通讯录姓氏多音字排序问题

通讯录姓氏多音字的问题解决
版权声明:本文为博主原创文章,遵循 CC 4.0 by-sa 版权协议,转载请附上原文出处链接和本声明。
本文链接:https://blog.csdn.net/shan286/article/details/79400558
之前做了通讯录的功能,发现多音字的姓氏在错误的字母下面,比如曾(zeng)在C字母下面,所以自己在原来的代码上做了优化,废话不多说,我这里也只是记录一下,所以直接贴代码。

package com.sherry.contact.util;

import android.text.TextUtils;
import android.util.Log;

import java.text.Collator;
import java.util.ArrayList;
import java.util.HashMap;
import java.util.Locale;
import java.util.Map;

/**

  • Created by shanxs on 2018/2/28.
    */

public class HanziNameToPinyin {

private static final String TAG = "HanziNameToPinyin";

// Turn on this flag when we want to check internal data structure.
private static final boolean DEBUG = false;

/**
 * Unihans array.
 *
 * Each unihans is the first one within same pinyin when collator is zh_CN.
 */
public static final char[] UNIHANS = { '\u963f', '\u54ce', '\u5b89',
        '\u80ae', '\u51f9', '\u516b', '\u6300', '\u6273', '\u90a6',
        '\u52f9', '\u9642', '\u5954', '\u4f3b', '\u5c44', '\u8fb9',
        '\u706c', '\u618b', '\u6c43', '\u51ab', '\u7676', '\u5cec',
        '\u5693', '\u5072', '\u53c2', '\u4ed3', '\u64a1', '\u518a',
        '\u5d7e', '\u66fd', '\u66fe', '\u5c64', '\u53c9', '\u8286',
        '\u8fbf', '\u4f25', '\u6284', '\u8f66', '\u62bb', '\u6c88',
        '\u6c89', '\u9637', '\u5403', '\u5145', '\u62bd', '\u51fa',
        '\u6b3b', '\u63e3', '\u5ddb', '\u5205', '\u5439', '\u65fe',
        '\u9034', '\u5472', '\u5306', '\u51d1', '\u7c97', '\u6c46',
        '\u5d14', '\u90a8', '\u6413', '\u5491', '\u5446', '\u4e39',
        '\u5f53', '\u5200', '\u561a', '\u6265', '\u706f', '\u6c10',
        '\u55f2', '\u7538', '\u5201', '\u7239', '\u4e01', '\u4e1f',
        '\u4e1c', '\u543a', '\u53be', '\u8011', '\u8968', '\u5428',
        '\u591a', '\u59b8', '\u8bf6', '\u5940', '\u97a5', '\u513f',
        '\u53d1', '\u5e06', '\u531a', '\u98de', '\u5206', '\u4e30',
        '\u8985', '\u4ecf', '\u7d11', '\u4f15', '\u65ee', '\u4f85',
        '\u7518', '\u5188', '\u768b', '\u6208', '\u7ed9', '\u6839',
        '\u522f', '\u5de5', '\u52fe', '\u4f30', '\u74dc', '\u4e56',
        '\u5173', '\u5149', '\u5f52', '\u4e28', '\u5459', '\u54c8',
        '\u548d', '\u4f44', '\u592f', '\u8320', '\u8bc3', '\u9ed2',
        '\u62eb', '\u4ea8', '\u5677', '\u53ff', '\u9f41', '\u4e6f',
        '\u82b1', '\u6000', '\u72bf', '\u5ddf', '\u7070', '\u660f',
        '\u5419', '\u4e0c', '\u52a0', '\u620b', '\u6c5f', '\u827d',
        '\u9636', '\u5dfe', '\u5755', '\u5182', '\u4e29', '\u51e5',
        '\u59e2', '\u5658', '\u519b', '\u5494', '\u5f00', '\u520a',
        '\u5ffc', '\u5c3b', '\u533c', '\u808e', '\u52a5', '\u7a7a',
        '\u62a0', '\u625d', '\u5938', '\u84af', '\u5bbd', '\u5321',
        '\u4e8f', '\u5764', '\u6269', '\u5783', '\u6765', '\u5170',
        '\u5577', '\u635e', '\u808b', '\u52d2', '\u5d1a', '\u5215',
