#写了一个用nodejs来拆词和问答的模块,跪求github点星，感谢好心人 #大哥们，谢谢你们了，跪求点赞，下面是链接 #https://github.com/zy445566/CustomerService #为了支持async和await，我这个也建议直接上node的7.7.1以上版本

CustomerService

Question & Answer'Ai

中文文档
english
API

中文文档

介绍
要求
安装
使用

介绍

CustomerService是一个可以从句子里面拆出词汇,并且可以实现问题和回答的相似度匹配的AI系统. 理论上只要词库(中文已有部分)和问答库(未导入)足够的大,就可以实现非常完美的问答.Power By NodeJs.

要求

由于使用了async和await,所以nodejs版本要在7.7.1以上.或使用babel.

安装

npm install customer-service

使用

初始化工具

const CustomerService = require("customer-service");
var customerService = new CustomerService('mylang',2,13);
var sw = customerService.getSplitWord();
var ie = customerService.getImportExport();
var qa = customerService.getQuestionAnswer();

拆词

sw.collisionWord(sw.sentenceToList('我在洗澡，你在干嘛'))
.then((res)=>{
console.log(res);
/* will print
[ { key: 'word:b7669c4218782c0035b6383623e19b29',
    word: '我在',
    questionList: {} },
  { key: 'word:5ac500cfcc1fe66f7243cad4039281d1',
    word: '洗澡',
    questionList: {} },
  { key: 'word:9d63b7094c5a502e66fccc79e5fe1f69',
    word: '你在',
    questionList: {} },
  { key: 'word:c44a42f209fa03606e607994c2321ebd',
    word: '干嘛',
    questionList: {} } ]
*/
});

库的导入和打印

//导入词库字符串
var chunk = "你好\r\n智能";
ie.stringToLevel(chunk,'word')
.then((res)=>{
	console.log(res);
})
.catch((err)=>{
	console.log(err);
});

//导入问答字符串
var chunk = "你在洗澡吗？===>>>然而并没有.\r\n你是屌丝吗？===>>>不是屌丝写什么代码.";
ie.stringToLevel(chunk,'word')
.then((res)=>{
	console.log(res);
})
.catch((err)=>{
	console.log(err);
});

//导入问答字符串

const path = require('path');
//导入词库
ie.readWordToLevel(path.join(__dirname,'word.log'),'word')
.then((res)=>{
	console.log(res);
})
.catch((err)=>{
	console.log(err);
});
//导入问答
ie.readWordToLevel(path.join(__dirname,'qa.log'),'qa')
.then((res)=>{
	console.log(res);
})
.catch((err)=>{
	console.log(err);
});
//print data
ie.printData({start:'word',limit:10});

获取回答

//获取最佳回答
qa.getAnswer('你是屌丝吗')
.then((res)=>{
	console.log(res);
})
.catch((err)=>{
	console.log(err);
});

Document

intro
req
install
use

intro

CustomerService is a down words from a sentence, and can realize question and answer the similarity matching of AI system. Theoretically as long as the word library (Chinese existing part) and q&a library (import) not enough big, can achieve very perfect answers. The Power By NodeJs.

req

nodejs'version>7.7.1 or use babel.

install

npm install customer-service

use

init

const CustomerService = require("customer-service");
var customerService = new CustomerService('mylang',2,13);
var sw = customerService.getSplitWord();
var ie = customerService.getImportExport();
var qa = customerService.getQuestionAnswer();

split word

sw.collisionWord(sw.sentenceToList('我在洗澡，你在干嘛'))
.then((res)=>{
console.log(res);
/* will print
[ { key: 'word:b7669c4218782c0035b6383623e19b29',
    word: '我在',
    questionList: {} },
  { key: 'word:5ac500cfcc1fe66f7243cad4039281d1',
    word: '洗澡',
    questionList: {} },
  { key: 'word:9d63b7094c5a502e66fccc79e5fe1f69',
    word: '你在',
    questionList: {} },
  { key: 'word:c44a42f209fa03606e607994c2321ebd',
    word: '干嘛',
    questionList: {} } ]
*/
});

lib import or print

//import word by string
var chunk = "你好\r\n智能";
ie.stringToLevel(chunk,'word')
.then((res)=>{
	console.log(res);
})
.catch((err)=>{
	console.log(err);
});

//import q&a by string
var chunk = "你在洗澡吗？===>>>然而并没有.\r\n你是屌丝吗？===>>>不是屌丝写什么代码.";
ie.stringToLevel(chunk,'word')
.then((res)=>{
	console.log(res);
})
.catch((err)=>{
	console.log(err);
});

//import word by file

const path = require('path');
//导入词库
ie.readWordToLevel(path.join(__dirname,'word.log'),'word')
.then((res)=>{
	console.log(res);
})
.catch((err)=>{
	console.log(err);
});

//import q&a by string

ie.readWordToLevel(path.join(__dirname,'qa.log'),'qa')
.then((res)=>{
	console.log(res);
})
.catch((err)=>{
	console.log(err);
});

//print data
ie.printData({start:'word',limit:10});

get answer

//get best answer
qa.getAnswer('你是屌丝吗')
.then((res)=>{
	console.log(res);
})
.catch((err)=>{
	console.log(err);
});

API

CustomerService
SplitWord
ImportExport
QuestionAnswer