/**
* 转换输入汉字的首字母，多个汉字是只反回第一个汉字的首字母
*
* @param chinese
* @return
*/
public static String getFirstLetter(String chines) {
chines = replaceString(chines);
String pinyinName = "";
char[] nameChar = chines.toCharArray();
HanyuPinyinOutputFormat defaultFormat = new HanyuPinyinOutputFormat();
defaultFormat.setCaseType(HanyuPinyinCaseType.UPPERCASE);
defaultFormat.setToneType(HanyuPinyinToneType.WITHOUT_TONE);
// 如果传入的第一个不是汉字 则取第二个汉字的首字母 否则只反回第个汉字
if (nameChar[0] > 128) {
try {
String[] pys = PinyinHelper.toHanyuPinyinStringArray(nameChar[0], defaultFormat);
if (pys != null) {
char nchar = pys[0].charAt(0);
pinyinName += nchar;
}
} catch (BadHanyuPinyinOutputFormatCombination e) {
e.printStackTrace();
}
} else if (nameChar[1] > 128) {
try {
String[] pys = PinyinHelper.toHanyuPinyinStringArray(nameChar[1], defaultFormat);
if (pys != null) {
char nchar = pys[0].charAt(0);
pinyinName += nchar;
}
} catch (BadHanyuPinyinOutputFormatCombination e) {
e.printStackTrace();
}
} else {
pinyinName += "#";
}
return pinyinName;
}
/**
* 汉字转拼音的方法
*
* @param name 汉字
* @return 拼音
*/
public static String HanyuToPinyin(String name) {
if (name == null || name.equals("")) {
return "";
}
String pinyinName = "";
/** 解决英文名字或首字母为英文的名字问题 如果名字首字母位英文，则直接返回首字母，不再进行转码 add by ChenPengzhen */
String az = "abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ";
if (az.contains(String.valueOf(name.charAt(0)))) {
return (name.charAt(0) + "").toUpperCase();
} else {
char[] nameChar = name.toCharArray();
HanyuPinyinOutputFormat defaultFormat = new HanyuPinyinOutputFormat();
defaultFormat.setCaseType(HanyuPinyinCaseType.LOWERCASE);
defaultFormat.setToneType(HanyuPinyinToneType.WITHOUT_TONE);
for (int i = 0; i < nameChar.length; i++) {
if (nameChar[i] > 128) {
try {
pinyinName += PinyinHelper.toHanyuPinyinStringArray(nameChar[i], defaultFormat)[0];
} catch (Exception e) {
e.printStackTrace();
}
}
}
if (pinyinName.equals("")) {
return "#";
}
return (pinyinName.charAt(0) + "").toUpperCase();
}
}
/**
* 汉字转换为汉语拼音首字母，英文字符不�?
*
* @author gzs
* @param chines 汉字
* @return 拼音
*/
public static String getPinyinFrt(String chines) {
String pinyinName = "";
// 转化为字�?
char[] nameChar = chines.toCharArray();
// 汉语拼音格式输出�?
HanyuPinyinOutputFormat defaultFormat = new HanyuPinyinOutputFormat();
// 输出设置,大小�?音标方式�?
defaultFormat.setCaseType(HanyuPinyinCaseType.UPPERCASE);
defaultFormat.setToneType(HanyuPinyinToneType.WITHOUT_TONE);
for (int i = 0; i < nameChar.length; i++) {
// 如果是中�?
if (nameChar[i] > 128) {
try {
String[] temp = PinyinHelper.toHanyuPinyinStringArray(nameChar[i], defaultFormat);
if (temp != null && temp.length > 0 && temp[0].length() > 0) {
pinyinName += temp[0].charAt(0);
}
} catch (BadHanyuPinyinOutputFormatCombination e) {
e.printStackTrace();
}
} else { // 为英文字�?
