Java 多线程断点下载文件

约定不等于承诺〃 2022-07-15 15:58 187阅读 0赞

#  #

# 基本原理： #

利用URLConnection获取要下载文件的长度、头部等相关信息，并设置响应的头部信息。并且通过URLConnection获取输入流，将文件分成指定的块，每一块单独开辟一个线程完成数据的读取、写入。通过输入流读取下载文件的信息，然后将读取的信息用RandomAccessFile随机写入到本地文件中。同时，每个线程写入的数据都文件指针也就是写入数据的长度，需要保存在一个临时文件中。这样当本次下载没有完成的时候，下次下载的时候就从这个文件中读取上一次下载的文件长度，然后继续接着上一次的位置开始下载。并且将本次下载的长度写入到这个文件中。

# 一、下载文件信息类、实体 #

封装即将下载资源的信息

package com.hoo.entity;
 
 /**
 * function: 下载文件信息类
 * @author hoojo
 * @createDate 2011-9-21 下午05:14:58
 * @file DownloadInfo.java
 * @package com.hoo.entity
 * @project MultiThreadDownLoad
 * @blog http://blog.csdn.net/IBM_hoojo
 * @email hoojo_@126.com
 * @version 1.0
 */
 public class DownloadInfo {
 //下载文件url
 private String url;
 //下载文件名称
 private String fileName;
 //下载文件路径
 private String filePath;
 //分成多少段下载， 每一段用一个线程完成下载
 private int splitter;
 
 //下载文件默认保存路径
 private final static String FILE_PATH = "C:/temp";
 //默认分块数、线程数
 private final static int SPLITTER_NUM = 5;
 
 public DownloadInfo() {
 super();
 }
 
 /**
 * @param url 下载地址
 */
 public DownloadInfo(String url) {
 this(url, null, null, SPLITTER_NUM);
 }
 
 /**
 * @param url 下载地址url
 * @param splitter 分成多少段或是多少个线程下载
 */
 public DownloadInfo(String url, int splitter) {
 this(url, null, null, splitter);
 }
 
 /***
 * @param url 下载地址
 * @param fileName 文件名称
 * @param filePath 文件保存路径
 * @param splitter 分成多少段或是多少个线程下载
 */
 public DownloadInfo(String url, String fileName, String filePath, int splitter) {
 super();
 if (url == null || "".equals(url)) {
 throw new RuntimeException("url is not null!");
 }
 this.url = url;
 this.fileName = (fileName == null || "".equals(fileName)) ? getFileName(url) : fileName;
 this.filePath = (filePath == null || "".equals(filePath)) ? FILE_PATH : filePath;
 this.splitter = (splitter < 1) ? SPLITTER_NUM : splitter;
 }
 
 /**
 * function: 通过url获得文件名称
 * @author hoojo
 * @createDate 2011-9-30 下午05:00:00
 * @param url
 * @return
 */
 private String getFileName(String url) {
 return url.substring(url.lastIndexOf("/") + 1, url.length());
 }
 
 public String getUrl() {
 return url;
 }
 
 public void setUrl(String url) {
 if (url == null || "".equals(url)) {
 throw new RuntimeException("url is not null!");
 }
 this.url = url;
 }
 
 public String getFileName() {
 return fileName;
 }
 
 public void setFileName(String fileName) {
 this.fileName = (fileName == null || "".equals(fileName)) ? getFileName(url) : fileName;
 }
 
 public String getFilePath() {
 return filePath;
 }
 
 public void setFilePath(String filePath) {
 this.filePath = (filePath == null || "".equals(filePath)) ? FILE_PATH : filePath;
 }
 
 public int getSplitter() {
 return splitter;
 }
 
 public void setSplitter(int splitter) {
 this.splitter = (splitter < 1) ? SPLITTER_NUM : splitter;
 }
 
 @Override
 public String toString() {
 return this.url + "#" + this.fileName + "#" + this.filePath + "#" + this.splitter;
 }
 }

# 二、随机写入一段文件 #

package com.hoo.download;
 
 import java.io.IOException;
 import java.io.RandomAccessFile;
 
 /**
 * function: 写入文件、保存文件
 * @author hoojo
 * @createDate 2011-9-21 下午05:44:02
 * @file SaveItemFile.java
 * @package com.hoo.download
 * @project MultiThreadDownLoad
 * @blog http://blog.csdn.net/IBM_hoojo
 * @email hoojo_@126.com
 * @version 1.0
 */
 public class SaveItemFile {
 //存储文件
 private RandomAccessFile itemFile;
 
 public SaveItemFile() throws IOException {
 this("", 0);
 }
 
 /**
 * @param name 文件路径、名称
 * @param pos 写入点位置 position
 * @throws IOException
 */
 public SaveItemFile(String name, long pos) throws IOException {
 itemFile = new RandomAccessFile(name, "rw");
 //在指定的pos位置开始写入数据
 itemFile.seek(pos);
 }
 
