scws

nodescws

scws

About

scws即Simple Chinese Word Segmentation。是C语言开发的基于词频词典的机械式中文分词引擎。scws的作者为hightman，采用BSD许可协议发布。nodescws的作者在libscws上添加功能(包括停用词、忽略符号、json格式配置等)并添加了node.js binding，除自己代码，不持有libscws著作权。

scws的主页: http://www.xunsearch.com/scws, GitHub: https://github.com/hightman/scws

nodescws

Current release: v0.5.1

• 项目主页: https://github.com/dotSlashLu/nodescws
• 使用问题，bug report: https://github.com/dotSlashLu/nodescws/issues

Install

npm install scws

Usage

var Scws = require("scws");
var scws = new Scws(settings); // NOTE: before v0.5.0, do new Scws.init(settings)
var results = scws.segment(text);
scws.destroy(); // DO NOT forget this or your memory may be corrupted

new Scws(settings)

注意，在v0.5.0之前，使用new Scws.init(settings)初始化。

• settings: Object, 分词设置, 支持charset, dicts, rule, ignorePunct, multi, debug：

  • charset: String, Optional

      采用的encoding，支持"utf8"，"gbk"， 默认值"utf8"
    

  • dicts: String, Required

      要采用的词典文件的filename，多个文件之间用':'分隔。
      支持xdb格式以及txt格式，自制词典请以".txt"作文件后缀。
      例如"./dicts/dict.utf8.xdb:./dicts/dict_cht.utf8.xdb:./dicts/dict.test.txt"
      scws自带的xdb格式词典附在该extension目录下(一般是node_modules/scws/)的./dicts/ ，
      有简体和繁体两种选择，如果该项缺失则默认使用自带utf8简体中文词典
    

  • rule: String, Optional

      要采用的规则文件，设置对应编码下的地名，人名，停用词等。
      详见该extension目录下(一般是node_modules/scws/)的rules/rules.utf8.ini。
      若该配置缺失则默认使用自带utf8的规则文件。
    
        v0.2.3添加了JSON支持，避免繁复的ini语法。
        若以.json结尾，则会解析对应的JSON rule文件，也可以直接传JSON string来进行配置。语法参考 ./rules/rules.utf8.json
    

  • ignorePunct: Bool, Optional

      是否忽略标点
    

  • multi: String, Optional

      是否进行长词复合切分，例如中国人这个词产生“中国人”，“中国”，“人”多个结果，可选值"short", "duality", "zmain", "zall":
          short: 短词
          duality: 组合相邻的两个单字
          zmain: 重要单字
          zall: 全部单字
    

  • debug: Bool, Optional

      是否以debug模式运行，若为true则输出scws的log, warning, error到stdout, defult为false
    

  • applyStopWord: Bool, Optional

      是否应用rule文件中[nostats]区块所规定的停用词，默认为true
    

scws.segment(text)

• text: String, 要切分的字符串

Return Array

[
    {
        word: '可读性',
        offset: 183, // 该词在文档中的位置
        length: 9, // byte
        attr: 'n', // 词性，采用《现代汉语语料库加工规范——词语切分与词性标注》标准，涵义请参考 http://blog.csdn.net/dbigbear/article/details/1488087
        idf: 7.800000190734863
    },
    ...
]

Example 用例

var fs   = require("fs")
    Scws = require("scws");

fs.readFile("./test_doc.txt", {
  encoding: "utf8"
}, function(err, data){
  if (err)
    return console.error(err);

  // initialize scws with config entries
  var scws = new Scws({
    charset: "utf8",
    //dicts: "./dicts/dict.utf8.xdb:./dicts/dict_cht.utf8.xdb:./dicts/dict.test.txt",
    dicts: "./dicts/dict.utf8.xdb",
    rule: "./rules/rules.utf8.ini",
    ignorePunct: true,
    multi: "duality",
    debug: true
  });

  // segment text
  res = scws.segment(data);
  res1 = scws.segment("大家好我来自德国，我是德国人");

  console.log(res, res1);

  // destroy scws, recollect memory
  scws.destroy();
})

更多请参考test/中的测试

Changelog

v0.5.1

• fix macOS build issue #18 thanks to agj

v0.5.0

• Update NAN, supports all major node.js versions
• New js API design
• Fix #11

v0.2.4

• Thanks to @mike820324 now scws supports io.js

v0.2.3

• Changed project structure
• Refactored node bindings
• Added rule setting by JSON file and JSON string thus making adding stop words more easier with node

v0.2.2

• Some small bug fixes, including issue #5(Thanks to @Frully)

v0.2.1

• Add stop words support
• Remove line endings when ignorePunct is set true

You can add your own stop words in the entry [nostats] in the rule file. Turn off stop words feature by setting applyStopWord false.

v0.2.0

New syntax to initialize scws: scws = new Scws(config); result = scws.segment(text); scws.destroy() so that we are able to reuse scws instance, thus gaining great improvement in perfermence when recurrently used(approximately 1/4 faster).

Added new setting entry debug. Setting config.debug = true will make scws output it's log, error, warning to stdout

v0.1.3

Published to npm registry. usage: scws(text, settings); available setting entries: charset, dicts, rule, ignorePunct, multi.

Contributors

• @mike820324