down.java

利用某网站的VOA听力进行批量下载

/**
* Download the file specified by an URL.
*/
//package fetchInformation;

import java.io.*;
import java.net.*;

/**
* @author caojinghua
*
*/
public class Down {

       public static void LoadFiles(String spec, File save)
      {//获取目标页面
try{
    URL url=new URL(spec);
    BufferedReader in=new BufferedReader(new InputStreamReader(url.openStream()));
    //output to a file
    BufferedWriter out=new BufferedWriter(new FileWriter(save));
    String line=null;
    while((line=in.readLine())!=null)
   {
         out.write(line);
   }
   if(in!=null)
         in.close();
   if(out!=null)
out.close();
}catch(MalformedURLException e)
{
     System.out.println(e);
}catch(IOException ioe)
{
     System.out.println(ioe);
}
      }
/**
* @param args
*/
public static void main(String[] args)throws IOException {
int pagenum=26700;
int flag=0;
File  file = new File( "config.ini" );

if( !(file.exists()))
{	//假如没有配置文件，就创建一个
	DataOutputStream newFw = new DataOutputStream( new FileOutputStream( "config.ini" ));
	newFw.writeInt(26815);
	newFw.close();
	}
//读入上次下载之后的位置，以便继续下载新的mp3
DataInputStream fr = new DataInputStream( new FileInputStream("config.ini"));
pagenum = fr.readInt();
fr.close();
System.out.println( "Download from :" +pagenum );
//file=new File(""+pagenum+".htm");
//while ((file.exists()))
//{
//	pagenum++;
//	file=new File(""+pagenum+".htm");
//	}
pagenum++;	//此时pagenum的值应该是最新的页面号码
//pagenum=26686;
while(1==1)
{
String savepath= "" + pagenum +".htm";
//URL补全
String url="http://www.51voa.com/VOA_Standard_English/VOA_Standard_English_" + pagenum + ".html";
try{
File savefile=new File(savepath);
LoadFiles(url, savefile);
}catch(NullPointerException e){
System.out.println("End");
System.exit(0);
}

try{
	File file1 = new File( savepath );
	BufferedReader reader = new BufferedReader(new FileReader(file1));
String str;
str=reader.readLine();
int index1=0,index2=0;
index1 = str.indexOf("<EMBED src=\"") + 12;
index2 = str.indexOf("\" width=", index1);
if( index1==-1 || index2==-1 )
	throw new IOException("Not a mp3 URL");
String newStr;
newStr="http://www.51voa.com" + str.substring(index1, index2);
//System.out.println(newStr);
System.out.println( "This is the URL:" + newStr );

//这里设定flashget的路径
String exePath="D:\\Program Files\\FlashGet Network\\Flashget\\flashget.exe";
String mp3Path="E:\\Downloads";
try{
	Runtime.getRuntime().exec(exePath+" "+newStr);
}catch(IOException e){
	//System.out.println(e);
	flag++;
}
}// end try
catch(IOException ee)
{
	System.out.println("End while exec"+pagenum);
	System.out.println(ee);
	flag++;
	//System.exit(0);
	}// end catch

pagenum++;
if(flag>10)
{	//这里是设定的用于修复页面号码不连续问题的，假如有10个连续页面不存在，说明后面确实是没有更新的内容了
	//将页面号码写回配置文件，以便下次继续下载新的内容而不是重复下载已经下载过的内容
	BufferedWriter fw = new BufferedWriter( new FileWriter("config.ini"));
	fw.write( "" + (pagenum - 11) );
	fw.close();
	System.out.println( "The last download is " + (pagenum -6) );
	System.exit(0);
}
	

}// end while
//System.out.println("");
//BufferedWriter fw = new BufferedWriter( new FileWriter("config.ini"));
//FileWriter fw = new FileWriter( "config.ini" );
//fw.write( "" + (pagenum - 5) );
//fw.close();
}// end method main 
}// end class Down