pinyinName += nameChar[i];
}
}
return pinyinName;
}
public static String[] toDiffHanyuPinyinStringArray(char ch, HanyuPinyinOutputFormat format)
throws BadHanyuPinyinOutputFormatCombination {
String[] pinyinStrArray = null;
if ('\u9fa5' < ch || '\u4e00' > ch) { // 不是汉字
return new String[] {("" + ch).toUpperCase()};
}
pinyinStrArray = PinyinHelper.toHanyuPinyinStringArray(ch, format);
int count = 0;
for (int i = 0; i < pinyinStrArray.length; i++) {
for (int j = i + 1; j < pinyinStrArray.length; j++) {
if (pinyinStrArray[i].equals(pinyinStrArray[j])) {
pinyinStrArray[j] = "null";
}
}
if (pinyinStrArray[i] != "null") {
++count;
}
}
String[] reStringArray = new String[count];
int j = 0;
for (int i = 0; i < pinyinStrArray.length; i++) {
if (!pinyinStrArray[i].equals("null")) {
reStringArray[j++] = pinyinStrArray[i];
}
}
return reStringArray;
}
public String toPinYin(String name) {
HanyuPinyinOutputFormat hanyuPinyin = new HanyuPinyinOutputFormat();
/** *大写 */
hanyuPinyin.setCaseType(HanyuPinyinCaseType.UPPERCASE);
/** *无音调 */
hanyuPinyin.setToneType(HanyuPinyinToneType.WITH_TONE_MARK);
/** *'¨¹' is "v" */
hanyuPinyin.setVCharType(HanyuPinyinVCharType.WITH_U_UNICODE);
String code = "";
String[] pinyinArray = null;
for (int i = 0; i < name.length(); i++) {
char hanzi = name.charAt(i);
try {
/** *是否在汉字范围内 */
if (hanzi >= 0x4e00 && hanzi <= 0x9fa5) {
pinyinArray = PinyinHelper.toHanyuPinyinStringArray(hanzi, hanyuPinyin);
code += pinyinArray[0].substring(0, 1);
} else if (hanzi >= 0x61 && hanzi <= 0x7a) {
code += String.valueOf((char) (hanzi - 32));
} else if (hanzi >= 0x41 && hanzi <= 0x5a) {
code += hanzi;
} else {
code += "#";
}
} catch (BadHanyuPinyinOutputFormatCombination e) {
e.printStackTrace();
}
}
return code;
}
/**
* 得到 全拼
*
* @param src
* @return
*/
public static String getPingYin(String src) {
char[] t1 = null;
t1 = src.toCharArray();
String[] t2 = new String[t1.length];
HanyuPinyinOutputFormat t3 = new HanyuPinyinOutputFormat();
t3.setCaseType(HanyuPinyinCaseType.LOWERCASE);
t3.setToneType(HanyuPinyinToneType.WITHOUT_TONE);
t3.setVCharType(HanyuPinyinVCharType.WITH_V);
String t4 = "";
int t0 = t1.length;
try {
for (int i = 0; i < t0; i++) {
// 判断是否为汉字字符
if (Character.toString(t1[i]).matches("[\\u4E00-\\u9FA5]+")) {
t2 = PinyinHelper.toHanyuPinyinStringArray(t1[i], t3);
t4 += t2[0];
} else {
t4 += Character.toString(t1[i]);
}
}
return t4;
} catch (BadHanyuPinyinOutputFormatCombination e1) {
e1.printStackTrace();
}
return t4;
}
/**
* 将汉字转换为全拼
*
* @param src
* @return String
*/
public static String getPinYin(String src) {
char[] t1 = null;
t1 = src.toCharArray();
// System.out.println(t1.length);
String[] t2 = new String[t1.length];
// System.out.println(t2.length);
// 设置汉字拼音输出的格式
HanyuPinyinOutputFormat t3 = new HanyuPinyinOutputFormat();
t3.setCaseType(HanyuPinyinCaseType.LOWERCASE);
t3.setToneType(HanyuPinyinToneType.WITHOUT_TONE);
t3.setVCharType(HanyuPinyinVCharType.WITH_V);
String t4 = "";
int t0 = t1.length;
try {
for (int i = 0; i < t0; i++) {
// 判断是否为汉字字符