        '\u4fe9', '\u5941', '\u826f', '\u64a9', '\u5217', '\u62ce',
        '\u5222', '\u6e9c', '\u56d6', '\u9f99', '\u779c', '\u565c',
        '\u5a08', '\u7567', '\u62a1', '\u7f57', '\u5463', '\u5988',
        '\u57cb', '\u5ada', '\u7264', '\u732b', '\u4e48', '\u5445',
        '\u95e8', '\u753f', '\u54aa', '\u5b80', '\u55b5', '\u4e5c',
        '\u6c11', '\u540d', '\u8c2c', '\u6478', '\u54de', '\u6bea',
        '\u55ef', '\u62cf', '\u8149', '\u56e1', '\u56d4', '\u5b6c',
        '\u7592', '\u5a1e', '\u6041', '\u80fd', '\u59ae', '\u62c8',
        '\u5b22', '\u9e1f', '\u634f', '\u56dc', '\u5b81', '\u599e',
        '\u519c', '\u7fba', '\u5974', '\u597b', '\u759f', '\u9ec1',
        '\u90cd', '\u5594', '\u8bb4', '\u5991', '\u62cd', '\u7705',
        '\u4e53', '\u629b', '\u5478', '\u55b7', '\u5309', '\u4e15',
        '\u56e8', '\u527d', '\u6c15', '\u59d8', '\u4e52', '\u948b',
        '\u5256', '\u4ec6', '\u4e03', '\u6390', '\u5343', '\u545b',
        '\u6084', '\u767f', '\u4eb2', '\u72c5', '\u828e', '\u4e18',
        '\u533a', '\u5cd1', '\u7f3a', '\u590b', '\u5465', '\u7a63',
        '\u5a06', '\u60f9', '\u4eba', '\u6254', '\u65e5', '\u8338',
        '\u53b9', '\u909a', '\u633c', '\u5827', '\u5a51', '\u77a4',
        '\u637c', '\u4ee8', '\u6be2', '\u4e09', '\u6852', '\u63bb',
        '\u95aa', '\u68ee', '\u50e7', '\u6740', '\u7b5b', '\u5c71',
        '\u4f24', '\u5f30', '\u5962', '\u7533', '\u8398', '\u6552',
        '\u5347', '\u5c38', '\u53ce', '\u4e66', '\u5237', '\u8870',
        '\u95e9', '\u53cc', '\u8c01', '\u542e', '\u8bf4', '\u53b6',
        '\u5fea', '\u635c', '\u82cf', '\u72fb', '\u590a', '\u5b59',
        '\u5506', '\u4ed6', '\u56fc', '\u574d', '\u6c64', '\u5932',
        '\u5fd1', '\u71a5', '\u5254', '\u5929', '\u65eb', '\u5e16',
        '\u5385', '\u56f2', '\u5077', '\u51f8', '\u6e4d', '\u63a8',
        '\u541e', '\u4e47', '\u7a75', '\u6b6a', '\u5f2f', '\u5c23',
        '\u5371', '\u6637', '\u7fc1', '\u631d', '\u4e4c', '\u5915',
        '\u8672', '\u4eda', '\u4e61', '\u7071', '\u4e9b', '\u5fc3',
        '\u661f', '\u51f6', '\u4f11', '\u5401', '\u5405', '\u524a',
        '\u5743', '\u4e2b', '\u6079', '\u592e', '\u5e7a', '\u503b',
        '\u4e00', '\u56d9', '\u5e94', '\u54df', '\u4f63', '\u4f18',
        '\u625c', '\u56e6', '\u66f0', '\u6655', '\u7b60', '\u7b7c',
        '\u5e00', '\u707d', '\u5142', '\u5328', '\u50ae', '\u5219',
        '\u8d3c', '\u600e', '\u5897', '\u624e', '\u635a', '\u6cbe',
        '\u5f20', '\u957f', '\u9577', '\u4f4b', '\u8707', '\u8d1e',
        '\u4e89', '\u4e4b', '\u5cd9', '\u5ea2', '\u4e2d', '\u5dde',
        '\u6731', '\u6293', '\u62fd', '\u4e13', '\u5986', '\u96b9',
        '\u5b92', '\u5353', '\u4e72', '\u5b97', '\u90b9', '\u79df',
        '\u94bb', '\u539c', '\u5c0a', '\u6628', '\u5159', '\u9fc3',
        '\u9fc4', };

