CIS 260 FindLinks.java

Code

/*
    FindLinks.java
    David Klick
    CIS 260
    2011-06-21

    Uses regular expressions to find links in web pages.
*/

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

public class FindLinks {
    public static void main(String[] args) {
        Pattern pat = Pattern.compile("<a\\s+href\\s*(\"[^\"]*\"|[\\S]*)\\s*>",
            Pattern.CASE_INSENSITIVE);
        Matcher matcher = null;
        String[] urls = null;
        int ch;
        
        if (args.length == 0) {
            Scanner in = new Scanner(System.in);
            System.out.print("Enter a space separated list of URLs: ");
            urls = in.nextLine().split("\\s+");
            if (urls.length == 0) {
                System.out.println("No URLs to process. Bye.");
                System.exit(0);
            }
        } else urls = args;

        StringBuilder page = new StringBuilder();
        InputStreamReader str = null;
        for (String url : urls) {
            try {
                if (!url.startsWith("http://")) url = "http://" + url;
                str = new InputStreamReader((new URL(url)).openStream());
                page.setLength(0);
                while ((ch = str.read()) != -1) page.append((char)ch);
                matcher = pat.matcher(page);
                System.out.println("\nLinks found in " + url);
                int matchNum = 0;
                while (matcher.find()) {
                    System.out.printf("%3d: %s\n", ++matchNum,
                        page.substring(matcher.start(), matcher.end()));
                }
                if (matchNum == 0) System.out.println("No links found on page.");
            } catch (IOException e) {
                System.out.println("\nError connecting to " + url);
            } finally {
                if (str != null) {
                    try {
                        str.close();
                    } catch (IOException e2) {}
                }
            }
        }
    }
}