// System.out.println(t1[i]);
if (Character.toString(t1[i]).matches("[\\u4E00-\\u9FA5]+")) {
t2 = PinyinHelper.toHanyuPinyinStringArray(t1[i], t3); // 将汉字的几种全拼都存到t2数组中
t4 += t2[0]; // 取出该汉字全拼的第一种读音并连接到字符串t4后
} else {
// 如果不是汉字字符，直接取出字符并连接到字符串t4后
t4 += Character.toString(t1[i]);
}
}
} catch (BadHanyuPinyinOutputFormatCombination e) {
// TODO Auto-generated catch block
e.printStackTrace();
}
return t4;
}
/**
* 把参数名转换为字段名，把中文转换为拼音，特殊字符转换为下划线
*
* @param string
* @return
*/
public static String pinyin(String string) {
Pattern pattern = Pattern.compile("[\\u4e00-\\u9fa5]");
Pattern pattern2 = Pattern.compile("^\\w+$");
char ch[] = string.toCharArray();
StringBuffer sb = new StringBuffer();
for (int i = 0; i < ch.length; i++) {
String t = String.valueOf(ch[i]);
Matcher matcher = pattern.matcher(t);
Matcher matcher2 = pattern2.matcher(t);
if (matcher.find()) { // 匹配中文
String pinyin =
PinyinHelper.toHanyuPinyinStringArray(matcher.group().toCharArray()[0])[0]
.replaceAll("\\d+", "")
.replace("u:", "v");
sb.append(pinyin);
} else {
if (matcher2.find()) { // 匹配字母数字下划线
} else {
t = "_";
}
sb.append(t);
}
}
return sb.toString();
}
/**
* 获取拼音
*
* @param chinese 含中文的字符串，若不含中文，原样输出
* @return 转换后的文本
* @throws BadHanyuPinyinOutputFormatCombination
*/
private Collection<String> GetPyString(String chinese)
throws BadHanyuPinyinOutputFormatCombination {
List<String[]> pinyinList = new ArrayList<>();
for (int i = 0; i < chinese.length(); i++) {
String[] pinyinArray =
PinyinHelper.toHanyuPinyinStringArray(chinese.charAt(i), this.outputFormat);
if (pinyinArray != null && pinyinArray.length > 0) {
pinyinList.add(pinyinArray);
}
}
Set<String> pinyins = null;
for (String[] array : pinyinList) {
if (pinyins == null || pinyins.isEmpty()) {
pinyins = new HashSet<>();
Collections.addAll(pinyins, array);
} else {
Set<String> pres = pinyins;
pinyins = new HashSet<>();
for (String pre : pres) {
for (String charPinyin : array) {
pinyins.add(pre + charPinyin);
}
}
}
}
return pinyins;
}
/**
* 转换输入汉字的全拼(小写) 多音字只反回第一个拼音,拼音首字母大写其它小写
*
* @param chines
* @return
*/
public static String converterToPingYingHeadUppercase(String chines) {
chines = replaceString(chines);
StringBuffer pinyinName = new StringBuffer();
char[] nameChar = chines.toCharArray();
HanyuPinyinOutputFormat defaultFormat = new HanyuPinyinOutputFormat();
defaultFormat.setCaseType(HanyuPinyinCaseType.LOWERCASE);
defaultFormat.setToneType(HanyuPinyinToneType.WITHOUT_TONE);
for (int i = 0; i < nameChar.length; i++) {
if (nameChar[i] > 128) {
try {
String[] pys = PinyinHelper.toHanyuPinyinStringArray(nameChar[i], defaultFormat);
String headPy = pys == null ? nameChar[i] + "" : pys[0];
pinyinName.append(
conversionHeadUppercase(headPy.charAt(0))
+ headPy.substring(1, headPy.length())
+ "-");
} catch (BadHanyuPinyinOutputFormatCombination e) {
e.printStackTrace();
}
} else {
pinyinName.append(nameChar[i]);
}
}
return pinyinName.toString();
}
@Test
public void testPinyinHelper() throws BadHanyuPinyinOutputFormatCombination {
// System.out.println(
