使用Node.js實作企業聊天機器人後端系統

現今企業為了減少客服人力，或是減輕內部人員繁忙的工作，許多企業都開始將自己的手機應用程式(APP)或是官方網站嵌入聊天機器人，所以本論文介紹聊天機器人的後端系統程式設計。本論文使用IBM Watson Assistant系統來設計聊天機器人的對話流程。後端伺服器的部分會使用Node.js去處理連接Watson Assistant、連接Line、實作登入驗證、以及審核機制等等。資料庫的部分是使用Microsoft SQL Server去儲存資料。整個後端的伺服器會使用Model-View-Controller(MVC)的架構去撰寫，目的是讓每個不同的程式有專屬的放置位置，以後在維護上也會比較方便。登入驗證會讓使用者輸入信箱，然後後端就會判斷有沒有登入。如果沒有登入就會使用信箱傳送驗證碼去進行登入驗證，直到使用者登入後才會去做後續的對話流程。對話的方面會讓使用者使用Line去跟聊天機器人去做對話。本論文提供使用者兩種系統：「請假系統」、「借會議室系統」可供操作。操作完也可以去查看紀錄，審核的人員也可以去進行審核的動作，審核完之後紀錄也會跟著作變更。

關鍵字

聊天機器人； Node.js ； Microsoft SQL Server ； IBM Watson Assistant ； Line ； Model-View-Controller (MVC)架構

並列摘要

Nowadays, in order to lower the size of the help desk workforce, or to reduce the work loading for company workers, many companies and enterprises are starting to embed Chatbot into their mobile phone apps or webpages. In this thesis, we will discuss the design and programming of the backend system supporting such Chatbot. We will use IBM Watson Assistant to design the dialog flows of Chatbot. The backend system we develop will be developed using Node.js. Its purpose is to provide a platform that we can use to connect to Watson Assistant, to LINE, and allow us to implement authentication and workflow procedures. We will use Microsoft SQL Server to store all the data of the backend systems. The backend system will be based on the Model-View-Controller (MVC) architecture. The goal is to ensure every piece of code will have a specific storage location and future maintenance will be as simple as possible. The authentication process will start by asking the user to provide an email. Then the backend system will check whether the user has logged in. If the user has not logged in, then a verification code will be sent to the user for verification purposes. Dialogs will begin after the authentication process is completed. In the dialog part, our design will let the user use LINE to communicate with Watson Assistant. We will have two workflow systems, namely the “Leave or absent” and “Reserve meeting room” system available as our examples. After a user finishing these functionalities, the user can also use the Chatbot to query the history of usage. Auditors can use the Chatbot to audit requests and check on the auditing records.