/**
 * Pinyin array.
 *
 * Each pinyin is corresponding to unihans of same offset in the unihans
 * array.
 */
public static final byte[][] PINYINS = { { 65, 0, 0, 0, 0, 0 },
        { 65, 73, 0, 0, 0, 0 }, { 65, 78, 0, 0, 0, 0 },
        { 65, 78, 71, 0, 0, 0 }, { 65, 79, 0, 0, 0, 0 },
        { 66, 65, 0, 0, 0, 0 }, { 66, 65, 73, 0, 0, 0 },
        { 66, 65, 78, 0, 0, 0 }, { 66, 65, 78, 71, 0, 0 },
        { 66, 65, 79, 0, 0, 0 }, { 66, 69, 73, 0, 0, 0 },
        { 66, 69, 78, 0, 0, 0 }, { 66, 69, 78, 71, 0, 0 },
        { 66, 73, 0, 0, 0, 0 }, { 66, 73, 65, 78, 0, 0 },
        { 66, 73, 65, 79, 0, 0 }, { 66, 73, 69, 0, 0, 0 },
        { 66, 73, 78, 0, 0, 0 }, { 66, 73, 78, 71, 0, 0 },
        { 66, 79, 0, 0, 0, 0 }, { 66, 85, 0, 0, 0, 0 },
        { 67, 65, 0, 0, 0, 0 }, { 67, 65, 73, 0, 0, 0 },
        { 67, 65, 78, 0, 0, 0 }, { 67, 65, 78, 71, 0, 0 },
        { 67, 65, 79, 0, 0, 0 }, { 67, 69, 0, 0, 0, 0 },
        { 67, 69, 78, 0, 0, 0 }, { 67, 69, 78, 71, 0, 0 },
        { 90, 69, 78, 71, 0, 0 }, { 67, 69, 78, 71, 0, 0 },
        { 67, 72, 65, 0, 0, 0 }, { 67, 72, 65, 73, 0, 0 },
        { 67, 72, 65, 78, 0, 0 }, { 67, 72, 65, 78, 71, 0 },
        { 67, 72, 65, 79, 0, 0 }, { 67, 72, 69, 0, 0, 0 },
        { 67, 72, 69, 78, 0, 0 }, { 83, 72, 69, 78, 0, 0 },
        { 67, 72, 69, 78, 0, 0 }, { 67, 72, 69, 78, 71, 0 },
        { 67, 72, 73, 0, 0, 0 }, { 67, 72, 79, 78, 71, 0 },
        { 67, 72, 79, 85, 0, 0 }, { 67, 72, 85, 0, 0, 0 },
        { 67, 72, 85, 65, 0, 0 }, { 67, 72, 85, 65, 73, 0 },
        { 67, 72, 85, 65, 78, 0 }, { 67, 72, 85, 65, 78, 71 },
        { 67, 72, 85, 73, 0, 0 }, { 67, 72, 85, 78, 0, 0 },
        { 67, 72, 85, 79, 0, 0 }, { 67, 73, 0, 0, 0, 0 },
        { 67, 79, 78, 71, 0, 0 }, { 67, 79, 85, 0, 0, 0 },
        { 67, 85, 0, 0, 0, 0 }, { 67, 85, 65, 78, 0, 0 },
        { 67, 85, 73, 0, 0, 0 }, { 67, 85, 78, 0, 0, 0 },
        { 67, 85, 79, 0, 0, 0 }, { 68, 65, 0, 0, 0, 0 },
        { 68, 65, 73, 0, 0, 0 }, { 68, 65, 78, 0, 0, 0 },
        { 68, 65, 78, 71, 0, 0 }, { 68, 65, 79, 0, 0, 0 },
        { 68, 69, 0, 0, 0, 0 }, { 68, 69, 78, 0, 0, 0 },
        { 68, 69, 78, 71, 0, 0 }, { 68, 73, 0, 0, 0, 0 },
        { 68, 73, 65, 0, 0, 0 }, { 68, 73, 65, 78, 0, 0 },
        { 68, 73, 65, 79, 0, 0 }, { 68, 73, 69, 0, 0, 0 },
        { 68, 73, 78, 71, 0, 0 }, { 68, 73, 85, 0, 0, 0 },
        { 68, 79, 78, 71, 0, 0 }, { 68, 79, 85, 0, 0, 0 },
        { 68, 85, 0, 0, 0, 0 }, { 68, 85, 65, 78, 0, 0 },
        { 68, 85, 73, 0, 0, 0 }, { 68, 85, 78, 0, 0, 0 },
        { 68, 85, 79, 0, 0, 0 }, { 69, 0, 0, 0, 0, 0 },
        { 69, 73, 0, 0, 0, 0 }, { 69, 78, 0, 0, 0, 0 },
        { 69, 78, 71, 0, 0, 0 }, { 69, 82, 0, 0, 0, 0 },
        { 70, 65, 0, 0, 0, 0 }, { 70, 65, 78, 0, 0, 0 },
        { 70, 65, 78, 71, 0, 0 }, { 70, 69, 73, 0, 0, 0 },
        { 70, 69, 78, 0, 0, 0 }, { 70, 69, 78, 71, 0, 0 },
        { 70, 73, 65, 79, 0, 0 }, { 70, 79, 0, 0, 0, 0 },
        { 70, 79, 85, 0, 0, 0 }, { 70, 85, 0, 0, 0, 0 },
        { 71, 65, 0, 0, 0, 0 }, { 71, 65, 73, 0, 0, 0 },
        { 71, 65, 78, 0, 0, 0 }, { 71, 65, 78, 71, 0, 0 },
        { 71, 65, 79, 0, 0, 0 }, { 71, 69, 0, 0, 0, 0 },
