ひらがなを音声認識アプリケーションJuliusで使われている音素表現(.htkdic)に変換する

Hiragana2Phoneme.java

package net.mekajiki;

import com.ibm.icu.text.Transliterator;
import java.util.ArrayList;
import java.util.List;

public class Hiragana2Phoneme {
    public static String hiragana2Phoneme(String text) {
        return romaji2Phoneme(hiragana2Romaji(text));
    }

    public static String hiragana2Romaji(String text) {
        Transliterator transliterator = Transliterator.getInstance("Hiragana-Latin");
        return transliterator.transliterate(text);
    }

    public static String hiragana2katakana(String text) {
        Transliterator transliterator = Transliterator.getInstance("Hiragana-Katakana");
        return transliterator.transliterate(text);
    }

    public static String romaji2Phoneme(String text) {
        List<String[]> replaceList = new ArrayList<String[]>();

        //撥音のN
        replaceList.add(new String[]{"n'", "N"});
        replaceList.add(new String[]{"n([^aiueo])", "N$1"});

        //おう→おーの音便
        replaceList.add(new String[]{"ou", "o:"});

        //長音
        replaceList.add(new String[]{"ā", "a:"});
        replaceList.add(new String[]{"ī", "i:"});
        replaceList.add(new String[]{"ū", "u:"});
        replaceList.add(new String[]{"ē", "e:"});
        replaceList.add(new String[]{"ō", "o:"});

        //同じ母音の連続→長音
        String[] vowels = {"a", "i", "u", "e", "o"};
        for(String vowel : vowels) {
            replaceList.add(new String[]{vowel + "{2,}", vowel + ":"});
        }

        //スペース区切り
        replaceList.add(new String[]{"[aiueoN]:?", "$0 "});
        replaceList.add(new String[]{"[^aiueoN]{1,4}", "$0 "});
        replaceList.add(new String[]{"\\s+", " "});

        //tchを促音とみなす
        replaceList.add(new String[]{"tch", "q ch"});

        //dzは音素として不正
        replaceList.add(new String[]{"dz", "z"});

        //同じ子音の連続を促音とみなす
        replaceList.add(new String[]{" ([^aiueo])$1", " q $1"});

        for(String[] map : replaceList) {
            text = text.replaceAll(map[0], map[1]);
        }

        return text;
    }
}