抓出soso音乐信息

import java.util.ArrayList;
import java.util.HashSet;
import java.util.Iterator;
import java.util.List;

public class SongList {
	//歌手列表
	private List<Song> songList;
	
	//当前页面最大分页数
	private int maxPage = 2;
	
	//是否取得页面最大分页数
	private boolean bool = false;
	
	//已添加的歌曲信息
	private HashSet<String> sondListed = new HashSet<String>();
	
	public SongList(String singer){
		this.songList = spiderBySinger(singer);
	}
	
	public List<Song> getSongList(){
		return songList;
	}

	/**
	 * 根据歌手名称去爬页面,取得歌曲信息
	 */
	private List<Song> spiderBySinger(String singer){
		ArrayList<Song> sosoList = new ArrayList<Song>();
		Spider spider = new Spider();
		for (int i = 1; i < maxPage ;i++){
			String content = spider.getURLContentByUrl("http://music.soso.com/music.cgi?w=" + singer + "&pl=&co=&ch=s.m.res&sc=mus&sz=&clz=wma&pg="+i);
			setMaxPageByContent(content);
			sosoList.addAll(this.getSongListByContent(content));
		}
		sondListed.clear();
		return sosoList;
	}
	
	/**
	 * 根据内容设置当前页面最大分页数
	 */
	private void setMaxPageByContent(String content){
		if (!bool){
			String s = ContentPattern.getStringByContentPattern(content, "<script language=\"javascript\">splitResNum(.*?);</script>");
			if (s != null && !s.equals("")){
				s = Replace.YYReplace(s, "(", "");
				s = Replace.YYReplace(s, ")", "");
				int i = Integer.parseInt(s)/20 + 1;
				if (i > 50){
					maxPage = 50;
				}else{
					maxPage = i;
				}
			}
			bool = true;
		} 
	}

	/**
	 * 获得歌曲列表
	 */
	private ArrayList<Song> getSongListByContent(String content){
		ArrayList<Song> list = new ArrayList<Song>();
		Iterator<String> iterator = ContentPattern.getListByContentPattern(content, "<form id=(.*?)<tr>").iterator();
		while(iterator.hasNext()){
			Song song = getSongByContent(iterator.next());
			if (song.getName() == null){
			}else{
				list.add(song);
			}
		}
		return list;
	}

	/**
	 * 根据页面内容提取歌曲信息
	 */
	private Song getSongByContent(String content){
		Song song = new Song();
		//获得歌曲名称
		String name = ContentPattern.getStringByContentPattern(content, "name=\"song\"\\s+value=\"(.*?)\">");
		//判断是否已添加此歌曲
		if(sondListed.contains(name.trim())){
			return song;
		}else{
			sondListed.add(name.trim());
			//获得歌曲文件地址
			String address = ContentPattern.getStringByContentPattern(content, "name=\"url\"\\s+value=\"(.*?)\">");
			//获得歌手
			String singer = ContentPattern.getStringByContentPattern(content, "name=\"singer\"\\s+value=\"(.*?)\">");
			//获得专辑
			String special = ContentPattern.getStringByContentPattern(content, "name=\"album\"\\s+value=\"(.*?)\">");
			song.setAddress(address);
			song.setName(name);
			song.setSpecial(special);
			song.setSinger(singer);
			return song;
		}
	}
}

import java.util.HashSet;
import java.util.regex.Matcher;
import java.util.regex.Pattern;

public class ContentPattern {
	/**
	 * 获得列表根据原内容和正则表达式
	 */
	public static HashSet<String> getListByContentPattern(String content, String pattern){
		HashSet<String> list = new HashSet<String>();
		// 用正则表达式编译链接的匹配模式。
		Pattern p = Pattern.compile(pattern, Pattern.CASE_INSENSITIVE);
		Matcher m = p.matcher(content);
		while(m.find()){
			String s = m.group(1).trim();
			if(!list.contains(s))
				list.add(s);
		}
		return list;
	}

	/**
	 * 内容,根据内容和正则表达式
	 */
	public static String getStringByContentPattern(String content, String pattern){
		String s = "";
		// 用正则表达式编译链接的匹配模式。
		Pattern p = Pattern.compile(pattern, Pattern.CASE_INSENSITIVE);
		Matcher m = p.matcher(content);
		if(m.find()){
			s = m.group(1).trim();
		}
		return s;
	}
}

public class Song {
	//歌曲名称
	private String name;
	
	//歌手名
	private String singer;
	
	//专辑
	private String special;
	
	//地址
	private String address;

	public String getName() {
		return name;
	}

	public void setName(String name) {
		this.name = name;
	}

	public String getSinger() {
		return singer;
	}

	public void setSinger(String singer) {
		this.singer = singer;
	}

	public String getSpecial() {
		return special;
	}

	public void setSpecial(String special) {
		this.special = special;
	}

	public String getAddress() {
		return address;
	}

	public void setAddress(String address) {
		this.address = address;
	}
}