📚 Multi-Modal ChatBot: RAG + CRAG with PostgreSQL

This project is a robust, scalable conversational AI system that integrates Retrieval-Augmented Generation (RAG) and Conversational RAG (CRAG) in a clean, togglable interface using Streamlit, PostgreSQL, and OpenAI or BGE embeddings.

Built as an evolution of a production-grade internal tool, this repo now serves as a high-quality, reusable personal project for document-based chatbot development.

---

🚀 Features

🔍 RAG (Retrieval-Augmented Generation)

• Embeds and retrieves top relevant document chunks for each user query.
• Supports both OpenAI and BGE embeddings.
• Stateless: each query is processed independently.

🧠 CRAG (Conversational RAG)

• Maintains a full user-bot conversation history (stored in PostgreSQL).
• Retrieval and response generation are contextualized using recent dialogue.
• Auto-formatted history embedded into the prompt dynamically.

🧰 Tooling

• Vector storage and history: PostgreSQL
• Embedding models: OpenAI / BAAI/bge-base-en-v1.5
• Streamlit UI with RAG/CRAG toggle
• File ingestion, chunking, and OCR-ready hooks

---

🧱 Architecture

                  ┌──────────────────────┐
                  │   Streamlit UI       │ ◄── User interaction (Chat.py)
                  └────────┬─────────────┘
                           │
             ┌─────────────▼──────────────┐
             │ RAG / CRAG Mode Selector   │ ◄── User selects mode
             └──────┬────────────┬────────┘
                    │            │
     ┌──────────────▼──┐     ┌───▼────────────────┐
     │  RAG Retriever  │     │  CRAG Retriever    │
     │ (rag_utils1.py) │     │ (crag_utils_pg.py) │
     └─────────────────┘     └────────────────────┘
             │                       │
        ┌────▼────┐            ┌─────▼──────┐
        │  Embeds │            │ History +  │
        │ Chunks  │            │ Contextual │
        └────┬────┘            │ Retrieval  │
             │                 └────────────┘
        ┌────▼──────────────────────────────┐
        │   PostgreSQL: chunks + history    │
        └───────────────────────────────────┘

---

💾 File Structure

my-chat-app/
├── Chat_with_CRAG_PostgreSQL.py   # Streamlit frontend with RAG/CRAG toggle
├── rag_utils1_cleaned_no_comments.py # RAG utilities (chunking, store, retrieve)
├── crag_utils_pg.py               # CRAG utilities with PostgreSQL backend
├── .env                           # Environment variables (API keys, DB config)
├── requirements.txt               # Python dependencies
└── README.md

---

⚙️ Setup Instructions

1. Clone the Repo

git clone https://github.com/YOUR_USERNAME/my-chat-app.git
cd my-chat-app

2. Install Dependencies

pip install -r requirements.txt

3. Configure Environment

Create a .env file in root:

OPENAI_API_KEY=your_openai_key
PG_HOST=localhost
PG_PORT=5432
PG_DB=ragdb
PG_USER=raguser
PG_PASS=ragpass

4. Run the App

streamlit run Chat_with_CRAG_PostgreSQL.py

---

🧪 Optional Enhancements

• Add login control (auth.py) for real user security
• Integrate OCR pipeline using pytesseract and PIL
• Serve on cloud platforms (e.g., Streamlit Cloud, Docker, GCP, AWS)

---

📜 License

This project is licensed under the MIT License — free to use, modify, and distribute.

---

🙋‍♂️ Author

Created by [Your Name] as a personal showcase of full-stack applied AI, with production-quality structure and deployment readiness.