 /**
 * function: 同步方法写入文件
 * @author hoojo
 * @createDate 2011-9-26 下午12:21:22
 * @param buff 缓冲数组
 * @param start 起始位置
 * @param length 长度
 * @return
 */
 public synchronized int write(byte[] buff, int start, int length) {
 int i = -1;
 try {
 itemFile.write(buff, start, length);
 i = length;
 } catch (IOException e) {
 e.printStackTrace();
 }
 return i;
 }
 
 public void close() throws IOException {
 if (itemFile != null) {
 itemFile.close();
 }
 }
 }

这个类主要是完成向本地的指定文件指针出开始写入文件，并返回当前写入文件的长度（文件指针）。这个类将被线程调用，文件被分成对应的块后，将被线程调用。每个线程都将会调用这个类完成文件的随机写入。

# 三、单个线程下载文件 #

package com.hoo.download;
 
 import java.io.IOException;
 import java.io.InputStream;
 import java.net.HttpURLConnection;
 import java.net.MalformedURLException;
 import java.net.URL;
 import java.net.URLConnection;
 import com.hoo.util.LogUtils;
 
 /**
 * function: 单线程下载文件
 * @author hoojo
 * @createDate 2011-9-22 下午02:55:10
 * @file DownloadFile.java
 * @package com.hoo.download
 * @project MultiThreadDownLoad
 * @blog http://blog.csdn.net/IBM_hoojo
 * @email hoojo_@126.com
 * @version 1.0
 */
 public class DownloadFile extends Thread {
 
 //下载文件url
 private String url;
 //下载文件起始位置 
 private long startPos;
 //下载文件结束位置
 private long endPos;
 //线程id
 private int threadId;
 
 //下载是否完成
 private boolean isDownloadOver = false;
 
 private SaveItemFile itemFile;
 
 private static final int BUFF_LENGTH = 1024 * 8;
 
 /**
 * @param url 下载文件url
 * @param name 文件名称
 * @param startPos 下载文件起点
 * @param endPos 下载文件结束点
 * @param threadId 线程id
 * @throws IOException
 */
 public DownloadFile(String url, String name, long startPos, long endPos, int threadId) throws IOException {
 super();
 this.url = url;
 this.startPos = startPos;
 this.endPos = endPos;
 this.threadId = threadId;
 //分块下载写入文件内容
 this.itemFile = new SaveItemFile(name, startPos);
 }
 
 
 @Override
 public void run() {
 while (endPos > startPos && !isDownloadOver) {
 try {
 URL url = new URL(this.url);
 HttpURLConnection conn = (HttpURLConnection) url.openConnection();
 
 // 设置连接超时时间为10000ms
 conn.setConnectTimeout(10000);
 // 设置读取数据超时时间为10000ms
 conn.setReadTimeout(10000);
 
 setHeader(conn);
 
 String property = "bytes=" + startPos + "-";
 conn.setRequestProperty("RANGE", property);
 
 //输出log信息
 LogUtils.log("开始 " + threadId + "：" + property + endPos);
 //printHeader(conn);
 
 //获取文件输入流，读取文件内容
 InputStream is = conn.getInputStream();
 
 byte[] buff = new byte[BUFF_LENGTH];
 int length = -1;
 LogUtils.log("#start#Thread: " + threadId + ", startPos: " + startPos + ", endPos: " + endPos);
 while ((length = is.read(buff)) > 0 && startPos < endPos && !isDownloadOver) {
 //写入文件内容，返回最后写入的长度
 startPos += itemFile.write(buff, 0, length);
 }
 LogUtils.log("#over#Thread: " + threadId + ", startPos: " + startPos + ", endPos: " + endPos);
 LogUtils.log("Thread " + threadId + " is execute over!");
 this.isDownloadOver = true;
 } catch (MalformedURLException e) {
 e.printStackTrace();
 } catch (IOException e) {
 e.printStackTrace();
 } finally {
 try {
 if (itemFile != null) {
 itemFile.close();
 }
 } catch (IOException e) {
 e.printStackTrace();
 }
 }
 }
 if (endPos < startPos && !isDownloadOver) {
 LogUtils.log("Thread " + threadId + " startPos > endPos, not need download file !");
 this.isDownloadOver = true;
 }
 if (endPos == startPos && !isDownloadOver) {
 LogUtils.log("Thread " + threadId + " startPos = endPos, not need download file !");
 this.isDownloadOver = true;
 }
 }
 
