package com.commtouch.drools;
import java.net.*;
import java.util.regex.Matcher;
import java.util.regex.Pattern;
public class Facts {
// return a file type of a url
public static class FileType {
private String type = "html";
public FileType (URL url) {
// get the path from the url
//TODO: use parser to do it?
String path = url.getPath();
if (path != null) {
// define regex to extract file from path
Pattern fileRegex = Pattern.compile("([^/]+)$");
Matcher fileMatcher = fileRegex.matcher(path);
// get regex capure $1 (filename)
String file = null;
while (fileMatcher.find())
file = fileMatcher.group(1);
String suffix = null;
if (file != null)
{
//try to extract suffix from file
Pattern suffixRegex = Pattern.compile("\\.([^\\.]+)$");
Matcher suffixMatcher = suffixRegex.matcher(file);
while (suffixMatcher.find())
suffix = suffixMatcher.group(1);
//verify that the suffix is a valid suffix
if (suffix != null && (suffix.length()> 1 && suffix.length() < 6))
setType(suffix);
}
}
}
public void setType(String type) {
this.type = type;
}
public String getType() {
return type;
}
}
}