        { 71, 69, 73, 0, 0, 0 }, { 71, 69, 78, 0, 0, 0 },
        { 71, 69, 78, 71, 0, 0 }, { 71, 79, 78, 71, 0, 0 },
        { 71, 79, 85, 0, 0, 0 }, { 71, 85, 0, 0, 0, 0 },
        { 71, 85, 65, 0, 0, 0 }, { 71, 85, 65, 73, 0, 0 },
        { 71, 85, 65, 78, 0, 0 }, { 71, 85, 65, 78, 71, 0 },
        { 71, 85, 73, 0, 0, 0 }, { 71, 85, 78, 0, 0, 0 },
        { 71, 85, 79, 0, 0, 0 }, { 72, 65, 0, 0, 0, 0 },
        { 72, 65, 73, 0, 0, 0 }, { 72, 65, 78, 0, 0, 0 },
        { 72, 65, 78, 71, 0, 0 }, { 72, 65, 79, 0, 0, 0 },
        { 72, 69, 0, 0, 0, 0 }, { 72, 69, 73, 0, 0, 0 },
        { 72, 69, 78, 0, 0, 0 }, { 72, 69, 78, 71, 0, 0 },
        { 72, 77, 0, 0, 0, 0 }, { 72, 79, 78, 71, 0, 0 },
        { 72, 79, 85, 0, 0, 0 }, { 72, 85, 0, 0, 0, 0 },
        { 72, 85, 65, 0, 0, 0 }, { 72, 85, 65, 73, 0, 0 },
        { 72, 85, 65, 78, 0, 0 }, { 72, 85, 65, 78, 71, 0 },
        { 72, 85, 73, 0, 0, 0 }, { 72, 85, 78, 0, 0, 0 },
        { 72, 85, 79, 0, 0, 0 }, { 74, 73, 0, 0, 0, 0 },
        { 74, 73, 65, 0, 0, 0 }, { 74, 73, 65, 78, 0, 0 },
        { 74, 73, 65, 78, 71, 0 }, { 74, 73, 65, 79, 0, 0 },
        { 74, 73, 69, 0, 0, 0 }, { 74, 73, 78, 0, 0, 0 },
        { 74, 73, 78, 71, 0, 0 }, { 74, 73, 79, 78, 71, 0 },
        { 74, 73, 85, 0, 0, 0 }, { 74, 85, 0, 0, 0, 0 },
        { 74, 85, 65, 78, 0, 0 }, { 74, 85, 69, 0, 0, 0 },
        { 74, 85, 78, 0, 0, 0 }, { 75, 65, 0, 0, 0, 0 },
        { 75, 65, 73, 0, 0, 0 }, { 75, 65, 78, 0, 0, 0 },
        { 75, 65, 78, 71, 0, 0 }, { 75, 65, 79, 0, 0, 0 },
        { 75, 69, 0, 0, 0, 0 }, { 75, 69, 78, 0, 0, 0 },
        { 75, 69, 78, 71, 0, 0 }, { 75, 79, 78, 71, 0, 0 },
        { 75, 79, 85, 0, 0, 0 }, { 75, 85, 0, 0, 0, 0 },
        { 75, 85, 65, 0, 0, 0 }, { 75, 85, 65, 73, 0, 0 },
        { 75, 85, 65, 78, 0, 0 }, { 75, 85, 65, 78, 71, 0 },
        { 75, 85, 73, 0, 0, 0 }, { 75, 85, 78, 0, 0, 0 },
        { 75, 85, 79, 0, 0, 0 }, { 76, 65, 0, 0, 0, 0 },
        { 76, 65, 73, 0, 0, 0 }, { 76, 65, 78, 0, 0, 0 },
        { 76, 65, 78, 71, 0, 0 }, { 76, 65, 79, 0, 0, 0 },
        { 76, 69, 0, 0, 0, 0 }, { 76, 69, 73, 0, 0, 0 },
        { 76, 69, 78, 71, 0, 0 }, { 76, 73, 0, 0, 0, 0 },
        { 76, 73, 65, 0, 0, 0 }, { 76, 73, 65, 78, 0, 0 },
        { 76, 73, 65, 78, 71, 0 }, { 76, 73, 65, 79, 0, 0 },
        { 76, 73, 69, 0, 0, 0 }, { 76, 73, 78, 0, 0, 0 },
        { 76, 73, 78, 71, 0, 0 }, { 76, 73, 85, 0, 0, 0 },
        { 76, 79, 0, 0, 0, 0 }, { 76, 79, 78, 71, 0, 0 },
        { 76, 79, 85, 0, 0, 0 }, { 76, 85, 0, 0, 0, 0 },
        { 76, 85, 65, 78, 0, 0 }, { 76, 85, 69, 0, 0, 0 },
        { 76, 85, 78, 0, 0, 0 }, { 76, 85, 79, 0, 0, 0 },
        { 77, 0, 0, 0, 0, 0 }, { 77, 65, 0, 0, 0, 0 },
        { 77, 65, 73, 0, 0, 0 }, { 77, 65, 78, 0, 0, 0 },
        { 77, 65, 78, 71, 0, 0 }, { 77, 65, 79, 0, 0, 0 },
        { 77, 69, 0, 0, 0, 0 }, { 77, 69, 73, 0, 0, 0 },
        { 77, 69, 78, 0, 0, 0 }, { 77, 69, 78, 71, 0, 0 },
        { 77, 73, 0, 0, 0, 0 }, { 77, 73, 65, 78, 0, 0 },
        { 77, 73, 65, 79, 0, 0 }, { 77, 73, 69, 0, 0, 0 },
        { 77, 73, 78, 0, 0, 0 }, { 77, 73, 78, 71, 0, 0 },
