segment/lib/module/ChsNameOptimizer.js

Chinese word segmentation 中文分词模块

'use strict';

/**
 * 人名优化模块
 *
 * @author 老雷<leizongmin@gmail.com>
 * @version 0.1
 */
 
var FAMILY_NAME_1 = require('./CHS_NAMES').FAMILY_NAME_1; 
var FAMILY_NAME_2 = require('./CHS_NAMES').FAMILY_NAME_2; 
var SINGLE_NAME = require('./CHS_NAMES').SINGLE_NAME;
var DOUBLE_NAME_1 = require('./CHS_NAMES').DOUBLE_NAME_1;
var DOUBLE_NAME_2 = require('./CHS_NAMES').DOUBLE_NAME_2;
var debug = console.log; 

/** 模块类型 */
exports.type = 'optimizer';

/**
 * 模块初始化
 * 
 * @param {Segment} segment 分词接口
 */
exports.init = function (segment) {
  exports.segment = segment;
};

/**
 * 对可能是人名的单词进行优化
 *
 * @param {array} words 单词数组
 * @return {array}
 */
exports.doOptimize = function (words) {
  //debug(words);
  var POSTAG = exports.segment.POSTAG;
  var i = 0;
  
  /* 第一遍扫描 */
  while (i < words.length) {
    var word = words[i];
    var nextword = words[i + 1];
    if (nextword) {
      //debug(nextword);
      // 如果为  "小|老" + 姓
      if (nextword && (word.w == '小' || word.w == '老') &&
      (nextword.w in FAMILY_NAME_1 || nextword.w in FAMILY_NAME_2)) {
        words.splice(i, 2, {
          w:  word.w + nextword.w,
          p:  POSTAG.A_NR
        }); 
        i++;
        continue;
      }
      
      // 如果是 姓 + 名（2字以内）
      if ((word.w in FAMILY_NAME_1 || word.w in FAMILY_NAME_2) &&
      ((nextword.p & POSTAG.A_NR) > 0 && nextword.w.length <= 2)) {
        words.splice(i, 2, {
          w:  word.w + nextword.w,
          p:  POSTAG.A_NR
        }); 
        i++;
        continue;
      }
      
      // 如果相邻两个均为单字且至少有一个字是未识别的，则尝试判断其是否为人名
      if (!word.p || !nextword.p) {
        if ((word.w in SINGLE_NAME && word.w == nextword.w) ||
        (word.w in DOUBLE_NAME_1 && nextword.w in DOUBLE_NAME_2)) {
          words.splice(i, 2, {
            w:  word.w + nextword.w,
            p:  POSTAG.A_NR
          });
          // 如果上一个单词可能是一个姓，则合并
          var preword = words[i - 1];
          if (preword &&
          (preword.w in FAMILY_NAME_1 || preword.w in FAMILY_NAME_2)) {
            words.splice(i - 1, 2, {
              w:  preword.w + word.w + nextword.w,
              p:  POSTAG.A_NR
            });
          } else {
            i++;
          }
          continue;
        }
      }

      // 如果为 无歧义的姓 + 名（2字以内） 且其中一个未未识别词
      if ((word.w in FAMILY_NAME_1 || word.w in FAMILY_NAME_2) &&
          (!word.p || !nextword.p)) {
        //debug(word, nextword);
        words.splice(i, 2, {
          w:  word.w + nextword.w,
          p:  POSTAG.A_NR
        });
      }
    }
    
    // 移到下一个单词
    i++;
  }
  
  /* 第二遍扫描 */
  i = 0;
  while (i < words.length) {
    var word = words[i];
    var nextword = words[i + 1];
    if (nextword) {
      // 如果为 姓 + 单字名
      if ((word.w in FAMILY_NAME_1 || word.w in FAMILY_NAME_2) &&
      nextword.w in SINGLE_NAME) {
        words.splice(i, 2, {
          w:  word.w + nextword.w,
          p:  POSTAG.A_NR
        });
        i++;
        continue;
      }
    }
    
    // 移到下一个单词
    i++;
  }
  
  return words;
};