// PinyinHelper.toHanyuPinyinString(test, new HanyuPinyinOutputFormat(), " ")
// );
HanyuPinyinOutputFormat FORMAT_DEFAULT = new HanyuPinyinOutputFormat();
FORMAT_DEFAULT.setToneType(HanyuPinyinToneType.WITHOUT_TONE);
LogUtil.show(PinyinHelper.toHanyuPinyinStringArray('中', FORMAT_DEFAULT));
}
/**
* 获取字符串的首字符拼音
*
* @param str
* @return
*/
public static String getPinYinHeadChar(String str) {
String convert = "";
char word = str.charAt(0);
// 提取汉字的首字母
String[] pinyinArray = PinyinHelper.toHanyuPinyinStringArray(word);
if (pinyinArray != null) {
convert += pinyinArray[0].charAt(0);
} else {
convert += word;
}
return convert;
}
private String[] pinyinSynonymArr(char c) throws BadHanyuPinyinOutputFormatCombination {
String[] pinyinArr = PinyinHelper.toHanyuPinyinStringArray(c, defaultFormat);
if (pinyinArr == null) {
return null;
}
String[] flattenPinyinArr = flattenPinyinArr(pinyinArr);
String[] withOrgCharPinyinArr = new String[flattenPinyinArr.length + 1];
withOrgCharPinyinArr[0] = Character.toString(c);
System.arraycopy(flattenPinyinArr, 0, withOrgCharPinyinArr, 1, flattenPinyinArr.length);
return withOrgCharPinyinArr;
}
public static String generateKey(String seq) {
StringBuffer letters = new StringBuffer(seq.length());
char[] chars = seq.toCharArray();
for (int i = 0; i < chars.length; i++) {
String[] pinYin = PinyinHelper.toHanyuPinyinStringArray(chars[i]);
if (null == pinYin || "none0".equals(pinYin[0])) {
letters.append(chars[i]);
} else {
letters.append(pinYin[0].charAt(0));
}
}
return letters.toString().toLowerCase();
}
/**
* 得到中文首字母
*
* @param str
* @return
*/
public static String getPinYinHeadChar(String str) {
String convert = "";
for (int j = 0; j < str.length(); j++) {
char word = str.charAt(j);
String[] pinyinArray = PinyinHelper.toHanyuPinyinStringArray(word);
if (pinyinArray != null) {
convert += pinyinArray[0].charAt(0);
} else {
convert += word;
}
}
return convert;
}
public String getCharacterPinYin(char c) {
HanyuPinyinOutputFormat format = null;
format = new HanyuPinyinOutputFormat();
format.setToneType(HanyuPinyinToneType.WITHOUT_TONE);
String[] pinyin = null;
try {
pinyin = PinyinHelper.toHanyuPinyinStringArray(c, format);
} catch (BadHanyuPinyinOutputFormatCombination e) {
e.printStackTrace();
}
// 如果c不是汉字，toHanyuPinyinStringArray会返回null
if (pinyin == null) return null;
// 只取一个发音，如果是多音字，仅取第一个发音
return pinyin[0];
}
/**
* 获取指定汉字的拼音数组<br>
* 注：对声调支持不是很好，有些会显示乱码，例如“很”
*
* @param ch 汉字字符
* @param isShowMark 是否显示声调
* @return 返回该汉字拼音字符串数组，非汉字返回null
*/
public static String[] getPinyin(char ch, boolean isShowMark) {
HanyuPinyinOutputFormat format = new HanyuPinyinOutputFormat();
// 为true则设置输出格式带声调
if (isShowMark) {
format.setToneType(HanyuPinyinToneType.WITH_TONE_MARK);
format.setVCharType(HanyuPinyinVCharType.WITH_U_UNICODE);
} else {
format.setToneType(HanyuPinyinToneType.WITHOUT_TONE);
}
String[] pinyins = null;
try {
pinyins = PinyinHelper.toHanyuPinyinStringArray(ch, format);
} catch (BadHanyuPinyinOutputFormatCombination e) {
e.printStackTrace();
}
return null == pinyins ? null : pinyins;