        { 77, 73, 85, 0, 0, 0 }, { 77, 79, 0, 0, 0, 0 },
        { 77, 79, 85, 0, 0, 0 }, { 77, 85, 0, 0, 0, 0 },
        { 78, 0, 0, 0, 0, 0 }, { 78, 65, 0, 0, 0, 0 },
        { 78, 65, 73, 0, 0, 0 }, { 78, 65, 78, 0, 0, 0 },
        { 78, 65, 78, 71, 0, 0 }, { 78, 65, 79, 0, 0, 0 },
        { 78, 69, 0, 0, 0, 0 }, { 78, 69, 73, 0, 0, 0 },
        { 78, 69, 78, 0, 0, 0 }, { 78, 69, 78, 71, 0, 0 },
        { 78, 73, 0, 0, 0, 0 }, { 78, 73, 65, 78, 0, 0 },
        { 78, 73, 65, 78, 71, 0 }, { 78, 73, 65, 79, 0, 0 },
        { 78, 73, 69, 0, 0, 0 }, { 78, 73, 78, 0, 0, 0 },
        { 78, 73, 78, 71, 0, 0 }, { 78, 73, 85, 0, 0, 0 },
        { 78, 79, 78, 71, 0, 0 }, { 78, 79, 85, 0, 0, 0 },
        { 78, 85, 0, 0, 0, 0 }, { 78, 85, 65, 78, 0, 0 },
        { 78, 85, 69, 0, 0, 0 }, { 78, 85, 78, 0, 0, 0 },
        { 78, 85, 79, 0, 0, 0 }, { 79, 0, 0, 0, 0, 0 },
        { 79, 85, 0, 0, 0, 0 }, { 80, 65, 0, 0, 0, 0 },
        { 80, 65, 73, 0, 0, 0 }, { 80, 65, 78, 0, 0, 0 },
        { 80, 65, 78, 71, 0, 0 }, { 80, 65, 79, 0, 0, 0 },
        { 80, 69, 73, 0, 0, 0 }, { 80, 69, 78, 0, 0, 0 },
        { 80, 69, 78, 71, 0, 0 }, { 80, 73, 0, 0, 0, 0 },
        { 80, 73, 65, 78, 0, 0 }, { 80, 73, 65, 79, 0, 0 },
        { 80, 73, 69, 0, 0, 0 }, { 80, 73, 78, 0, 0, 0 },
        { 80, 73, 78, 71, 0, 0 }, { 80, 79, 0, 0, 0, 0 },
        { 80, 79, 85, 0, 0, 0 }, { 80, 85, 0, 0, 0, 0 },
        { 81, 73, 0, 0, 0, 0 }, { 81, 73, 65, 0, 0, 0 },
        { 81, 73, 65, 78, 0, 0 }, { 81, 73, 65, 78, 71, 0 },
        { 81, 73, 65, 79, 0, 0 }, { 81, 73, 69, 0, 0, 0 },
        { 81, 73, 78, 0, 0, 0 }, { 81, 73, 78, 71, 0, 0 },
        { 81, 73, 79, 78, 71, 0 }, { 81, 73, 85, 0, 0, 0 },
        { 81, 85, 0, 0, 0, 0 }, { 81, 85, 65, 78, 0, 0 },
        { 81, 85, 69, 0, 0, 0 }, { 81, 85, 78, 0, 0, 0 },
        { 82, 65, 78, 0, 0, 0 }, { 82, 65, 78, 71, 0, 0 },
        { 82, 65, 79, 0, 0, 0 }, { 82, 69, 0, 0, 0, 0 },
        { 82, 69, 78, 0, 0, 0 }, { 82, 69, 78, 71, 0, 0 },
        { 82, 73, 0, 0, 0, 0 }, { 82, 79, 78, 71, 0, 0 },
        { 82, 79, 85, 0, 0, 0 }, { 82, 85, 0, 0, 0, 0 },
        { 82, 85, 65, 0, 0, 0 }, { 82, 85, 65, 78, 0, 0 },
        { 82, 85, 73, 0, 0, 0 }, { 82, 85, 78, 0, 0, 0 },
        { 82, 85, 79, 0, 0, 0 }, { 83, 65, 0, 0, 0, 0 },
        { 83, 65, 73, 0, 0, 0 }, { 83, 65, 78, 0, 0, 0 },
        { 83, 65, 78, 71, 0, 0 }, { 83, 65, 79, 0, 0, 0 },
        { 83, 69, 0, 0, 0, 0 }, { 83, 69, 78, 0, 0, 0 },
        { 83, 69, 78, 71, 0, 0 }, { 83, 72, 65, 0, 0, 0 },
        { 83, 72, 65, 73, 0, 0 }, { 83, 72, 65, 78, 0, 0 },
        { 83, 72, 65, 78, 71, 0 }, { 83, 72, 65, 79, 0, 0 },
        { 83, 72, 69, 0, 0, 0 }, { 83, 72, 69, 78, 0, 0 },
        { 88, 73, 78, 0, 0, 0 }, { 83, 72, 69, 78, 0, 0 },
        { 83, 72, 69, 78, 71, 0 }, { 83, 72, 73, 0, 0, 0 },
        { 83, 72, 79, 85, 0, 0 }, { 83, 72, 85, 0, 0, 0 },
        { 83, 72, 85, 65, 0, 0 }, { 83, 72, 85, 65, 73, 0 },
        { 83, 72, 85, 65, 78, 0 }, { 83, 72, 85, 65, 78, 71 },
        { 83, 72, 85, 73, 0, 0 }, { 83, 72, 85, 78, 0, 0 },
        { 83, 72, 85, 79, 0, 0 }, { 83, 73, 0, 0, 0, 0 },
        { 83, 79, 78, 71, 0, 0 }, { 83, 79, 85, 0, 0, 0 },
