urllinkexample.java

java 完全探索的随书源码

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

public class URLLinkExample
{

  // Default Constructor
  public URLLinkExample()
  {
    super();
  }

  // Does the token have a "http:" substirng within it
  private boolean hasMatch( String token )
  {
    return token.indexOf( "http:" ) != -1;
  }

  // Trim the string to something respectful to print out
  private String trimURL( String url )
  {
    String tempStr = null;
    int beginIndex = url.indexOf( "http" );
    int endIndex = url.length();
    tempStr = url.substring( beginIndex, endIndex );
    endIndex = tempStr.indexOf( '"' );
    if ( endIndex == -1 )
      endIndex = tempStr.length();
    return tempStr.substring( 0, endIndex );
  }

  // Go through all of the text returned from a web site and search for links
  public Collection searchURL( String urlString )
  {
    URL url = null;
    URLConnection conn = null;
    String nextLine = null;
    StringTokenizer tokenizer = null;
    Collection urlCollection = new ArrayList();

    try
    {
      // Get a new URL object on the url string passed in
      url = new URL( urlString );
      // open the connection
      conn = url.openConnection();
      conn.setDoOutput( true );
      // Complete the connection
      conn.connect();
      BufferedReader reader = new BufferedReader( new InputStreamReader( conn.getInputStream() ));

      // Go through all the text and check it for being a link to another page
      while( (nextLine = reader.readLine()) != null )
      {
        // Create a tokenizer on each text line
        tokenizer = new StringTokenizer( nextLine );
        while( tokenizer.hasMoreTokens() )
        {
          String urlToken = tokenizer.nextToken();
          // If the token is a link, add it to a collection
          if ( hasMatch( urlToken) )
            urlCollection.add( trimURL( urlToken ) );
        }
      }
    }
    catch( MalformedURLException ex )
    {
      System.out.println( "You did not provide a valid URL" );
      ex.printStackTrace();
    }
    catch( IOException ex )
    {
      ex.printStackTrace();
    }
    return urlCollection;
  }

  public static void main(String[] args)
  {
    if( args.length != 1 )
    {
      System.out.println( "Usage: java URLLinkExample <http://url>" );
      System.exit( -1 );
    }

    // Get the url from the command line arguments
    String url = args[0];

    System.out.println( "Searching web site: " + url );
    URLLinkExample example = new URLLinkExample();
    Collection urlCollection = example.searchURL( url );

    // Print out the candiate links
    Iterator iter = urlCollection.iterator();
    while( iter.hasNext() )
    {
      System.out.println( iter.next() );
    }
  }
}