 /**
 * function: 打印下载文件头部信息
 * @author hoojo
 * @createDate 2011-9-22 下午05:44:35
 * @param conn HttpURLConnection
 */
 public static void printHeader(URLConnection conn) {
 int i = 1;
 while (true) {
 String header = conn.getHeaderFieldKey(i);
 i++;
 if (header != null) {
 LogUtils.info(header + ":" + conn.getHeaderField(i));
 } else {
 break;
 }
 }
 }
 
 /**
 * function: 设置URLConnection的头部信息，伪装请求信息
 * @author hoojo
 * @createDate 2011-9-28 下午05:29:43
 * @param con
 */
 public static void setHeader(URLConnection conn) {
 conn.setRequestProperty("User-Agent", "Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9.0.3) Gecko/2008092510 Ubuntu/8.04 (hardy) Firefox/3.0.3");
 conn.setRequestProperty("Accept-Language", "en-us,en;q=0.7,zh-cn;q=0.3");
 conn.setRequestProperty("Accept-Encoding", "utf-8");
 conn.setRequestProperty("Accept-Charset", "ISO-8859-1,utf-8;q=0.7,*;q=0.7");
 conn.setRequestProperty("Keep-Alive", "300");
 conn.setRequestProperty("connnection", "keep-alive");
 conn.setRequestProperty("If-Modified-Since", "Fri, 02 Jan 2009 17:00:05 GMT");
 conn.setRequestProperty("If-None-Match", "\"1261d8-4290-df64d224\"");
 conn.setRequestProperty("Cache-conntrol", "max-age=0");
 conn.setRequestProperty("Referer", "http://www.baidu.com");
 }
 
 public boolean isDownloadOver() {
 return isDownloadOver;
 }
 
 public long getStartPos() {
 return startPos;
 }
 
 public long getEndPos() {
 return endPos;
 }
 }

这个类主要是完成单个线程的文件下载，将通过URLConnection读取指定url的资源信息。然后用InputStream读取文件内容，然后调用调用SaveItemFile类，向本地写入当前要读取的块的内容。

# 四、分段多线程写入文件内容 #

package com.hoo.download;
 
 import java.io.DataInputStream;
 import java.io.DataOutputStream;
 import java.io.File;
 import java.io.FileInputStream;
 import java.io.FileOutputStream;
 import java.io.IOException;
 import java.net.HttpURLConnection;
 import java.net.MalformedURLException;
 import java.net.URL;
 import com.hoo.entity.DownloadInfo;
 import com.hoo.util.LogUtils;
 
 /**
 * function: 分批量下载文件
 * @author hoojo
 * @createDate 2011-9-22 下午05:51:54
 * @file BatchDownloadFile.java
 * @package com.hoo.download
 * @project MultiThreadDownLoad
 * @blog http://blog.csdn.net/IBM_hoojo
 * @email hoojo_@126.com
 * @version 1.0
 */
 public class BatchDownloadFile implements Runnable {
 //下载文件信息 
 private DownloadInfo downloadInfo;
 //一组开始下载位置
 private long[] startPos;
 //一组结束下载位置
 private long[] endPos;
 //休眠时间
 private static final int SLEEP_SECONDS = 500;
 //子线程下载
 private DownloadFile[] fileItem;
 //文件长度
 private int length;
 //是否第一个文件
 private boolean first = true;
 //是否停止下载
 private boolean stop = false;
 //临时文件信息
 private File tempFile;
 
 public BatchDownloadFile(DownloadInfo downloadInfo) {
 this.downloadInfo = downloadInfo;
 String tempPath = this.downloadInfo.getFilePath() + File.separator + downloadInfo.getFileName() + ".position";
 tempFile = new File(tempPath);
 //如果存在读入点位置的文件
 if (tempFile.exists()) {
 first = false;
 //就直接读取内容
 try {
 readPosInfo();
 } catch (IOException e) {
 e.printStackTrace();
 }
 } else {
 //数组的长度就要分成多少段的数量
 startPos = new long[downloadInfo.getSplitter()];
 endPos = new long[downloadInfo.getSplitter()];
 }
 }
 
