//    Any other category: NewLine, MidNum, CR, LF, Double_Quote, Katakana, and
//    Other.
//    These categories should always result in a break between two cased letters.
//    Rule: Always break.
//
// Note 1: the Katakana and MidNum categories can, in esoteric cases, result in
// preventing a break between two cased letters. For now we will ignore this
// (e.g. [ALetter] [ExtendNumLet] [Katakana] [ExtendNumLet] [ALetter] and

32D6          ; mapped                 ; 30AD          # 1.1  CIRCLED KATAKANA KI
32D7          ; mapped                 ; 30AF          # 1.1  CIRCLED KATAKANA KU
32D8          ; mapped                 ; 30B1          # 1.1  CIRCLED KATAKANA KE
32D9          ; mapped                 ; 30B3          # 1.1  CIRCLED KATAKANA KO
32DA          ; mapped                 ; 30B5          # 1.1  CIRCLED KATAKANA SA

    protected final Client client;
    private final SuggestSettings settings;
    protected final AnalyzerSettings analyzerSettings;

    protected final Transliterator transliterator = Transliterator.getInstance("Hiragana-Katakana");

    public AnalyzerConverter(final Client client, final SuggestSettings settings) {
        this.client = client;
        this.settings = settings;
        this.analyzerSettings = settings.analyzer();
    }

import com.ibm.icu.text.Transliterator;

public class KatakanaConverter implements ReadingConverter {

    protected final Transliterator transliterator = Transliterator.getInstance("Hiragana-Katakana");

    protected volatile boolean initialized = false;

    protected TokenizerFactory tokenizerFactory = null;

    public KatakanaConverter() {
        // nothing
    }

        final UnicodeBlock block = UnicodeBlock.of(text.codePointAt(0));
        if (block == UnicodeBlock.CJK_UNIFIED_IDEOGRAPHS //
                || block == UnicodeBlock.HIRAGANA //
                || block == UnicodeBlock.KATAKANA //
                || block == UnicodeBlock.HANGUL_SYLLABLES //
        ) {
            return QueryBuilders.prefixQuery(f, text);
        }
        return QueryBuilders.matchPhraseQuery(f, text);
    }

func IsTitle (r <type -21>) <type -15>;
func IsUpper (r <type -21>) <type -15>;
var Javanese <type 1>;
var Join_Control <type 1>;
var Kaithi <type 1>;
var Kannada <type 1>;
var Katakana <type 1>;
var Kayah_Li <type 1>;
var Kharoshthi <type 1>;
var Khmer <type 1>;
var Khojki <type 1>;
var Khudawadi <type 1>;
var L <type 1>;
var Lao <type 1>;
var Latin <type 1>;
var Lepcha <type 1>;

	Kannada                = _Kannada                // Kannada is the set of Unicode characters in script Kannada.
	Katakana               = _Katakana               // Katakana is the set of Unicode characters in script Katakana.
	Kawi                   = _Kawi                   // Kawi is the set of Unicode characters in script Kawi.

		{"IsSymbol", Func, 0},
		{"IsTitle", Func, 0},
		{"IsUpper", Func, 0},
		{"Javanese", Var, 0},
		{"Join_Control", Var, 0},
		{"Kaithi", Var, 0},
		{"Kannada", Var, 0},
		{"Katakana", Var, 0},
		{"Kawi", Var, 21},
		{"Kayah_Li", Var, 0},
		{"Kharoshthi", Var, 0},
		{"Khitan_Small_Script", Var, 16},
		{"Khmer", Var, 0},
		{"Khojki", Var, 4},
		{"Khudawadi", Var, 4},

pkg unicode, var Javanese *RangeTable
pkg unicode, var Join_Control *RangeTable
pkg unicode, var Kaithi *RangeTable
pkg unicode, var Kannada *RangeTable
pkg unicode, var Katakana *RangeTable
pkg unicode, var Kayah_Li *RangeTable
pkg unicode, var Kharoshthi *RangeTable
pkg unicode, var Khmer *RangeTable
pkg unicode, var L *RangeTable
pkg unicode, var Lao *RangeTable