edu.cmu.sphinx.trainer.TrainerDictionary Maven / Gradle / Ivy

Go to download
Show more of this group Show more artifacts with this name
Show all versions of sphinx4-core Show documentation
Sphinx4 core
The newest version!
/*
 * Copyright 1999-2002 Carnegie Mellon University.  
 * Portions Copyright 2002 Sun Microsystems, Inc.  
 * Portions Copyright 2002 Mitsubishi Electric Research Laboratories.
 * All Rights Reserved.  Use is subject to license terms.
 * 
 * See the file "license.terms" for information on usage and
 * redistribution of this file, and for a DISCLAIMER OF ALL 
 * WARRANTIES.
 *
 */

package edu.cmu.sphinx.trainer;

import edu.cmu.sphinx.linguist.acoustic.Unit;
import edu.cmu.sphinx.linguist.dictionary.TextDictionary;
import edu.cmu.sphinx.linguist.dictionary.Pronunciation;

/** Dummy trainer dictionary. */
public class TrainerDictionary extends TextDictionary {

    static final String UTTERANCE_BEGIN_SYMBOL = "";
    static final String UTTERANCE_END_SYMBOL = "";
    static final String SILENCE_SYMBOL = "SIL";


    /**
     * Gets a word pronunciation graph. Dummy initial and final states optional.
     *
     * @param word     the word
     * @param hasDummy if true, the graph will have dummy initial and final states
     * @return the graph
     */
    public Graph getWordGraph(String word, boolean hasDummy) {
        Graph wordGraph = new Graph();
        Pronunciation[] pronunciations;
        Unit[] units;
        Node prevNode;
        Node wordNode = null;
        int pronunciationID = 0;
        String wordWithoutParentheses = word.replaceFirst("\\(.*\\)", "");

        if (word.equals(wordWithoutParentheses)) {
            pronunciationID = 0;
        } else {
            String number =
                    word.replaceFirst(".*\\(", "").replaceFirst("\\)", "");
            try {
                pronunciationID = Integer.parseInt(number);
            } catch (NumberFormatException nfe) {
                throw new Error("Word with invalid pronunciation ID", nfe);
            }
        }
        pronunciations
                = getWord(wordWithoutParentheses).getPronunciations();
        if (pronunciations == null) {
            System.out.println("Pronunciation not found for word " +
                    wordWithoutParentheses);
            return null;
        }
        if (pronunciationID >= pronunciations.length) {
            System.out.println("Dictionary has only " +
                    pronunciations.length +
                    " for word " + word);
            return null;
        }
        units = pronunciations[pronunciationID].getUnits();
        assert units != null : "units is empty: problem with dictionary?";

        // Now, create the graph, where each node contains a single unit
        if (hasDummy) {
            Node initialNode = new Node(NodeType.DUMMY);
            wordGraph.addNode(initialNode);
            wordGraph.setInitialNode(initialNode);
            prevNode = initialNode;
        } else {
            prevNode = null;
        }
        for (Unit unit : units) {
            // create a new node for the next unit
            wordNode = new Node(NodeType.PHONE, unit.getName());
            if (prevNode == null) {
                wordGraph.addNode(wordNode);
                wordGraph.setInitialNode(wordNode);
            } else {
                // Link the new node into the graph
                wordGraph.linkNodes(prevNode, wordNode);
            }
            prevNode = wordNode;
        }
        // All words are done. Just add the final node
        if (hasDummy) {
            wordNode = new Node(NodeType.DUMMY);
            wordGraph.linkNodes(prevNode, wordNode);
        }
        assert wordNode != null;
        wordGraph.setFinalNode(wordNode);

        return wordGraph;
    }


    /** Prints out dictionary as a string. */
    @Override
    public String toString() {
        return "DEFAULT";
    }
}