}
public static String getFirstSpellByString(String string) {
HanyuPinyinOutputFormat format = new HanyuPinyinOutputFormat();
format.setToneType(HanyuPinyinToneType.WITHOUT_TONE);
char[] chars = string.toCharArray();
StringBuffer stringBuffer = new StringBuffer();
for (char c : chars) {
try {
String[] strings = PinyinHelper.toHanyuPinyinStringArray(c, format);
if (strings != null && strings.length > 0) {
stringBuffer.append(strings[0].charAt(0));
}
} catch (BadHanyuPinyinOutputFormatCombination e) {
e.printStackTrace();
}
}
return stringBuffer.toString();
}
private static String getFullSpell(String chinese) {
StringBuffer pybf = new StringBuffer();
char[] arr = chinese.toCharArray();
HanyuPinyinOutputFormat defaultFormat = new HanyuPinyinOutputFormat();
defaultFormat.setCaseType(HanyuPinyinCaseType.UPPERCASE);
defaultFormat.setToneType(HanyuPinyinToneType.WITHOUT_TONE);
for (int i = 0; i < arr.length; i++) {
if (arr[i] > 128) {
try {
pybf.append(PinyinHelper.toHanyuPinyinStringArray(arr[i], defaultFormat)[0]);
} catch (BadHanyuPinyinOutputFormatCombination e) {
e.printStackTrace();
}
} else {
pybf.append(arr[i]);
}
}
return pybf.toString();
}
/**
* 汉字转换位汉语拼音，英文字符不变
*
* @param chines 汉字
* @return 拼音
*/
public static String converterToSpell(String chines) {
String pinyinName = "";
char[] nameChar = chines.toCharArray();
HanyuPinyinOutputFormat defaultFormat = new HanyuPinyinOutputFormat();
defaultFormat.setCaseType(HanyuPinyinCaseType.LOWERCASE);
defaultFormat.setToneType(HanyuPinyinToneType.WITHOUT_TONE);
for (int i = 0; i < nameChar.length; i++) {
if (nameChar[i] > 128) {
try {
pinyinName += PinyinHelper.toHanyuPinyinStringArray(nameChar[i], defaultFormat)[0];
} catch (BadHanyuPinyinOutputFormatCombination e) {
e.printStackTrace();
}
} else {
pinyinName += nameChar[i];
}
}
return pinyinName;
}
public static String getPinYin(String text) {
if (!text.equalsIgnoreCase("")) {
// check if English string
boolean valid = text.matches("^[\u0000-\u0080]+$");
if (!valid) {
String[] pinyinArray;
List<String> pinyin = new ArrayList<String>();
for (int i = 0; i < text.length(); i++) {
pinyin.add(PinyinHelper.toHanyuPinyinStringArray(text.charAt(i))[0]);
}
text = pinyin.get(0);
for (int i = 1; i < pinyin.size(); i++) {
text += (" " + pinyin.get(i));
}
return text;
} else return "";
}
return null;
}
/**
* 获取指定字符串的拼音字符串<br>
* 注：对声调支持不是很好，有些会显示乱码，例如“很”
*
* @param str 字符串
* @param isShowMark 是否显示声调
* @return 返回拼音字符串，非汉字则照原样输出
*/
public static String getPinyins(String str, boolean isShowMark) {
StringBuffer pinyinStr = new StringBuffer();
HanyuPinyinOutputFormat format = new HanyuPinyinOutputFormat();
// 为true则设置输出格式带声调
if (isShowMark) {
format.setToneType(HanyuPinyinToneType.WITH_TONE_MARK);
format.setVCharType(HanyuPinyinVCharType.WITH_U_UNICODE);
} else {
format.setToneType(HanyuPinyinToneType.WITHOUT_TONE);
}
for (int i = 0; i < str.length(); i++) {
try {
String[] pinyins = PinyinHelper.toHanyuPinyinStringArray(str.charAt(i), format);
if (null == pinyins) pinyinStr.append(str.charAt(i));
