InputNormalizer.java

/*
 * This program is free software; you can redistribute it and/or modify it under
 * the terms of the GNU General Public License as published by the Free Software
 * Foundation; either version 2 of the License, or (at your option) any later
 * version. You should have received a copy of the GNU General Public License
 * along with this program; if not, write to the Free Software Foundation, Inc.,
 * 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
 */

package org.aitools.programd.util;

import java.util.ArrayList;
import java.util.List;
import java.util.regex.Matcher;
import java.util.regex.Pattern;

import org.aitools.util.Lists;
import org.aitools.util.xml.Characters;
import org.jdom.Text;

/**
 * <code>InputNormalizer</code> replaces <code>Substituter</code> as the utility class for performing various stages of
 * <a href="http://aitools.org/aiml/TR/2001/WD-aiml/#section-input-normalization">input normalization </a>.
 * Substitutions of other types are now handled independently by their respective processors.
 * 
 */
public class InputNormalizer {

  /** A regex to match characters that are not legal in AIML patterns. */
  private static final java.util.regex.Pattern ILLEGAL_CHARACTERS = java.util.regex.Pattern
      .compile("[^\\p{javaUpperCase}\\p{Digit} \\*_]+");

  /** A regex to match characters that are not legal in AIML patterns, ignoring case. */
  private static final java.util.regex.Pattern ILLEGAL_CHARACTERS_IGNORE_CASE = java.util.regex.Pattern
      .compile("[^\\p{javaLetter}\\p{Digit} \\*_]+");

  /**
   * <p>
   * Performs <a href="http://aitools.org/aiml/TR/2001/WD-aiml/#section-pattern-fitting-normalizations">pattern-fitting
   * </a> normalization on an input.
   * </p>
   * <p>
   * This is best used to produce a caseless representation of the effective match string when presenting match results
   * to a user; however, if used when actually performing the match it will result in the case of wildcard-captured
   * values being lost. Some amendment of the specification is probably in order.
   * </p>
   * 
   * @param input the string to pattern-fit
   * @return the pattern-fitted input
   */
  public static String patternFit(String input) {
    return Text.normalizeString(ILLEGAL_CHARACTERS.matcher(Characters.removeMarkup(input)).replaceAll(" ")).trim();
  }

  /**
   * <p>
   * Performs a partial <a
   * href="http://aitools.org/aiml/TR/2001/WD-aiml/#section-pattern-fitting-normalizations">pattern-fitting </a>
   * normalization on an input -- partial because it does <i>not </i> convert letters to uppercase.
   * </p>
   * <p>
   * This is used when sending patterns to the Graphmaster so that wildcard contents can be captured and case retained.
   * Some amendment of the specification is probably in order.
   * </p>
   * 
   * @param input the string to pattern-fit
   * @return the pattern-fitted input
   */
  public static String patternFitIgnoreCase(String input) {
    return Text.normalizeString(ILLEGAL_CHARACTERS_IGNORE_CASE.matcher(Characters.removeMarkup(input)).replaceAll(" "))
        .trim();
  }

  /**
   * Splits an input into sentences, as defined by the <code>sentenceSplitters</code>.
   * 
   * @param sentenceSplitters the sentence splitters to use
   * @param input the input to split
   * @return the input split into sentences
   */
  public static List<String> sentenceSplit(Pattern sentenceSplitters, String input) {
    List<String> result = new ArrayList<String>();
    Matcher matcher = sentenceSplitters.matcher(input);
    if (matcher.find()) {
      do {
        result.add(input.substring(matcher.start(), matcher.end()));
      } while (matcher.find());
      if (!matcher.hitEnd()) {
        result.add(input.substring(matcher.end()));
      }
      return result;
    }
    // else
    return Lists.singleItem(input);
  }
}