【Android开发经验】关于“多线程断点续传下载”功能的一个简单实现和讲解

    转载请注明出处：http://blog.csdn.net/zhaokaiqiang1992
    上班第一天，在技术群里面和大家闲扯，无意中谈到了关于框架的使用，一个同学说为了用xUtils的断线续传下载功能，把整个库引入到了项目中，在google的官方建议中，是非常不建议这种做法的，集合框架虽然把很多功能集成起来，但是代码越多，出现问题的可能越大，而且无形之中增加了APK的大小，因此，得不偿失。所以，这篇文章主要就“断线续传”下载功能，简单的说下思路和代码实现，因为这类代码比较多，所以找了一个写的不错的demo，简单优化了一下。
    在贴代码之前，我们先分析一下需求和解决思路。首先是下载功能，我们简单的使用HttpURLConnection就可以了，没有引入框架的必要，然后就是断点续传了，其实断点续传指的就是我们可以随时停止我们的下载任务，当下次再次开始的时候，可以从上次下载到的位置继续下载，节省下载时间，很方便也很实用，做法无非就是在下载的过程中，纪录下下载到的位置，当再次开始下载的时候，我们从上一次的位置继续请求服务器即可。说到这里，有个类不得不提，那就是RandomAccessFile，这个类是实现断点续传功能的核心类，RandomAccessFile允许我们从我们想要的位置进行读写操作，因此，我们可以把我们要下载的文件切分成几部分，然后开启多个线程，分别从文件不同的位置进行下载，这样等所有的部分都下载完成之后，我们就能够得到一个完整的文件了，这就是多线程下载的原理，完成上面几个步骤，我们的多线程断线续传下载功能就基本完成了，下面是在网上找的一个Demo，我对代码进行了部分修改，从代码里面，我们看一下如何进行代码的实现。
    首先，如果要实现断点续传，我们就要纪录每个线程下载的文件的位置，可以使用文件，也可以使用sp，也可以使用DB，这个Demo里面使用的DB，我们首先看一下数据库的Helper实现类，里面存储主键、线程号、开始位置、结束位置、完成位置和下载地址即可。
    DownLoadHelper.java
public class DownLoadHelper extends SQLiteOpenHelper {

private static final String SQL_NAME = "download.db";
private static final int DOWNLOAD_VERSION = 1;

public DownLoadHelper(Context context) {
super(context, SQL_NAME, null, DOWNLOAD_VERSION);
}

/**
* 在download.db数据库下创建一个download_info表存储下载信息
*/
@Override
public void onCreate(SQLiteDatabase db) {
db.execSQL("create table download_info(_id integer PRIMARY KEY AUTOINCREMENT, thread_id integer, "
+ "start_pos integer, end_pos integer, compelete_size integer,url char)");
}

@Override
public void onUpgrade(SQLiteDatabase db, int oldVersion, int newVersion) {

}

}

    有了Helper之后，我们在创建一个Sql工具类，完成对表的数据操作
    DownlaodSqlTool.java
public class DownlaodSqlTool {

private DownLoadHelper dbHelper;

public DownlaodSqlTool(Context context) {
dbHelper = new DownLoadHelper(context);
}

/**
* 创建下载的具体信息
*/
public void insertInfos(List<DownloadInfo> infos) {
SQLiteDatabase database = dbHelper.getWritableDatabase();
for (DownloadInfo info : infos) {
String sql = "insert into download_info(thread_id,start_pos, end_pos,compelete_size,url) values (?,?,?,?,?)";
Object[] bindArgs = { info.getThreadId(), info.getStartPos(),
info.getEndPos(), info.getCompeleteSize(), info.getUrl() };
database.execSQL(sql, bindArgs);
}
}

/**
* 得到下载具体信息
*/
public List<DownloadInfo> getInfos(String urlstr) {
List<DownloadInfo> list = new ArrayList<DownloadInfo>();
SQLiteDatabase database = dbHelper.getWritableDatabase();
String sql = "select thread_id, start_pos, end_pos,compelete_size,url from download_info where url=?";
Cursor cursor = database.rawQuery(sql, new String[] { urlstr });
while (cursor.moveToNext()) {
DownloadInfo info = new DownloadInfo(cursor.getInt(0),
cursor.getInt(1), cursor.getInt(2), cursor.getInt(3),
cursor.getString(4));
list.add(info);
}
return list;
}