else pinyinStr.append(pinyins[0]);
} catch (BadHanyuPinyinOutputFormatCombination e) {
e.printStackTrace();
}
}
return pinyinStr.toString();
}
/**
* 将字符串中的中文转化为拼音,其他字符不变
*
* @param inputString
* @return
*/
public static String getPingYin(String inputString) {
HanyuPinyinOutputFormat format = new HanyuPinyinOutputFormat();
format.setCaseType(HanyuPinyinCaseType.LOWERCASE);
format.setToneType(HanyuPinyinToneType.WITHOUT_TONE);
format.setVCharType(HanyuPinyinVCharType.WITH_V);
char[] input = inputString.trim().toCharArray();
String output = "";
try {
for (int i = 0; i < input.length; i++) {
if (java.lang.Character.toString(input[i]).matches("[\\u4E00-\\u9FA5]+")) {
String[] temp = PinyinHelper.toHanyuPinyinStringArray(input[i], format);
output += temp[0];
} else output += java.lang.Character.toString(input[i]);
}
} catch (BadHanyuPinyinOutputFormatCombination e) {
e.printStackTrace();
}
return output;
}
/**
* 获取拼音集合
*
* @author wyh
* @param src
* @return Set<String>
*/
public static Set<String> getPinyin(String src) {
if (src != null && !src.trim().equalsIgnoreCase("")) {
char[] srcChar;
srcChar = src.toCharArray();
// 汉语拼音格式输出类
HanyuPinyinOutputFormat hanYuPinOutputFormat = new HanyuPinyinOutputFormat();
// 输出设置，大小写，音标方式等
hanYuPinOutputFormat.setCaseType(HanyuPinyinCaseType.LOWERCASE);
hanYuPinOutputFormat.setToneType(HanyuPinyinToneType.WITHOUT_TONE);
hanYuPinOutputFormat.setVCharType(HanyuPinyinVCharType.WITH_V);
String[][] temp = new String[src.length()][];
for (int i = 0; i < srcChar.length; i++) {
char c = srcChar[i];
// 是中文或者a-z或者A-Z转换拼音(我的需求，是保留中文或者a-z或者A-Z)
if (String.valueOf(c).matches("[\\u4E00-\\u9FA5]+")) {
try {
temp[i] = PinyinHelper.toHanyuPinyinStringArray(srcChar[i], hanYuPinOutputFormat);
} catch (BadHanyuPinyinOutputFormatCombination e) {
e.printStackTrace();
}
} else if (((int) c >= 65 && (int) c <= 90)
|| ((int) c >= 97 && (int) c <= 122)
|| (int) c <= '9' && (int) c >= '0') {
temp[i] = new String[] {String.valueOf(srcChar[i])};
} else {
temp[i] = new String[] {""};
}
}
String[] pingyinArray = Exchange(temp);
Set<String> pinyinSet = new HashSet<String>();
for (int i = 0; i < pingyinArray.length; i++) {
pinyinSet.add(pingyinArray[i]);
}
return pinyinSet;
}
return null;
}
private static String getFirstSpell(String chinese) {
StringBuffer pybf = new StringBuffer();
char[] arr = chinese.toCharArray();
HanyuPinyinOutputFormat defaultFormat = new HanyuPinyinOutputFormat();
defaultFormat.setCaseType(HanyuPinyinCaseType.UPPERCASE);
defaultFormat.setToneType(HanyuPinyinToneType.WITHOUT_TONE);
for (int i = 0; i < arr.length; i++) {
if (arr[i] > 128) {
try {
String[] temp = PinyinHelper.toHanyuPinyinStringArray(arr[i], defaultFormat);
if (temp != null) {
pybf.append(temp[0].charAt(0));
}
} catch (BadHanyuPinyinOutputFormatCombination e) {
e.printStackTrace();
}
} else {
pybf.append(arr[i]);
}
}
return pybf.toString().replaceAll("\\W", "").trim();
}
public static String getPinYin4J(String input) {
if (mFormat == null) {
mFormat = new HanyuPinyinOutputFormat();
