private Token getToken(char character) {
Token token = new Token();
final String letter = Character.toString(character);
token.source = letter;
int offset = -1;
int cmp;
if (character < 256) {
token.type = Token.LATIN;
token.target = letter;
return token;
} else if (character < FIRST_UNIHAN) {
token.type = Token.UNKNOWN;
token.target = letter;
return token;
} else {
cmp = COLLATOR.compare(letter, FIRST_PINYIN_UNIHAN);
if (cmp < 0) {
token.type = Token.UNKNOWN;
token.target = letter;
return token;
} else if (cmp == 0) {
token.type = Token.PINYIN;
offset = 0;
} else {
cmp = COLLATOR.compare(letter, LAST_PINYIN_UNIHAN);
if (cmp > 0) {
token.type = Token.UNKNOWN;
token.target = letter;
return token;
} else if (cmp == 0) {
token.type = Token.PINYIN;
offset = UNIHANS.length - 1;
}
}
}
token.type = Token.PINYIN;
if (offset < 0) {
int begin = 0;
int end = UNIHANS.length - 1;
while (begin <= end) {
offset = (begin + end) / 2;
final String unihan = Character.toString(UNIHANS[offset]);
cmp = COLLATOR.compare(letter, unihan);
if (cmp == 0) {
break;
} else if (cmp > 0) {
begin = offset + 1;
} else {
end = offset - 1;
}
}
}
if (cmp < 0) {
offset--;
}
StringBuilder pinyin = new StringBuilder();
for (int j = 0; j < PINYINS[offset].length && PINYINS[offset][j] != 0; j++) {
pinyin.append((char) PINYINS[offset][j]);
}
token.target = pinyin.toString();
return token;
}