 @Override
 public void run() {
 //首次下载，获取下载文件长度
 if (first) {
 length = this.getFileSize();//获取文件长度
 if (length == -1) {
 LogUtils.log("file length is know!");
 stop = true;
 } else if (length == -2) {
 LogUtils.log("read file length is error!");
 stop = true;
 } else if (length > 0) {
 /**
 * eg 
 * start: 1, 3, 5, 7, 9
 * end: 3, 5, 7, 9, length
 */
 for (int i = 0, len = startPos.length; i < len; i++) {
 int size = i * (length / len);
 startPos[i] = size;
 
 //设置最后一个结束点的位置
 if (i == len - 1) {
 endPos[i] = length;
 } else {
 size = (i + 1) * (length / len);
 endPos[i] = size;
 }
 LogUtils.log("start-end Position[" + i + "]: " + startPos[i] + "-" + endPos[i]);
 }
 } else {
 LogUtils.log("get file length is error, download is stop!");
 stop = true;
 }
 }
 
 //子线程开始下载
 if (!stop) {
 //创建单线程下载对象数组
 fileItem = new DownloadFile[startPos.length];//startPos.length = downloadInfo.getSplitter()
 for (int i = 0; i < startPos.length; i++) {
 try {
 //创建指定个数单线程下载对象，每个线程独立完成指定块内容的下载
 fileItem[i] = new DownloadFile(
 downloadInfo.getUrl(), 
 this.downloadInfo.getFilePath() + File.separator + downloadInfo.getFileName(), 
 startPos[i], endPos[i], i
 );
 fileItem[i].start();//启动线程，开始下载
 LogUtils.log("Thread: " + i + ", startPos: " + startPos[i] + ", endPos: " + endPos[i]);
 } catch (IOException e) {
 e.printStackTrace();
 }
 }
 
 //循环写入下载文件长度信息
 while (!stop) {
 try {
 writePosInfo();
 LogUtils.log("downloading……");
 Thread.sleep(SLEEP_SECONDS);
 stop = true;
 } catch (IOException e) {
 e.printStackTrace();
 } catch (InterruptedException e) {
 e.printStackTrace();
 }
 for (int i = 0; i < startPos.length; i++) {
 if (!fileItem[i].isDownloadOver()) {
 stop = false;
 break;
 }
 }
 }
 LogUtils.info("Download task is finished!");
 }
 }
 
 /**
 * 将写入点数据保存在临时文件中
 * @author hoojo
 * @createDate 2011-9-23 下午05:25:37
 * @throws IOException
 */
 private void writePosInfo() throws IOException {
 DataOutputStream dos = new DataOutputStream(new FileOutputStream(tempFile));
 dos.writeInt(startPos.length);
 for (int i = 0; i < startPos.length; i++) {
 dos.writeLong(fileItem[i].getStartPos());
 dos.writeLong(fileItem[i].getEndPos());
 //LogUtils.info("[" + fileItem[i].getStartPos() + "#" + fileItem[i].getEndPos() + "]");
 }
 dos.close();
 }
 
 /**
 * function:读取写入点的位置信息
 * @author hoojo
 * @createDate 2011-9-23 下午05:30:29
 * @throws IOException
 */
 private void readPosInfo() throws IOException {
 DataInputStream dis = new DataInputStream(new FileInputStream(tempFile));
 int startPosLength = dis.readInt();
 startPos = new long[startPosLength];
 endPos = new long[startPosLength];
 for (int i = 0; i < startPosLength; i++) {
 startPos[i] = dis.readLong();
 endPos[i] = dis.readLong();
 }
 dis.close();
 }
 
 /**
 * function: 获取下载文件的长度
 * @author hoojo
 * @createDate 2011-9-26 下午12:15:08
 * @return
 */
 private int getFileSize() {
 int fileLength = -1;
 try {
 URL url = new URL(this.downloadInfo.getUrl());
 HttpURLConnection conn = (HttpURLConnection) url.openConnection();
 
 DownloadFile.setHeader(conn);
 
 int stateCode = conn.getResponseCode();
 //判断http status是否为HTTP/1.1 206 Partial Content或者200 OK
 if (stateCode != HttpURLConnection.HTTP_OK && stateCode != HttpURLConnection.HTTP_PARTIAL) {
 LogUtils.log("Error Code: " + stateCode);
 return -2;
 } else if (stateCode >= 400) {
 LogUtils.log("Error Code: " + stateCode);
 return -2;
 } else {
 //获取长度
 fileLength = conn.getContentLength();
 LogUtils.log("FileLength: " + fileLength);
 }
 