mFormat.setCaseType(HanyuPinyinCaseType.LOWERCASE);
mFormat.setToneType(HanyuPinyinToneType.WITHOUT_TONE);
}
StringBuilder builder = new StringBuilder();
int size = input.length();
for (int i = 0; i < size; i++) {
String[] array;
try {
char c = input.charAt(i);
array = PinyinHelper.toHanyuPinyinStringArray(c, mFormat);
if (array == null) {
builder.append(c);
} else {
builder.append(array[0] != null ? array[0] : c + "");
}
} catch (BadHanyuPinyinOutputFormatCombination e) {
}
}
return builder.toString();
}
/**
* 反回输入字符串的所有首字母(小写)并去除非汉字
*
* @param chines
* @return
*/
public static String converterToAllFirstSpellsLowercase(String chines) {
chines = replaceString(chines);
StringBuffer pinyinName = new StringBuffer();
char[] nameChar = chines.toCharArray();
HanyuPinyinOutputFormat defaultFormat = new HanyuPinyinOutputFormat();
defaultFormat.setCaseType(HanyuPinyinCaseType.LOWERCASE);
defaultFormat.setToneType(HanyuPinyinToneType.WITHOUT_TONE);
for (int i = 0; i < nameChar.length; i++) {
if (nameChar[i] > 128) {
try {
String[] pys = PinyinHelper.toHanyuPinyinStringArray(nameChar[i], defaultFormat);
if (pys != null) {
char nchar = pys[0].charAt(0);
pinyinName.append(nchar);
}
} catch (BadHanyuPinyinOutputFormatCombination e) {
e.printStackTrace();
}
} else {
pinyinName.append(nameChar[i]);
}
}
return pinyinName.toString();
}
// 通用方法有了。。不过太耗内存。
public static String Name2Pinyin(String name) throws BadHanyuPinyinOutputFormatCombination {
// 注意在调用此方法是不能有空格。在前面去掉..这里就不去了。。看着费劲
if (name == null || name.length() == 0 || name.length() > 4) {
return null;
}
HanyuPinyinOutputFormat format = new HanyuPinyinOutputFormat();
format.setCaseType(HanyuPinyinCaseType.UPPERCASE);
format.setToneType(HanyuPinyinToneType.WITHOUT_TONE);
format.setVCharType(HanyuPinyinVCharType.WITH_V);
char fourChar = 0;
if (name.length() == 4) {
fourChar = name.charAt(3);
name = name.substring(0, 3);
}
char[] eachs = name.toCharArray();
StringBuffer buffer = null;
if (eachs.length == 1) {
buffer = new StringBuffer();
String[] s = PinYinUtil.toDiffHanyuPinyinStringArray(eachs[0], format);
for (int i = 0; i < s.length; i++) {
buffer.append(s[i]);
if (s.length != 1 && i != s.length - 1) {
buffer.append(";");
}
}
return buffer.toString().replaceAll(" ", "");
} else if (eachs.length == 2) {
buffer = new StringBuffer();
String[] firstNameArrays = PinYinUtil.toDiffHanyuPinyinStringArray(eachs[0], format);
String[] secondNameArrays = PinYinUtil.toDiffHanyuPinyinStringArray(eachs[1], format);
for (int i = 0; i < firstNameArrays.length; ) {
for (int j = 0; j < secondNameArrays.length; j++) {
buffer.append(firstNameArrays[i] + " ");
buffer.append(secondNameArrays[j] + ";");
}
i++;
}
if (buffer.lastIndexOf(";") == buffer.length() - 1) {
return buffer.substring(0, buffer.length() - 1).replaceAll(" ", "");
}
} else if (eachs.length == 3) {
buffer = new StringBuffer();
String[] firstNameArrays = PinYinUtil.toDiffHanyuPinyinStringArray(eachs[0], format);
String[] secondNameArrays = PinYinUtil.toDiffHanyuPinyinStringArray(eachs[1], format);
