Keyword Correlation Model (KCM 關鍵字相關性模型)

KCM API are also available now：API Online Version

不知道要寫啥
Now three languages are available：

• Chinese
• English

Getting Started

These instructions will get you a copy of the project up and running on your local machine for development and testing purposes. See deployment for notes on how to deploy the project on a live system.

Prerequisities

1. OS：Ubuntu / OSX would be nice
2. environment：need python3 sudo apt-get update; sudo apt-get install; python3 python3-dev
3. (optional) 簡繁轉換：sudo apt-get install -y opencc
4. (optional) 英文版Wiki，請先下載nltk：python3 -m nltk.downloader -d /usr/local/share/nltk_data all
5. (optional) 用各論壇資料建立KCM模型：請先參照這篇產出輸入檔KCM Data Source Extractor

Installing

• Install By Docker：
  1. You can only run this command in directory which has Dockerfile：sudo docker build -t kcm .
• Install manually：
  1. 使用虛擬環境 Use virtualenv is recommended：
  2. virtualenv venv
  3. 啟動方法 How to activate virtualenv 1. for Linux：. venv/bin/activate 2. for Windows：venv\Scripts\activate
  4. 安裝 Install：pip install -e git://github.com/UDICatNCHU/KCM.git@master#egg=KCM

Running & Testing

Run

• 整體使用方法 Usage：

  • KCM物件的參數：
    • lang：cht 中文、英文 eng
    • io_dir：讀取建立model用的input檔的路徑
    • thread_count：輸入整數，為啟動的執行緒數量
    • uri：Mongo連接網址，mongodb://資料庫IP
  • Instance Method：
    • main：建立模型的函式
    • setLang：選擇語系的函式，build會根據設定的語言到相對應的資料夾讀取input檔
    • removeDB：清空MongoDB
    • get：到MongoDB查詢結果  * 執行方法（二選一）：
    1. command line執行：python3 src/kcm/KCM/__main__.py(或是你放kcm/__main__.py的路徑) -p 你放要建立模型輸入檔的地方

    from KCM.__main__ import KCM
    k = KCM('cht', '含有輸入檔的資料夾路徑')
    k.removeDB()
    k.main()
    
    print(k.get('美國隊長', 10))
    

Results

the result of querying KCM model with keyword "Captain America"：

[
  ["電影",93],
  ["復仇者",78],
  ["戰士",55],
  ["英雄",55],
  ["鋼鐵",52],
  ["內戰",50],
  ["復仇者聯盟",44],
  ["漫畫",42],
  ["酷寒",39],
  ["巴基",36]
]

Break down into end to end tests

1. 執行全部的測試 run whole test：python3 run_tests.py

And coding style tests

目前沒有coding style tests...

Deployment

目前只是一般的python程式，pip install 即可

Built With

• python3.5

Versioning

For the versions available, see the tags on this repository.

Contributors

• 范耀中 教授
• 黃思穎
• 陳聖軒
• Yen-Ju Lee
• 張泰瑋 david

License

Acknowledgments

• WikiExtractor
• 感謝fxsjy的jieba斷詞系統 Thanks python library jieba from fxsjy
• 感謝Google工程師釋出的word2vec Thanks word2vec from Google