pangilinan-patrick / mydbuddy-channel-web

🤖 Repository for the MyDBuddy Tech Support Chatbot built with Botpress for the De La Salle Lipa MyDCampus Portal

🤖 MyDBuddy Channel-Web/Examples Directory

This is the repository for the MyDCampus Botpress Chatbot Channel-Web Examples directory.

The Channel-Web module in Botpress enables the integration of the chat bot into the website or web application, allowing users to interact with the bot through a chat widget.

The Channel-Web module also includes features such as message history, user authentication, and custom styling options (which is extensively used in this project) for the chat widget.

A repository specifically for the examples directory in this module was created because the source control within the Botpress Admin Panel didn't quite suit our development needs. We needed a way to be able to easily track the changes made to the UI of the chatbot to always have the most updated view of what the chatbot would look like. This repository also stores the NLP and Rule-Based flow of the chatbot exported from the admin panel.

🪛 Installation and usage

You can directly clone this repo into the examples directory in the channel-web directory which can be located at ../botpress/data/assets/modules/channel-web/examples/ (the directory needs to be empty first). After that, the files should be inside the examples directory.

cd ../botpress/data/assets/modules/channel-web/examples/ # ../ is where you have the botpress binary downloaded
rm -r * # deletes all the content in the examples directory
git clone https://github.com/pangilinan-patrick/MyDCampus-Channel-Web.git . # clones into the current directory

The site is then tested with this link (Botpress must be running and the .tgz file has been imported using the Botpress admin site):

http://localhost:3000/assets/modules/channel-web/examples/MyDCampusPortal.html?botId=draft

Please note that this project was developed using Botpress 12.30.6. You can download the latest binaries here: https://v12.botpress.com/

📁 Important Files

📂 The My DCampus Portal_Files directory contains the assets for the MyDCampusPortal.html website. The main CSS file for the website inside the directory is app.min.css

    📄 The chat bot injection script is in MyDCampusPortal.html found around line 520, or you can just Ctrl + F + "chatbot script" to find it. You can also make edits to the sidebar navigation around line 140 which contains the settings front-end for the chat bot settings.

    📄 The main chatbot script and other related scripts are located around line 650.

📂 The Themes directory contains the CSS file for each theme that's used by the chat bot.

📂 The Chatbot tgz Import Files directory contains the chatbot NLP and Rule Based flow and is to be imported using the Botpress Admin Panel to work with an up to date chat bot at all times. The current latest version of the chatbot is v1.

📂 The Database Files directory contains the IT support knowledgebase in the MyDCampus portal website to aid in the chatbot's ability to answer user's queries. The database is to be imported into PostgreSQL for the reason that it performs better than MySQL which is used by default.

/* Example of how to import the database from the Database Files folder. A graphical tool can also be used. */
createdb mydbuddydb
pg_restore -d mydbuddydb -h 127.0.0.1 -U postgres mydbuddydb.sql

A .env file is to then be created in the Botpress directory and should be set accordingly:

# .env file
DATABASE_URL=postgres
DATABASE_URL=postgres://postgres:[email protected]:5432/mydbuddydb

✅ Features

UI Features 🖌️	Status
Themes DLSL Colors 1 DLSL Colors 2 Neutral Textbox Centered Botpress Branding Tall Chatbox	✅
Accessibility	✅
Dark Theme	✅

Chatbot Features 🦾	Status
NLP Flow	✅
Rule-Based Flow	✅

🤖 Chatbot Capabilities

NLP/NLU

• Intent Classification - recognizing what the user wants by categorizing phrases by meaning.
• Entity Extraction - pulling out structured data or information from unstructured text with reference to predefined categories, such as dates, time, cities, and names.
• Slot Tagging - identifying contiguous spans of words in an input that correspond to certain parameters, labeling a user query with semantic meaning.
• Language Identification - knowing which language hte user query is written.
• Spell Checking - making sure that the user input is spelt correctly by fixing spelling errors.
• Out of Scope Recognition - identifies instances of which user input is not within the scope of the chatbot.

Rule-Based System

• User Interface - the user can interact with the rule-based system through this component. The UI of the system takes queries and displays outputs in a readable format, allowing non-expert users to communicate easily.
• Interface or Rules Engine - this component is the brain of the rule-based system, as it serves as the main processing unit of the system. This engine applies rules to the knowledge base and extracts knowledge from it in order to derive a conclusion.
• Knowledge Base - a type of storage that stores information acquired from the different experts of the domain that the system is particularly focused on. The more content the knowledge base has, the more precise the rule-based system can be. The knowledge base used for the chatbot is the existing knowledge base available in DLSL.

More details in our thesis manuscript.

🎨 Themes and UI

This showcases the themes for the chatbot and the website.

Light Theme	Dark Theme

DLSL Colors 1	DLSL Colors 2

Neutral Textbox	Tall Chatbox

Sidebar Options

♿ Accessibility Features

Zooming In

A zoom in feature can be applied on the chatbox for users with poor eyesight.

High Contrast Theme

The high contrast theme colors are based on Contrast Ratios compliant with WCAG (Web Content Accessibility Guidance)

Black Foreground on DLSL Colors Background	DLSL Colors Foreground on Black Foreground

Gray Foreground on Black Background

High Contrast Theme

🎉 Credits

• The DLSL MyDCampus portal is developed by the ICT team at DLSL the original site can be found here (Only members of DLSL can access): https://mydcampus.dlsl.edu.ph

• Botpress is an easy to use chatbot editor powered by generative AI. Their website can be found here: https://botpress.com/

• The DLSL logos, including the BR. BUKNOY and Mascot which is used as the logo for the chatbot are intellectual properties of De La Salle Lipa. More information can be found here: https://scholarum.dlsl.edu.ph/intellectual-property/