String[] thirdNameArrays = PinYinUtil.toDiffHanyuPinyinStringArray(eachs[2], format);
for (int i = 0; i < firstNameArrays.length; ) {
for (int j = 0; j < secondNameArrays.length; ) {
for (int k = 0; k < thirdNameArrays.length; k++) {
buffer.append(firstNameArrays[i] + " ");
buffer.append(secondNameArrays[j] + " ");
buffer.append(thirdNameArrays[k]);
if (fourChar != 0) {
if (("" + fourChar).matches("[a-zA-Z]")) {
buffer.append(("" + fourChar).toUpperCase());
} else if ('\u9fa5' < fourChar || '\u4e00' > fourChar) {
buffer.append(("" + fourChar));
} else {
buffer.append(
PinyinHelper.toHanyuPinyinStringArray(fourChar)[0]
.toUpperCase()
.replaceAll("[^a-zA-Z]", ""));
}
}
buffer.append(";");
}
j++;
}
i++;
}
if (buffer.lastIndexOf(";") == buffer.length() - 1) {
return buffer.substring(0, buffer.length() - 1).replaceAll(" ", "");
}
}
return null;
}
/**
* 获取指定汉字的拼音首字母
*
* @param ch 要获取的汉字
* @return 返回汉字首字母，如果非汉字返回null
*/
public static String getHeadChar(char ch) {
String[] pinyins = PinyinHelper.toHanyuPinyinStringArray(ch);
return null == pinyins ? null : String.valueOf(pinyins[0].charAt(0));
}
@Override
public final boolean incrementToken() throws IOException {
clearAttributes();
if (!done) {
done = true;
int upto = 0;
char[] buffer = termAtt.buffer();
while (true) {
final int length = input.read(buffer, upto, buffer.length - upto);
if (length == -1) break;
upto += length;
if (upto == buffer.length) buffer = termAtt.resizeBuffer(1 + buffer.length);
}
termAtt.setLength(upto);
String str = termAtt.toString();
termAtt.setEmpty();
StringBuilder stringBuilder = new StringBuilder();
StringBuilder firstLetters = new StringBuilder();
for (int i = 0; i < str.length(); i++) {
char c = str.charAt(i);
if (c < 128) {
stringBuilder.append(c);
} else {
try {
String[] strs = PinyinHelper.toHanyuPinyinStringArray(c, format);
if (strs != null) {
// get first result by default
String first_value = strs[0];
// TODO more than one pinyin
stringBuilder.append(first_value);
if (this.padding_char.length() > 0) {
stringBuilder.append(this.padding_char);
}
firstLetters.append(first_value.charAt(0));
}
} catch (BadHanyuPinyinOutputFormatCombination badHanyuPinyinOutputFormatCombination) {
badHanyuPinyinOutputFormatCombination.printStackTrace();
}
}
}
// let's join them
if (first_letter.equals("prefix")) {
termAtt.append(firstLetters.toString());
if (this.padding_char.length() > 0) {
termAtt.append(this.padding_char); // TODO splitter
}
termAtt.append(stringBuilder.toString());
} else if (first_letter.equals("append")) {
termAtt.append(stringBuilder.toString());
if (this.padding_char.length() > 0) {
if (!stringBuilder.toString().endsWith(this.padding_char)) {
termAtt.append(this.padding_char);
}
}
termAtt.append(firstLetters.toString());
} else if (first_letter.equals("none")) {
termAtt.append(stringBuilder.toString());
} else if (first_letter.equals("only")) {
termAtt.append(firstLetters.toString());
}
finalOffset = correctOffset(upto);
offsetAtt.setOffset(correctOffset(0), finalOffset);
return true;
}
return false;
}