2-1-11

package task;

public class EntityBean {
    //成员变量列表
    private String name;
    private String tsxx;
    private String pf;
    private String plrs;
    
    //get()方法和set()方法
    public String getName() {
        return name;
    }
    public void setName(String name) {
        this.name = name;
    }
    public String getTsxx() {
        return tsxx;
    }
    public void setTsxx(String tsxx) {
        this.tsxx = tsxx;
    }
    public String getPf() {
        return pf;
    }
    public void setPf(String pf) {
        this.pf = pf;
    }
    public String getPlrs() {
        return plrs;
    }
    public void setPlrs(String plrs) {
        this.plrs = plrs;
    }
    
    //有参构造方法
    public EntityBean(String name, String tsxx, String pf, String plrs) {
        super();
        this.name = name;
        this.tsxx = tsxx;
        this.pf = pf;
        this.plrs = plrs;
    }
    
    //无参构造方法
    public EntityBean() {
        super();
    }
    
    //重写toString()方法
    @Override
    public String toString() {
        return "EntityBean [name=" + name + ", tsxx=" + tsxx + ", pf=" + pf + ", plrs=" + plrs + "]";
    }
    
    
}

package task;

import java.io.File;
import java.io.FileNotFoundException;
import java.io.IOException;
import java.io.PrintWriter;
import java.util.ArrayList;

import org.jsoup.Jsoup;
import org.jsoup.nodes.Document;
import org.jsoup.nodes.Element;
import org.jsoup.select.Elements;

public class Crawler {
    
    //1.成员变量
    private static String outputfile = "output\\豆瓣读书数据.txt";
    private static String sourcefile = "resource\\豆瓣读书 Top 250.htm";
    
    
    //2.解析本地网页，返回Document
    public static Document getHtml() throws IOException {
        File f=new File(sourcefile);  //创建文件对象
        //解析本地文件，第一个参数是文件对象
        //第二个参数是网页的字符编码格式
        Document doc=Jsoup.parse(f,"utf-8");  
        return doc;
    }
    
    //3.提取数据，返回集合
    public static ArrayList<EntityBean> getData(Document doc) throws IOException {
        Elements elements=doc.select("#content > div > div.article > div > table > tbody > tr > td");
        ArrayList<EntityBean> list=new ArrayList<EntityBean>();//创建集合对象
        for(Element e :elements ) {  //10部电影信息
            String name=e.select("div.pl2").text();
            String tsxx=e.select("p.pl").text();
            String pf=e.select("span.rating_nums").text();
            String plrs=e.select("span.pl").text().replace("(","").replace(")", "").replace("评价", "").trim();
            //System.out.println(name+" "+tsxx+" "+pf+" "+plrs);
            EntityBean eb=new EntityBean(name,tsxx,pf,plrs);
            list.add(eb);
        }
        return list;
    }
    
    //4.写出到本地文件系统
    public static void wirteToFile(ArrayList<EntityBean>  list) throws FileNotFoundException {
        PrintWriter pw=new PrintWriter(outputfile);//创建字节打印流对象
        for(EntityBean eb : list) {  //10
            pw.println(eb.getName()+"\t"+eb.getTsxx()+"\t"+eb.getPf()+
                    "\t"+eb.getPlrs());
        }
        pw.close();
    }
    //5.主方法
    public static void main(String[] args) throws IOException {
        Document doc = Crawler.getHtml();//解析网页，得到doc对象
        System.out.println("网页标题："+doc.title());
        ArrayList<EntityBean> entitys = Crawler.getData(doc);  //提取数据
        Crawler.wirteToFile(entitys);
    }
}