JAVA 多线程爬虫实例详解

import java.util.ArrayList;
import java.util.List;
import java.util.concurrent.ExecutorService;
import java.util.concurrent.Executors;
import java.util.concurrent.locks.Lock;
import java.util.concurrent.locks.ReentrantLock;

/*
 * 获取京东评论url列表
 */

public class MyThreading {
  private static String p_id = null;
  private static Url urls = null;

  public MyThreading(String p_id){
    this.p_id = p_id ;   // 京东商品的id
    urls = new Url(p_id); 
  }

  public List<String> getUriList(){
    ExecutorService executor = Executors.newCachedThreadPool();
    for (int i = 0 ; i < 600 ; i ++){
      executor.execute(new AddUrl(i));    // 添加任务到线程池
    }
    executor.shutdown();
    while (!executor.isTerminated()){}
    return urls.getList();
  }

  public static class AddUrl implements Runnable{
    int page;
    public AddUrl(int page){
      this.page = page;
    }
    public void run(){
      urls.addList(page);   // 启动多线程任务
    }
  }

  public static class Url {

    private static Lock lock = new ReentrantLock();     // 开启显式家锁
    private static List<String> urlList = new ArrayList();     
    private String p_id;

    public Url(String p_id ){
      this.p_id = p_id ;
    }
    public List<String> getList(){
      return urlList;
    }
    public void addList(int page){
      lock.lock();
      try{
        String url = "http://club.jd.com/productpage/p-" + p_id + "-s-0-t-0-p-" + String.valueOf(page) + ".html";
//       Thread.sleep(5);
        urlList.add(url);    //添加url到url列表
      }catch(Exception ex ){
      }
      finally {
        lock.unlock();     // 解锁
      }

    }
  }
  public static void main(String[] args) {
    String p_id = "2441288";
    MyThreading myThreading = new MyThreading(p_id);
    List <String> urlList = myThreading.getUriList();
    for(String url : urlList){
      System.out.println(url);
    }
    System.out.println(urlList.size());
  }
}

import java.io.BufferedReader;
import java.io.InputStreamReader;
import java.net.URL;
import java.net.URLConnection;
import java.util.ArrayList;
import java.util.List;
import java.util.concurrent.ExecutorService;
import java.util.concurrent.Executors;
import java.util.concurrent.locks.Lock;
import java.util.concurrent.locks.ReentrantLock;
import java.util.regex.Matcher;
import java.util.regex.Pattern;


public class ThreadingCrawel {
  private static Content content = null;
  private static List<String> urlList = null;

  public ThreadingCrawel(List<String> urlList){
    this.urlList = urlList;
    content = new Content();
  }
  public List<String> getContent(){

    ExecutorService executor = Executors.newCachedThreadPool();
    for (String url : urlList){
      executor.execute(new AddContent(url));
    }
    executor.shutdown();
    while(!executor.isTerminated()){}
    return content.getContent();

  }

  public static class AddContent implements Runnable{
    String url;
    public AddContent(String url){
      this.url = url;
    }
    public void run(){
      content.addContent(url);
    }
  }

  public static class Content {

    private static Lock lock = new ReentrantLock();
    private static List<String> contentList = new ArrayList();

    public void addContent(String url){

      String content = "";
      BufferedReader in = null;
      try{
        URL realUrl = new URL(url);
        URLConnection connection = realUrl.openConnection();
        in = new BufferedReader(new InputStreamReader(connection.getInputStream(), "gbk"));
        String line;
        while( (line = in.readLine()) != null){
          content += line +"\n";
        }
      }catch(Exception e){
        e.printStackTrace();
      }
      finally{
        try{
          if (in != null){
            in.close();
          }
        }catch(Exception e2){
          e2.printStackTrace();
        }
      }

      Pattern p = Pattern.compile("content\":\".*?\"");
      Matcher match = p.matcher(content);
      String tmp;
       lock.lock();
      while(match.find()){
        tmp = match.group();
        tmp = tmp.replaceAll("\"", "");
        tmp = tmp.replace("content:", "");
        tmp = tmp.replaceAll("<.*?>", "");
        contentList.add(tmp);
        try {
          Thread.sleep(1);
        } catch (InterruptedException e) {
          // TODO Auto-generated catch block
          e.printStackTrace();
        }
      }
      lock.unlock();

    }
    public List getContent(){
      return contentList;
    }
  }
  public static void main(String[] args){
    long start = System.currentTimeMillis();
    String p_id = "2441288";
    MyThreading myThreading = new MyThreading(p_id);
    List <String> urlList = myThreading.getUriList();
    ThreadingCrawel threadingCrawel = new ThreadingCrawel(urlList);
    List <String> contentList = threadingCrawel.getContent();
    for(String content : contentList){
      System.out.println(content);
    }
    long end = System.currentTimeMillis();
    System.out.println(end - start);
  }
}