        { 83, 85, 0, 0, 0, 0 }, { 83, 85, 65, 78, 0, 0 },
        { 83, 85, 73, 0, 0, 0 }, { 83, 85, 78, 0, 0, 0 },
        { 83, 85, 79, 0, 0, 0 }, { 84, 65, 0, 0, 0, 0 },
        { 84, 65, 73, 0, 0, 0 }, { 84, 65, 78, 0, 0, 0 },
        { 84, 65, 78, 71, 0, 0 }, { 84, 65, 79, 0, 0, 0 },
        { 84, 69, 0, 0, 0, 0 }, { 84, 69, 78, 71, 0, 0 },
        { 84, 73, 0, 0, 0, 0 }, { 84, 73, 65, 78, 0, 0 },
        { 84, 73, 65, 79, 0, 0 }, { 84, 73, 69, 0, 0, 0 },
        { 84, 73, 78, 71, 0, 0 }, { 84, 79, 78, 71, 0, 0 },
        { 84, 79, 85, 0, 0, 0 }, { 84, 85, 0, 0, 0, 0 },
        { 84, 85, 65, 78, 0, 0 }, { 84, 85, 73, 0, 0, 0 },
        { 84, 85, 78, 0, 0, 0 }, { 84, 85, 79, 0, 0, 0 },
        { 87, 65, 0, 0, 0, 0 }, { 87, 65, 73, 0, 0, 0 },
        { 87, 65, 78, 0, 0, 0 }, { 87, 65, 78, 71, 0, 0 },
        { 87, 69, 73, 0, 0, 0 }, { 87, 69, 78, 0, 0, 0 },
        { 87, 69, 78, 71, 0, 0 }, { 87, 79, 0, 0, 0, 0 },
        { 87, 85, 0, 0, 0, 0 }, { 88, 73, 0, 0, 0, 0 },
        { 88, 73, 65, 0, 0, 0 }, { 88, 73, 65, 78, 0, 0 },
        { 88, 73, 65, 78, 71, 0 }, { 88, 73, 65, 79, 0, 0 },
        { 88, 73, 69, 0, 0, 0 }, { 88, 73, 78, 0, 0, 0 },
        { 88, 73, 78, 71, 0, 0 }, { 88, 73, 79, 78, 71, 0 },
        { 88, 73, 85, 0, 0, 0 }, { 88, 85, 0, 0, 0, 0 },
        { 88, 85, 65, 78, 0, 0 }, { 88, 85, 69, 0, 0, 0 },
        { 88, 85, 78, 0, 0, 0 }, { 89, 65, 0, 0, 0, 0 },
        { 89, 65, 78, 0, 0, 0 }, { 89, 65, 78, 71, 0, 0 },
        { 89, 65, 79, 0, 0, 0 }, { 89, 69, 0, 0, 0, 0 },
        { 89, 73, 0, 0, 0, 0 }, { 89, 73, 78, 0, 0, 0 },
        { 89, 73, 78, 71, 0, 0 }, { 89, 79, 0, 0, 0, 0 },
        { 89, 79, 78, 71, 0, 0 }, { 89, 79, 85, 0, 0, 0 },
        { 89, 85, 0, 0, 0, 0 }, { 89, 85, 65, 78, 0, 0 },
        { 89, 85, 69, 0, 0, 0 }, { 89, 85, 78, 0, 0, 0 },
        { 74, 85, 78, 0, 0, 0 }, { 89, 85, 78, 0, 0, 0 },
        { 90, 65, 0, 0, 0, 0 }, { 90, 65, 73, 0, 0, 0 },
        { 90, 65, 78, 0, 0, 0 }, { 90, 65, 78, 71, 0, 0 },
        { 90, 65, 79, 0, 0, 0 }, { 90, 69, 0, 0, 0, 0 },
        { 90, 69, 73, 0, 0, 0 }, { 90, 69, 78, 0, 0, 0 },
        { 90, 69, 78, 71, 0, 0 }, { 90, 72, 65, 0, 0, 0 },
        { 90, 72, 65, 73, 0, 0 }, { 90, 72, 65, 78, 0, 0 },
        { 90, 72, 65, 78, 71, 0 }, { 67, 72, 65, 78, 71, 0 },
        { 90, 72, 65, 78, 71, 0 }, { 90, 72, 65, 79, 0, 0 },
        { 90, 72, 69, 0, 0, 0 }, { 90, 72, 69, 78, 0, 0 },
        { 90, 72, 69, 78, 71, 0 }, { 90, 72, 73, 0, 0, 0 },
        { 83, 72, 73, 0, 0, 0 }, { 90, 72, 73, 0, 0, 0 },
        { 90, 72, 79, 78, 71, 0 }, { 90, 72, 79, 85, 0, 0 },
        { 90, 72, 85, 0, 0, 0 }, { 90, 72, 85, 65, 0, 0 },
        { 90, 72, 85, 65, 73, 0 }, { 90, 72, 85, 65, 78, 0 },
        { 90, 72, 85, 65, 78, 71 }, { 90, 72, 85, 73, 0, 0 },
        { 90, 72, 85, 78, 0, 0 }, { 90, 72, 85, 79, 0, 0 },
        { 90, 73, 0, 0, 0, 0 }, { 90, 79, 78, 71, 0, 0 },
        { 90, 79, 85, 0, 0, 0 }, { 90, 85, 0, 0, 0, 0 },
        { 90, 85, 65, 78, 0, 0 }, { 90, 85, 73, 0, 0, 0 },
        { 90, 85, 78, 0, 0, 0 }, { 90, 85, 79, 0, 0, 0 },
        { 0, 0, 0, 0, 0, 0 }, { 83, 72, 65, 78, 0, 0 },
        { 0, 0, 0, 0, 0, 0 }, };