/**
* 更新数据库中的下载信息
*/
public void updataInfos(int threadId, int compeleteSize, String urlstr) {
SQLiteDatabase database = dbHelper.getWritableDatabase();
String sql = "update download_info set compelete_size=? where thread_id=? and url=?";
Object[] bindArgs = { compeleteSize, threadId, urlstr };
database.execSQL(sql, bindArgs);
}

/**
* 关闭数据库
*/
public void closeDb() {
dbHelper.close();
}

/**
* 下载完成后删除数据库中的数据
*/
public void delete(String url) {
SQLiteDatabase database = dbHelper.getWritableDatabase();
database.delete("download_info", "url=?", new String[] { url });
}
}
    数据库相关的类就这些，断点续传的功能已经完成，下面看下载如何实现。
    首先，为了操作方便，我们对下载的文件抽取实体类
    DownloadInfo.java
public class DownloadInfo {

private int threadId;// 下载器id
private int startPos;// 开始点
private int endPos;// 结束点
private int compeleteSize;// 完成度
private String url;// 下载文件的URL地址

public DownloadInfo(int threadId, int startPos, int endPos,
int compeleteSize, String url) {
this.threadId = threadId;
this.startPos = startPos;
this.endPos = endPos;
this.compeleteSize = compeleteSize;
this.url = url;
}

public DownloadInfo() {
}

public String getUrl() {
return url;
}

public void setUrl(String url) {
this.url = url;
}

public int getThreadId() {
return threadId;
}

public void setThreadId(int threadId) {
this.threadId = threadId;
}

public int getStartPos() {
return startPos;
}

public void setStartPos(int startPos) {
this.startPos = startPos;
}

public int getEndPos() {
return endPos;
}

public void setEndPos(int endPos) {
this.endPos = endPos;
}

public int getCompeleteSize() {
return compeleteSize;
}

public void setCompeleteSize(int compeleteSize) {
this.compeleteSize = compeleteSize;
}

@Override
public String toString() {
return "DownloadInfo [threadId=" + threadId + ", startPos=" + startPos
+ ", endPos=" + endPos + ", compeleteSize=" + compeleteSize
+ "]";
}
}

    实体类抽取之后，我们就可以实现下载功能了，DownloadHttpTool是实现下载功能的主要类
    DownloadHttpTool.java
public class DownloadHttpTool {

private static final String TAG = DownloadHttpTool.class.getSimpleName();
// 线程数量
private int threadCount;
// URL地址
private String urlstr;
private Context mContext;
private Handler mHandler;
// 保存下载信息的类
private List<DownloadInfo> downloadInfos;
// 目录
private String localPath;
// 文件名
private String fileName;
private int fileSize;
// 文件信息保存的数据库操作类
private DownlaodSqlTool sqlTool;

// 利用枚举表示下载的三种状态
private enum Download_State {
Downloading, Pause, Ready, Delete;
}

// 当前下载状态
private Download_State state = Download_State.Ready;
// 所有线程下载的总数
private int globalCompelete = 0;

public DownloadHttpTool(int threadCount, String urlString,
String localPath, String fileName, Context context, Handler handler) {
super();
this.threadCount = threadCount;
this.urlstr = urlString;
this.localPath = localPath;
this.mContext = context;
this.mHandler = handler;
this.fileName = fileName;
sqlTool = new DownlaodSqlTool(mContext);
}

// 在开始下载之前需要调用ready方法进行配置
public void ready() {
Log.w(TAG, "ready");
globalCompelete = 0;
downloadInfos = sqlTool.getInfos(urlstr);
if (downloadInfos.size() == 0) {
initFirst();
} else {
File file = new File(localPath + "/" + fileName);
if (!file.exists()) {
sqlTool.delete(urlstr);
initFirst();
} else {
fileSize = downloadInfos.get(downloadInfos.size() - 1)
.getEndPos();
for (DownloadInfo info : downloadInfos) {
globalCompelete += info.getCompeleteSize();
}
Log.w(TAG, "globalCompelete:::" + globalCompelete);
}
}
}

