断文件是否完整，并删除临时文件
if (file.length() == fileLength) {
if (tmpfile.exists()) {
System.out.println("删除临时文件!!");
tmpfile.delete();
}
}
}
/*
* 断点设置方法，当有临时文件时，直接在临时文件中读取上次下载中断时的断点位置。没有临时文件，即第一次下载时，重新设置断点。
*
* Rantmpfile.seek()跳转到一个位置的目的是为了让各个断点存储的位置尽量分开。
*
* 这是实现断点续传的重要基础。
*/
private void setBreakPoint(long[] startPos, long[] endPos, File tmpfile) {
RandomAccessFile rantmpfile = null;
try {
if (tmpfile.exists()) {
System.out.println("继续下载!!");
rantmpfile = new RandomAccessFile(tmpfile, "rw");
for (int i = 0; i < threadNum; i++) {
rantmpfile.seek(8 * i + 8);
startPos[i] = rantmpfile.readLong();
rantmpfile.seek(8 * (i + 1000) + 16);
endPos[i] = rantmpfile.readLong();
System.out.println("the Array content in the exit file: ");
System.out.println("thre thread" + (i + 1) + " startPos:"
+ startPos[i] + ", endPos: " + endPos[i]);
}
} else {
System.out.println("the tmpfile is not available!!");
rantmpfile = new RandomAccessFile(tmpfile, "rw");
//最后一个线程的截止位置大小为请求资源的大小
for (int i = 0; i < threadNum; i++) {
startPos[i] = threadLength * i;
if (i == threadNum - 1) {
endPos[i] = fileLength;
} else {
endPos[i] = threadLength * (i + 1) - 1;
}
rantmpfile.seek(8 * i + 8);
rantmpfile.writeLong(startPos[i]);
rantmpfile.seek(8 * (i + 1000) + 16);
rantmpfile.writeLong(endPos[i]);
System.out.println("the Array content: ");
System.out.println("thre thread" + (i + 1) + " startPos:"
+ startPos[i] + ", endPos: " + endPos[i]);
}
}
} catch (FileNotFoundException e) {
e.printStackTrace();
} catch (IOException e) {
e.printStackTrace();
} finally {
try {
if (rantmpfile != null) {
rantmpfile.close();
}
} catch (IOException e) {
e.printStackTrace();
}
}
}
/*
* 实现下载功能的内部类，通过读取断点来设置向服务器请求的数据区间。
*/
class DownLoadThread implements Runnable {
private long startPos;
private long endPos;
private MultiTheradDownLoad task = null;
private RandomAccessFile downloadfile = null;
private int id;
private File tmpfile = null;
private RandomAccessFile rantmpfile = null;
private CountDownLatch latch = null;
public DownLoadThread(long startPos, long endPos,
MultiTheradDownLoad task, int id, File tmpfile,
CountDownLatch latch) {
this.startPos = startPos;
this.endPos = endPos;
this.task = task;
this.tmpfile = tmpfile;
try {
//
this.downloadfile = new RandomAccessFile(this.task.filename,"rw");
this.rantmpfile = new RandomAccessFile(this.tmpfile, "rw");
} catch (FileNotFoundException e) {
e.printStackTrace();
}
this.id = id;
this.latch = latch;
}
@Override
public void run() {
HttpURLConnection httpcon = null;
InputStream is = null;
int length = 0;
System.out.println("线程" + id + " 开始下载!!");
while (true) {
try {
httpcon = (HttpURLConnection) task.url.openConnection();
httpcon.setRequestMethod("GET");
//防止网络阻塞，设置指定的超时时间；单位都是ms。超过指定时间，就会抛出异常
httpcon.setReadTimeout(20000);//读取数据的超时设置
httpcon.setConnectTimeout(20000);//连接的超时设置
if (startPos < endPos) {
//向服务器请求指定区间段的数据，这是实现断点续传的根本。
httpcon.setRequestProperty("Range", "bytes=" + startPos+ "-" + endPos);
System.out.println("线程 " + id+ " 长度:---- "+ (endPos - startPos));
downloadfile.seek(startPos);
if (httpcon.getResponseCode() != HttpURLConnection.HTTP_OK
&& httpcon.getResponseCode() != HttpURLConnection.HTTP_PARTIAL) {
this.task.bool = true;
httpcon.disconnect();
downloadfile.close();
System.out.println("线程 ---" + id + " 下载完成!!");
latch.countDown();//计数器自减
break;
}
is = httpcon.getInputStream();//获取服务器返回的资源流
long count = 0l;
byte[] buf = new byte[1024];
while (!this.task.bool && (length = is.read(buf)) != -1) {
count += length;
downloadfile.write(buf, 0, length);
//不断更新每个线程下载资源的起始位置，并写入临时文件；为断点续传做准备
startPos += length;
rantmpfile.seek(8 * id + 8);
rantmpfile.writeLong(startPos);
}
System.out.println("线程 " + id
+ " 总下载大小: " + count);
//关闭流
is.close();
httpcon.disconnect();
downloadfile.close();
rantmpfile.close();
}
latch.countDown();//计数器自减
System.out.println("线程 " + id + " 下载完成!!");
break;
} catch (IOException e) {
e.printStackTrace();
} finally {
try {
if (is != null) {
is.close();
}
} catch (IOException e) {
e.printStackTrace();
}
}
}
}
}
}
其中设置HTTP访问代理在公网上不需要。可以看出来，整个类中分为三大部分：下载功能方法downloadPart();设置断点的私有方法setBreakPoint(long[] startPos, long[] endPos, File tmpfile)；以及实现下载功能的内部类，通过读取断点来设置向服务器请求的数据区间class DownLoadThread,其中重写了线程的run()方法。
二、测试类
package Download;
public class DownLoadTest {
public static void main(String[] args) {
//http://down.sandai.net/thunder9/Thunder9.1.40.898.exe
//http://download.firefox.com.cn/releases-sha2/stub/official/zh-CN/Firefox-latest.exe
//http://p0.ifengimg.com/pmop/2017/0830/4AB0880C38002DC837856D274469E3D09BDF43B9_size28_w500_h314.jpeg
//http://img.article.pchome.net/00/22/70/23/pic_lib/wm/03.jpg
int threadNum = 4;
String filepath = "http://down.sandai.net/thunder9/Thunder9.1.40.898.exe";
MultiTheradDownLoad load = new MultiTheradDownLoad(filepath ,threadNum);
load.downloadPart();
}
}
扫一扫
303
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