/**
 * 姓氏的多音字处理
 */
public static Map<String, Character> specialHanzi = new HashMap<>();

static {

    specialHanzi.put("8d3e", '\u7532'); // 贾--甲

    specialHanzi.put("5355", '\u5584'); // 单--善

    specialHanzi.put("6c88", '\u6df1'); // 沈--深

    specialHanzi.put("4ec7", '\u6c42'); // 仇--求

    specialHanzi.put("89e3", '\u8c22'); // 解--谢

    specialHanzi.put("7fdf", '\u5b85'); // 翟--宅

    specialHanzi.put("67e5", '\u4e4d'); // 查--乍

    specialHanzi.put("66fe", '\u589e'); // 曾--增

    specialHanzi.put("665f", '\u6210'); // 晟--成

    specialHanzi.put("4e50", '\u6708'); // 乐--月

    specialHanzi.put("533a", '\u6b27'); // 区--欧

    specialHanzi.put("51af", '\u9022'); // 冯--逢

    specialHanzi.put("7e41", '\u5a46'); // 繁--婆

    specialHanzi.put("957f", '\u8d26'); // 长--帐

    specialHanzi.put("77f3", '\u65f6'); // 石--时

    specialHanzi.put("67cf", '\u6446'); // 柏--摆

    specialHanzi.put("6734", '\u666e'); // 朴--普

    specialHanzi.put("7f2a", '\u79d2'); // 缪--秒

}

/**
 * First and last Chinese character with known Pinyin according to zh
 * collation
 */
private static final String FIRST_PINYIN_UNIHAN = "\u963F";
private static final String LAST_PINYIN_UNIHAN = "\u9FFF";

private static final Collator COLLATOR = Collator.getInstance(Locale.CHINA);

private static HanziNameToPinyin sInstance;
private final boolean mHasChinaCollator;

public static class Token {
    /**
     * Separator between target string for each source char
     */
    public static final String SEPARATOR = " ";

    public static final int LATIN = 1;
    public static final int PINYIN = 2;
    public static final int UNKNOWN = 3;

    public Token() {
    }

    public Token(int type, String source, String target) {
        this.type = type;
        this.source = source;
        this.target = target;
    }