public void start() {
Log.w(TAG, "start");
if (downloadInfos != null) {
if (state == Download_State.Downloading) {
return;
}
state = Download_State.Downloading;
for (DownloadInfo info : downloadInfos) {
Log.v(TAG, "startThread");
new DownloadThread(info.getThreadId(), info.getStartPos(),
info.getEndPos(), info.getCompeleteSize(),
info.getUrl()).start();
}
}
}

public void pause() {
state = Download_State.Pause;
sqlTool.closeDb();
}

public void delete() {
state = Download_State.Delete;
compelete();
new File(localPath + File.separator + fileName).delete();
}

public void compelete() {
sqlTool.delete(urlstr);
sqlTool.closeDb();
}

public int getFileSize() {
return fileSize;
}

public int getCompeleteSize() {
return globalCompelete;
}

/**
* 第一次下载初始化
*/
private void initFirst() {
Log.w(TAG, "initFirst");
try {
URL url = new URL(urlstr);
HttpURLConnection connection = (HttpURLConnection) url
.openConnection();
connection.setConnectTimeout(5000);
connection.setRequestMethod("GET");
fileSize = connection.getContentLength();
Log.w(TAG, "fileSize::" + fileSize);
File fileParent = new File(localPath);
if (!fileParent.exists()) {
fileParent.mkdir();
}
File file = new File(fileParent, fileName);
if (!file.exists()) {
file.createNewFile();
}
// 本地访问文件
RandomAccessFile accessFile = new RandomAccessFile(file, "rwd");
accessFile.setLength(fileSize);
accessFile.close();
connection.disconnect();
} catch (Exception e) {
e.printStackTrace();
}
int range = fileSize / threadCount;
downloadInfos = new ArrayList<DownloadInfo>();
for (int i = 0; i < threadCount - 1; i++) {
DownloadInfo info = new DownloadInfo(i, i * range, (i + 1) * range
- 1, 0, urlstr);
downloadInfos.add(info);
}
DownloadInfo info = new DownloadInfo(threadCount - 1, (threadCount - 1)
* range, fileSize - 1, 0, urlstr);
downloadInfos.add(info);
sqlTool.insertInfos(downloadInfos);
}

/**
* 自定义下载线程
*
* @author zhaokaiqiang
* @time 2015-2-25下午5:52:28
*/
private class DownloadThread extends Thread {

private int threadId;
private int startPos;
private int endPos;
private int compeleteSize;
private String urlstr;
private int totalThreadSize;

public DownloadThread(int threadId, int startPos, int endPos,
int compeleteSize, String urlstr) {
this.threadId = threadId;
this.startPos = startPos;
this.endPos = endPos;
totalThreadSize = endPos - startPos + 1;
this.urlstr = urlstr;
this.compeleteSize = compeleteSize;
}

@Override
public void run() {
HttpURLConnection connection = null;
RandomAccessFile randomAccessFile = null;
InputStream is = null;
try {
randomAccessFile = new RandomAccessFile(localPath
+ File.separator + fileName, "rwd");
randomAccessFile.seek(startPos + compeleteSize);
URL url = new URL(urlstr);
connection = (HttpURLConnection) url.openConnection();
connection.setConnectTimeout(5000);
connection.setRequestMethod("GET");
connection.setRequestProperty("Range", "bytes="
+ (startPos + compeleteSize) + "-" + endPos);
is = connection.getInputStream();
byte[] buffer = new byte[1024];
int length = -1;
while ((length = is.read(buffer)) != -1) {
randomAccessFile.write(buffer, 0, length);
compeleteSize += length;
Message message = Message.obtain();
message.what = threadId;
message.obj = urlstr;
message.arg1 = length;
mHandler.sendMessage(message);
Log.w(TAG, "Threadid::" + threadId + " compelete::"
+ compeleteSize + " total::" + totalThreadSize);
// 当程序不再是下载状态的时候，纪录当前的下载进度
if ((state != Download_State.Downloading)
|| (compeleteSize >= totalThreadSize)) {
sqlTool.updataInfos(threadId, compeleteSize, urlstr);
break;
}
}