 //读取文件长度
 /*for (int i = 1; ; i++) {
 String header = conn.getHeaderFieldKey(i);
 if (header != null) {
 if ("Content-Length".equals(header)) {
 fileLength = Integer.parseInt(conn.getHeaderField(i));
 break;
 }
 } else {
 break;
 }
 }
 */
 
 DownloadFile.printHeader(conn);
 } catch (MalformedURLException e) {
 e.printStackTrace();
 } catch (IOException e) {
 e.printStackTrace();
 }
 return fileLength;
 }
 }

这个类主要是完成读取指定url资源的内容，获取该资源的长度。然后将该资源分成指定的块数，将每块的起始下载位置、结束下载位置，分别保存在一个数组中。每块都单独开辟一个独立线程开始下载。在开始下载之前，需要创建一个临时文件，写入当前下载线程的开始下载指针位置和结束下载指针位置。

# 五、工具类、测试类 #

日志工具类

package com.hoo.util;
 
 /**
 * function: 日志工具类
 * @author hoojo
 * @createDate 2011-9-21 下午05:21:27
 * @file LogUtils.java
 * @package com.hoo.util
 * @project MultiThreadDownLoad
 * @blog http://blog.csdn.net/IBM_hoojo
 * @email hoojo_@126.com
 * @version 1.0
 */
 public abstract class LogUtils {
 
 public static void log(Object message) {
 System.err.println(message);
 }
 
 public static void log(String message) {
 System.err.println(message);
 }
 
 public static void log(int message) {
 System.err.println(message);
 }
 
 public static void info(Object message) {
 System.out.println(message);
 }
 
 public static void info(String message) {
 System.out.println(message);
 }
 
 public static void info(int message) {
 System.out.println(message);
 }
 }
 
 
 
 下载工具类
 
 
 package com.hoo.util;
 
 import com.hoo.download.BatchDownloadFile;
 import com.hoo.entity.DownloadInfo;
 
 /**
 * function: 分块多线程下载工具类
 * @author hoojo
 * @createDate 2011-9-28 下午05:22:18
 * @file DownloadUtils.java
 * @package com.hoo.util
 * @project MultiThreadDownLoad
 * @blog http://blog.csdn.net/IBM_hoojo
 * @email hoojo_@126.com
 * @version 1.0
 */
 public abstract class DownloadUtils {
 
 public static void download(String url) {
 DownloadInfo bean = new DownloadInfo(url);
 LogUtils.info(bean);
 BatchDownloadFile down = new BatchDownloadFile(bean);
 new Thread(down).start();
 }
 
 public static void download(String url, int threadNum) {
 DownloadInfo bean = new DownloadInfo(url, threadNum);
 LogUtils.info(bean);
 BatchDownloadFile down = new BatchDownloadFile(bean);
 new Thread(down).start();
 }
 
 public static void download(String url, String fileName, String filePath, int threadNum) {
 DownloadInfo bean = new DownloadInfo(url, fileName, filePath, threadNum);
 LogUtils.info(bean);
 BatchDownloadFile down = new BatchDownloadFile(bean);
 new Thread(down).start();
 }
 }
 
 
 
 下载测试类
 
 
 package com.hoo.test;
 
 import com.hoo.util.DownloadUtils;
 
 /**
 * function: 下载测试
 * @author hoojo
 * @createDate 2011-9-23 下午05:49:46
 * @file TestDownloadMain.java
 * @package com.hoo.download
 * @project MultiThreadDownLoad
 * @blog http://blog.csdn.net/IBM_hoojo
 * @email hoojo_@126.com
 * @version 1.0
 */
 public class TestDownloadMain {
 
 public static void main(String[] args) {
 /*DownloadInfo bean = new DownloadInfo("http://i7.meishichina.com/Health/UploadFiles/201109/2011092116224363.jpg");
 System.out.println(bean);
 BatchDownloadFile down = new BatchDownloadFile(bean);
 new Thread(down).start();*/
 
 //DownloadUtils.download("http://i7.meishichina.com/Health/UploadFiles/201109/2011092116224363.jpg");
 DownloadUtils.download("http://mp3.baidu.com/j?j=2&url=http%3A%2F%2Fzhangmenshiting2.baidu.com%2Fdata%2Fmusic%2F1669425%2F%25E9%2599%25B7%25E5%2585%25A5%25E7%2588%25B1%25E9%2587%258C%25E9%259D%25A2.mp3%3Fxcode%3D2ff36fb70737c816553396c56deab3f1", "aa.mp3", "c:/temp", 5);
 }
 }

多线程下载主要在第三部和第四部，其他的地方还是很好理解。源码中提供相应的注释了，便于理解。