Pinyin Input Method Engine

汉语拼音输入法引擎 JavaScript 实现。

快速使用

安装

npm install pinyin-input-method-engine

拼音音节拆分

import { splitAsYinJie } from "pinyin-input-method-engine";
const res = splitAsYinJie("tichulebucuodejianyi");

构建拼音查询算法实例

import {
  DirectedAcyclicGraph,
  HiddenMarkovModel,
} from "pinyin-input-method-engine";
// 为了增加灵活性，使逻辑控制与数据分离，字典数据文件在根路径下的 dict 目录中。
// 可通过 fetch 静态资源的方式加载文件，并通过 JSON.parse() 方法进行解析；
// 或者通过 Typescript "resolveJsonModule": true 直接引用。
import charDict from "pinyin-input-method-engine/dict/dag_char.json";
import phraseDict from "pinyin-input-method-engine/dict/dag_phrase.json";
import pinyinDict from "pinyin-input-method-engine/dict/hmm_py2hz.json";
import startDict from "pinyin-input-method-engine/dict/hmm_start.json";
import emissionDict from "pinyin-input-method-engine/dict/hmm_emission.json";
import transitionDict from "pinyin-input-method-engine/dict/hmm_transition.json";

const dag = new DirectedAcyclicGraph(
  charDict as unknown as TDagDict,
  phraseDict as unknown as TDagDict
);

const hmm = new HiddenMarkovModel(
  pinyinDict as unknown as TPinyinDict,
  startDict as unknown as IHmmStartDict,
  emissionDict as unknown as IHmmDict,
  transitionDict as unknown as IHmmDict
);

使用 DAG 查询

// 不限制查询数量
const res = dag.query({
  yinJieList: ['ni','hao'];
});

// 制定最大查询数量
const res1 = dag.query({
  yinJieList: ['ni','hao'];
  maxNum: 1
});

使用 HMM 查询

// 不限制查询数量
const res = hmm.query({
  yinJieList: ['ni','hao'];
});

// 制定最大查询数量
const res1 = hmm.query({
  yinJieList: ['ni','hao'];
  maxNum: 1
});

测试

npm run test

构建

npm run build

参考引用

本库代码实现参考以下文章及代码实现，对作者的辛勤付出和开源精神表示万分感谢。

1. 如何实现拼音与汉字的互相转换.
2. 拼音转汉字（拼音输入法引擎）.
3. 汉字转拼音.