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;	}}