nani/app

Julien LepillerSat Jul 31 14:40:47+0200 2021

5ee3eb5

Improve furigana matching

app/src/main/java/eu/lepiller/nani/result/Result.java

         // split the text into kanji / not kanji portions
         ArrayList<String> portions = new ArrayList<>();
         ArrayList<String> portionsMatcher = new ArrayList<>();
         StringBuilder current = new StringBuilder();
         StringBuilder currentMatcher = new StringBuilder();
         Character.UnicodeBlock b = CJK_UNIFIED_IDEOGRAPHS;
         MojiConverter converter = new MojiConverter();
                 // if the headwork contains katakana, convert it to hiragana to match pronunciation
                 // better.
                 current.append(txt.charAt(i));
                 if(b2 == KATAKANA) {
                     String s = new String(new char[]{txt.charAt(i)});
                     String hiragana = converter.convertRomajiToHiragana(converter.convertKanaToRomaji(s));
                     currentMatcher.append(hiragana.charAt(0));
                 } else {
                     currentMatcher.append(txt.charAt(i));
                 }
             } else {
                 String s = current.toString();
                 if(!s.isEmpty())
                     portions.add(s);
                 s = currentMatcher.toString();
                 if(!s.isEmpty())
                     portionsMatcher.add(s);
                 current = new StringBuilder();
                 currentMatcher = new StringBuilder();
                 current.append(txt.charAt(i));
                 if(b2 == KATAKANA) {
                     String katakana = new String(new char[]{txt.charAt(i)});
                     String hiragana = converter.convertRomajiToHiragana(converter.convertKanaToRomaji(katakana));
                     currentMatcher.append(hiragana.charAt(0));
                 } else {
                     currentMatcher.append(txt.charAt(i));
                 }
             }
             b = b2;
         String str = current.toString();
         if(!str.isEmpty())
             portions.add(str);
         str = currentMatcher.toString();
         if(!str.isEmpty()) {
             portionsMatcher.add(str);
         }
         // Create a regexp to match kanji places
         current = new StringBuilder();
         current.append("^");
         for(String s: portionsMatcher) {
         for(String s: portions) {
             if(Character.UnicodeBlock.of(s.charAt(0)) == CJK_UNIFIED_IDEOGRAPHS) {
                 current.append("(.*)");
             } else {
                 current.append(s);
                 for(Character c: s.toCharArray()) {
                     if(Character.UnicodeBlock.of(c) == KATAKANA) {
                         current.append("[");
                         current.append(c);
                         current.append(converter.convertRomajiToHiragana(converter.convertKanaToRomaji(new String(new char[]{c}))));
                         current.append("]");
                     } else {
                         current.append(c);
                     }
                 }
             }
         }
         current.append("$");