    /**
     * Type of this token, ASCII, PINYIN or UNKNOWN.
     */
    public int type;
    /**
     * Original string before translation.
     */
    public String source;
    /**
     * Translated string of source. For Han, target is corresponding Pinyin.
     * Otherwise target is original string in source.
     */
    public String target;
}

protected HanziNameToPinyin(boolean hasChinaCollator) {
    mHasChinaCollator = hasChinaCollator;
}

public static HanziNameToPinyin getInstance() {
    synchronized (HanziNameToPinyin.class) {
        if (sInstance != null) {
            return sInstance;
        }
        // Check if zh_CN collation data is available
        final Locale locale[] = Collator.getAvailableLocales();

        // 增加的代码,增强。
        final Locale chinaAddition = new Locale("zh");

        for (int i = 0; i < locale.length; i++) {
            if (locale[i].equals(Locale.CHINA)
                    || locale[i].equals(chinaAddition)) {
                // Do self validation just once.
                if (DEBUG) {
                    Log.d(TAG, "Self validation. Result: "
                            + doSelfValidation());
                }
                sInstance = new HanziNameToPinyin(true);
                return sInstance;
            }
        }
        Log.w(TAG,
                "There is no Chinese collator, HanziToPinyin is disabled");
        sInstance = new HanziNameToPinyin(false);
        return sInstance;
    }
}

/**
 * Validate if our internal table has some wrong value.
 *
 * @return true when the table looks correct.
 */
private static boolean doSelfValidation() {
    char lastChar = UNIHANS[0];
    String lastString = Character.toString(lastChar);
    for (char c : UNIHANS) {
        if (lastChar == c) {
            continue;
        }
        final String curString = Character.toString(c);
        int cmp = COLLATOR.compare(lastString, curString);
        if (cmp >= 0) {
            Log.e(TAG, "Internal error in Unihan table. "
                    + "The last string \"" + lastString
                    + "\" is greater than current string \"" + curString
                    + "\".");
            return false;
        }
        lastString = curString;
    }
    return true;
}

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 {
        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();
    if (TextUtils.isEmpty(token.target)) {
        token.type = Token.UNKNOWN;
        token.target = token.source;
    }
    return token;
}

/**
 * Convert the input to a array of tokens. The sequence of ASCII or Unknown
 * characters without space will be put into a Token, One Hanzi character
 * which has pinyin will be treated as a Token. If these is no China
 * collator, the empty token array is returned.
 */
public ArrayList<Token> get(final String input) {
    ArrayList<Token> tokens = new ArrayList<Token>();
    if (!mHasChinaCollator || TextUtils.isEmpty(input)) {
        // return empty tokens.
        return tokens;
    }
    final int inputLength = input.length();
    final StringBuilder sb = new StringBuilder();
    int tokenType = Token.LATIN;
    // Go through the input, create a new token when
    // a. Token type changed
    // b. Get the Pinyin of current charater.
    // c. current character is space.
    for (int i = 0; i < inputLength; i++) {
        char character = input.charAt(i);
        if (character == ' ') {
            if (sb.length() > 0) {
                addToken(sb, tokens, tokenType);
            }
        } else if (character < 256) {
            if (tokenType != Token.LATIN && sb.length() > 0) {
                addToken(sb, tokens, tokenType);
            }
            tokenType = Token.LATIN;
            sb.append(character);
        } else {
            if (i == 0) {
                // 如果是名字的第一个字就进行多音字姓氏判断,也就这句代码关键
                if (specialHanzi.get(Integer.toHexString(character)) != null) {
                    character = specialHanzi.get(Integer.toHexString(character));
                }
            }
            Token t = getToken(character);
            if (t.type == Token.PINYIN) {
                if (sb.length() > 0) {
                    addToken(sb, tokens, tokenType);
                }
                tokens.add(t);
                tokenType = Token.PINYIN;
            } else {
                if (tokenType != t.type && sb.length() > 0) {
                    addToken(sb, tokens, tokenType);
                }
                tokenType = t.type;
                sb.append(character);
            }
        }
    }
    if (sb.length() > 0) {
        addToken(sb, tokens, tokenType);
    }
    return tokens;
}

private void addToken(final StringBuilder sb,
                      final ArrayList<Token> tokens, final int tokenType) {
    String str = sb.toString();
    tokens.add(new Token(tokenType, str, str));
    sb.setLength(0);
}

/**
 *  输入汉字返回拼音
 */
public static String getPinYin(String hanzi) {
    ArrayList<Token> tokens = HanziNameToPinyin.getInstance().get(hanzi);
    StringBuilder sb = new StringBuilder();
    if (tokens != null && tokens.size() > 0) {
        for (Token token : tokens) {
            if (Token.PINYIN == token.type) {
                sb.append(token.target);
            } else {
                sb.append(token.source);
            }
        }
    }
    return sb.toString();
}

}

  • 0
    点赞
  • 1
    收藏
    觉得还不错? 一键收藏
  • 0
    评论

“相关推荐”对你有帮助么?

  • 非常没帮助
  • 没帮助
  • 一般
  • 有帮助
  • 非常有帮助
提交
评论
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值