AI Chatbot with STM & LTM Memory

This project is created using Short Term Memory (STM) and Long Term Memory (LTM) AI Chatbot. FastAPI API management stores and consumes LTM with Streamlit front-end interface and PostgreSQL database. The system generates responses using Google Gemini and provides embedding-based memory management. The project is designed to work in isolation with Docker containers.

---

Features

• Short Term Memory (STM): Stores session-based user data, resets on page refresh.
• Long Term Memory (LTM): Stores persistent user data across sessions for personalized responses.
• Embedding Model: Utilizes Google’s Generative API to generate embeddings and store them in the database.
• Interactive Chat Interface: Provides a user-friendly interface through Streamlit for real-time conversations.
• Dockerized Services: All services, including FastAPI and Streamlit, are containerized using Docker for easy deployment.

---

Technologies Used

• FastAPI: High-performance web framework for building the backend APIs.
• Streamlit: Interactive UI framework for building user interfaces.
• Google Generative API: For generating text embeddings.
• PostgreSQL: Used for storing LTM data with vector-based storage.
• Docker: For containerizing the application components.
• Docker Compose: For managing multi-container Docker applications.

---

Setup and Installation

• Docker and Docker Compose must be installed.

Installation Steps

1. Clone the repository:

   github.com/TarikYil/AI_Chatbot_Memory.git
   cd AI_Chatbot_Memory.git
   

2. Build and start the containers using Docker Compose:

   docker-compose up --build -d

3. To access the Streamlit app, open your browser and go to:

   http://localhost:8502
   

Project Setup

1. Create a .env file in the root of the project with the following variables:

   GOOGLE_API_KEY=your_google_api_key
   POSTGRES_HOST=localhost
   POSTGRES_PORT=5432
   POSTGRES_DB=mydatabase
   POSTGRES_USER=myuser
   POSTGRES_PASSWORD=mypassword

Usage

• The chatbot will respond based on the memory type selected (STM or LTM).
• STM: Stores conversation history only for the session, resets on page reload.
• LTM: Stores and retrieves data persistently, providing personalized responses based on long-term memory.

Project Structure

AI_CHATBOT_MEMORY/
├── app/
│   ├── __pycache__/
│   ├── embeddings/
│   │   ├── __pycache__/
│   │   └── google_embeddings.py      # Google Embeddings integration
│   ├── llm/
│   │   └── gemini.py                 # Gemini LLM model
│   ├── memory/
│   │   ├── __pycache__/
│   │   ├── ltm.py                    # Long Term Memory functionality
│   │   ├── stm.py                    # Short Term Memory functionality
│   │   ├── main.py                   # Main application logic
│   │   └── models.py                 # Data models (e.g., ChatMessage)
│   ├── .env                          # Environment variables
│   └── main.py                       # Entry point for the application
├── ui/
│   └── streamlit_app.py              # Streamlit frontend UI
├── .gitignore
├── docker-compose.yml                # Docker Compose configuration
├── Dockerfile.api                    # Dockerfile for FastAPI service
├── Dockerfile.streamlit              # Dockerfile for Streamlit UI
├── README.md                         # Project documentation
└── requirements.txt                  # Python dependencies

Create and configure the .env file.