RAG Apartment Chatbot - Installation & Usage Guide

🎯 Project Overview

This is a sophisticated RAG (Retrieval-Augmented Generation) chatbot for apartment rentals with dual interfaces:

• Interactive Mode: Terminal-based chat
• Web Mode: Beautiful Streamlit web interface

🔧 Installation

1. Install Dependencies

pip install streamlit google-genai chromadb pymongo python-dotenv sentence-transformers

2. Configure Environment Variables

Create a .env file in your project root with the following content:

# API Keys
GEMINI_API_KEY="your_gemini_api_key"
GEMINI_MODEL_NAME="gemini-2.5-flash-lite"

# MongoDB Configuration
MONGO_URI="your_mongodb_uri"
MONGO_DB_NAME="chat_sessions"
MONGO_COLLECTION_NAME="chatbot_context"

# ChromaDB Configuration
CHROMA_DB_PATH="./chroma_storage"
CHROMA_COLLECTION_NAME="json_docs"
CHROMA_EMBEDDING_MODEL_NAME="all-MiniLM-L6-v2"

# Data Configuration
JSON_FILE_PATH="./data_ingestion/apartments.json"

# Application Configuration
DEFAULT_USER_ID="default_user"
RAG_CONTEXT_LIMIT="5"
APP_MODE=web
WEB_HOST="127.0.0.1"
WEB_PORT="8501"

🚀 Usage

Interactive Mode

1. Start the FastAPI backend:

uvicorn api_server:app --host 127.0.0.1 --port 8000 --reload

2. Start the Streamlit frontend:

streamlit run frontend/app.py --server.port 8501

Make sure your .env has APP_MODE=web. Open: http://127.0.0.1:8501 in your browser.

🏗️ Architecture

Complete RAG Workflow:

┌─────────────────┐    ┌──────────────────┐    ┌─────────────────┐
│   User Query    │───▶│   ChromaDB       │───▶│   Context       │
│                 │    │   (Apartments)   │    │   Retrieved     │
└─────────────────┘    └──────────────────┘    └─────────────────┘
                                                         │
┌─────────────────┐    ┌──────────────────┐            ▼
│   Final         │◀───│   Gemini AI      │◀───┌─────────────────┐
│   Response      │    │   Generation     │    │   Enhanced      │
└─────────────────┘    └──────────────────┘    │   Prompt        │
         │                                      └─────────────────┘
         ▼                                               ▲
┌─────────────────┐    ┌──────────────────┐            │
│   MongoDB       │◀───│   Conversation   │────────────┘
│   Storage       │    │   History        │
└─────────────────┘    └──────────────────┘

🌐 Streamlit Web Interface Features

🎨 Beautiful UI Components:

• ✅ Real-time chat interface with message bubbles
• ✅ Session management with user/session IDs
• ✅ Sidebar with controls and system status
• ✅ Responsive design for mobile/desktop
• ✅ Custom CSS styling with gradients
• ✅ Loading indicators and error handling

🛠️ Interactive Features:

• ✅ New Session: Start fresh conversation
• ✅ Clear Chat: Remove history from UI and database
• ✅ System Status: Check component health
• ✅ Example Queries: Quick-start with sample questions
• ✅ Auto-scroll: Latest messages always visible

📊 Session Information:

• User ID and Session ID display
• Message count tracking
• Connection status indicator
• Real-time updates

📊 Data Loading

• First Run: Automatically loads apartments.json into ChromaDB
• Subsequent Runs: Skips loading if data already exists
• Persistent Storage: Data survives application restarts

🔄 Mode Switching

Switch between modes by changing the APP_MODE environment variable:

# Interactive mode
export APP_MODE="interactive"
python main.py

# Web mode (Streamlit)
export APP_MODE="web"
python main.py

🛠️ Streamlit Advantages

🚀 Why Streamlit over FastAPI?

• Simpler Development: No HTML/CSS/JavaScript needed
• Built for AI: Perfect for chatbots and ML applications
• Native Chat Support: Built-in chat message handling
• Rapid Prototyping: Get beautiful UI in minutes
• Session Management: Built-in session state handling
• Live Updates: Auto-refresh and real-time updates

🎯 Perfect for RAG Applications:

• Chat interfaces are first-class citizens
• Easy data visualization and status display
• Simple deployment and sharing
• Great for internal tools and demos

🛠️ Troubleshooting

Common Issues:

1. Import Error: pip install streamlit
2. Port Conflict: Change WEB_PORT in .env
3. Database Connection: Check MongoDB URI
4. API Limits: Verify Gemini API key

Streamlit-Specific:

• Browser not opening: Use --server.headless true
• Port issues: Use --server.port XXXX
• Performance: Use st.cache_data() for large data

🎨 Customization

Frontend Styling:

• Edit CSS in frontend/app.py
• Use st.markdown() with custom CSS
• Modify color schemes and layouts

Add New Features:

• File upload for apartment data
• Charts and visualizations
• Export conversation history
• Multi-language support

🚀 Deployment Options

Local Development:

streamlit run frontend/app.py

Streamlit Cloud:

• Push to GitHub
• Connect repository to Streamlit Cloud
• Deploy with one click