} catch (Exception e) {
e.printStackTrace();
sqlTool.updataInfos(threadId, compeleteSize, urlstr);
} finally {
try {
if (is != null) {
is.close();
}
randomAccessFile.close();
connection.disconnect();
} catch (Exception e) {
e.printStackTrace();
}
}
}
}
}
    在上面的代码中，我们定义了一个线程，进行文件的多线程下载，并且在退出下载状态和完成下载的时候，纪录下载的位置，存到数据库中。在原来的代码中，是没获取一次数据，存取一次数据库，大大的增加了数据库的操作频率，降低了效率，原先下载420k左右的文件，需要操作420次数据库，现在只要一次即可。    在initFirst()里面，首先进行了初始化，根据下载文件的大小和开启线程的数量，对下载实体类进行了初始化和赋值。在RandomAccessFile创建完毕，DownloadInfo初始化完毕之后，就可以通过start()进行文件的下载了。
    其实到这里，基本的功能已经实现了。为了使得我们的操作更加的方便，同时可以监控到下载的进度，我们对下载类进行一次封装，代码如下：
    DownloadUtil.javapublic class DownloadUtil {

private DownloadHttpTool mDownloadHttpTool;
private OnDownloadListener onDownloadListener;

private int fileSize;
private int downloadedSize = 0;

@SuppressLint("HandlerLeak")
private Handler mHandler = new Handler() {

@Override
public void handleMessage(Message msg) {
int length = msg.arg1;
synchronized (this) {// 加锁保证已下载的正确性
downloadedSize += length;

}
if (onDownloadListener != null) {
onDownloadListener.downloadProgress(downloadedSize);
}
if (downloadedSize >= fileSize) {
mDownloadHttpTool.compelete();
if (onDownloadListener != null) {
onDownloadListener.downloadEnd();
}
}
}

};

public DownloadUtil(int threadCount, String filePath, String filename,
String urlString, Context context) {

mDownloadHttpTool = new DownloadHttpTool(threadCount, urlString,
filePath, filename, context, mHandler);
}

// 下载之前首先异步线程调用ready方法获得文件大小信息，之后调用开始方法
public void start() {
new AsyncTask<Void, Void, Void>() {

@Override
protected Void doInBackground(Void... arg0) {
mDownloadHttpTool.ready();
return null;
}

@Override
protected void onPostExecute(Void result) {
fileSize = mDownloadHttpTool.getFileSize();
downloadedSize = mDownloadHttpTool.getCompeleteSize();
Log.w("Tag", "downloadedSize::" + downloadedSize);
if (onDownloadListener != null) {
onDownloadListener.downloadStart(fileSize);
}
mDownloadHttpTool.start();
}
}.execute();
}

public void pause() {
mDownloadHttpTool.pause();
}

public void delete() {
mDownloadHttpTool.delete();
}

public void reset() {
mDownloadHttpTool.delete();
start();
}

public void setOnDownloadListener(OnDownloadListener onDownloadListener) {
this.onDownloadListener = onDownloadListener;
}

// 下载回调接口
public interface OnDownloadListener {
public void downloadStart(int fileSize);

public void downloadProgress(int downloadedSize);

public void downloadEnd();
}
}
      通过对外暴露接口，我们可以实现下载进度的监听了！    用的时候也很简单，像下面这样就ok了
String urlString = "http://bbra.cn/Uploadfiles/imgs/20110303/fengjin/013.jpg";
final String localPath = Environment.getExternalStorageDirectory()
.getAbsolutePath() + "/ADownLoadTest";
mDownloadUtil = new DownloadUtil(2, localPath, "abc.jpg", urlString,
this);
mDownloadUtil.setOnDownloadListener(new OnDownloadListener() {

@Override
public void downloadStart(int fileSize) {
max = fileSize;
mProgressBar.setMax(fileSize);
}

@Override
public void downloadProgress(int downloadedSize) {
mProgressBar.setProgress(downloadedSize);
total.setText((int) downloadedSize * 100 / max + "%");
}

@Override
public void downloadEnd() {
Bitmap bitmap = decodeSampledBitmapFromResource(localPath
+ File.separator + "abc.jpg", 200, 200);
image.setImageBitmap(bitmap);
}
});
    GitHub的Demo地址：https://github.com/ZhaoKaiQiang/MultiThreadDownLoadDemo