nlp_sf

natural language processing

description

the Natural Language Processing tools under the nodeJs running time, including Chinese word segmentation, sentiment analysis and so on.

simple to use

let {oops} = require('nlp_sf');
console.log(oops.divide('这是待分词文本，调用oops接口直接分词并且统计词频。'));

installation

npm install nlp_sf

file structure

|
|
+--lib----系统支持库
|
|
+--assets----测试文件
|
|
+--examples----测试程序
|
|
+--index.js----程序入口
|
|
+--package.json
|
|
+--readme.md

api

• nlpir
• oops

nlpir

description

the native Natural Language Processing tools.

loading
first you need to load nlpir. this is the preferred method:

let nlp = require('nlp_sf');
let {nlpir} = nlp;

optionally, you also can load nlpir by the method:

let {nlpir} = require('nlp_sf');

.init()

nlpir.init();

.exit()

nlpir.exit();

oops

description

the interface of the Natural Language Processing.

loading
first you need to load oops. this is the preferred method:

let nlp = require('nlp_sf');
let {oops} = nlp;

optionally, you also can load oops by the method:

let {oops} = require('nlp_sf');

.divide(params, error)
simply, method to divide a string and return array consist of the top 100 words sort by frequency.
default error = false, not throw error.

let arr = oops.divide('这是待分词文本，调用oops接口直接分词并且统计词频。');

custom divide params:
default

dict = {},
query = {
  top: 100,
  freq: 0,
  len: 1,
  join: false,
  tags: ['n', 'a']
}

set top is -999, export all keywords
text is either String or array
tags

let path = require('path');
let text = '这是待分词文本，调用oops接口直接分词并且统计词频。';
let userDict = '../assets/userDict.txt';
userDict = path.join(__dirname, userDict);
let dict = {
  userDict: userDict,
  userWord: ['两国关系 n']
};
let query = {
  top: 20,
  freq: 0,
  len: 1,
  join: true,
  tags: ['n']
};
let params = {
  text: text,
  dict: dict,
  query: query
}
let arr = oops.divide(params);

.sense(params)
simply, method to sense a string and return object consist of positive point and negative point.

let arr = oops.sense('这是待情感分析文本，调用oops接口分析并且返回情感分析结果。');

custom sense params:

let path = require('path');
let title = '情感分析';
let content = '这是待情感分析文本，调用oops接口分析并且返回情感分析结果。';
let target = '情感';
let userDict = '../assets/userDict.txt';
userDict = path.join(__dirname, userDict);
let params = {
  content: content,
  target: target,
  title: title,
  userDict: title
}
let obj = oops.sense(params);

.distinguish(array)
first update dict, second use distinguish
folder /lib/lexical store .yml dict, updateDict read the .yml dict , and update .txt dict store /lib/dict folder

let arr = oops.divide(params);
oops.updateDict();
console.log(oops.distinguish(arr));

.divdis(params)
divide + distinguish
first update dict, second use divdis
folder /lib/lexical store .yml dict, updateDict read the .yml dict , and update .txt dict store /lib/dict folder

oops.updateDict();
console.log(oops.divdis(arr));