urlmap.java

来自「一个简单的在互联网上抓包的程序」· Java 代码 · 共 19 行

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

public class UrlMap extends HashMap {
  Vector urlList = new Vector();
  String urlFileName;
  RandomAccessFile urlFile;
  
  
  public static void main(String[] args) throws Exception {
    String path = url2path("http://www.kmit.edu.tw/ccc/Carwler.htm");
    System.out.println(path);
    System.out.println("normalize:"+normalize("http://www.kmit.edu.tw/ccc/Carwler.htm"));
    System.out.println(fullUrl("http://www.kmit.edu.tw/abc/def/ghi.htm", "123/456.htm"));
    UrlData[] urls = html2urls(STR.file2text("test1.htm"));
    System.out.println(Arrays.asList(urls));
/*    UrlMap urlMap = new UrlMap("urlMap.txt");
    urlMap.add("http://www.kmit.edu.tw/ccc/Carwler.htm", "虏