AGC Document Chatbot

A Streamlit-based web application that provides intelligent search and chat capabilities for Attorney General's Chambers (AGC) documents. The system uses Retrieval-Augmented Generation (RAG) to enhance search accuracy and provide context-aware responses.

Features

• Document Browsing: Browse through all available AGC documents with filtering by document type and title/content search
• Enhanced RAG Search: Search documents using AI-enhanced query understanding
• Document Detail View: View full document details with contextual information
• Chat Interface: Chat with AI about document content with context-awareness

Architecture

The application consists of several key components:

• Web Interface: Built with Streamlit
• Document Database: MySQL database for storing document metadata and content
• Embedding Services: Vector embeddings for semantic search capabilities
• RAG Enhancement: Improved search using OpenAI's capabilities

Project Structure

.
├── app.py                  # Main Streamlit application
├── config.py               # Configuration settings
├── db/                     # Database utilities
│   └── import_lkk_data.py  # Script for importing LKK data
├── embedding/              # Embedding and RAG services
│   ├── embedding_service.py
│   ├── enhanced_rag_service.py
│   └── rag_service.py
├── utils/                  # Utility functions
├── Data/                   # Document data
└── requirements.txt        # Python dependencies

Installation

1. Clone the repository
2. Create a virtual environment:

   python -m venv venv
   source venv/bin/activate  # On Windows: venv\Scripts\activate
   

3. Install dependencies:

   pip install -r requirements.txt
   

4. Configure environment variables (create a .env file based on requirements)

Database Setup

1. Install and run XAMPP:

   • Download XAMPP from https://www.apachefriends.org/
   • Install and launch XAMPP Control Panel
   • Start the Apache and MySQL services
   • Access phpMyAdmin at http://localhost/phpmyadmin
   • Create a new database named agc

2. Configure the database connection in your .env file:

   MYSQL_HOST=localhost
   MYSQL_USER=root
   MYSQL_PASSWORD=
   MYSQL_DATABASE=agc
   

Data Import (Optional)

To import LKK (Laporan Keputusan Kes) data into the system:

python -m db.import_lkk_data

This script will:

• Set up required database tables
• Import available data from SQL or PDF files in the Data directory
• Generate document embeddings for search functionality

Usage

Run the Streamlit application:

streamlit run app.py

The application will be available at http://localhost:8501 by default.

Requirements

• Python 3.7+
• MySQL database (via XAMPP)
• OpenAI API key (for embedding and RAG features)