ChatBot Memory

An intelligent conversational AI system with persistent memory, document understanding, and personalized context management. Built with LangChain, ChromaDB, and advanced retrieval techniques.

Features

Persistent Memory System

• Long-term Memory: Remembers conversations across sessions using Mem0
• User Context: Customizable user profiles and preferences
• Conversation History: Maintains context throughout interactions

Document Intelligence

• PDF Processing: Load and analyze PDF documents
• Smart Retrieval: Hybrid search combining BM25 and semantic similarity
• Document Classification: Automatically categorizes uploaded content
• Contextual Relevance: Finds information relevant to current conversation

Quick Start

Prerequisites

• Python 3.8+
• CUDA-compatible GPU (optional, for faster embeddings)

Installation

1. Clone the repository

   git clone https://github.com/adeliie/ChatBot_memory.git
   cd ChatBot_memory

2. Install dependencies

   pip install -r requirements.txt

3. Install Ollama and download model

   # Install Ollama (visit https://ollama.ai for platform-specific instructions)
   ollama pull llama3

4. Configure API keys

   cp .env.example .env
   # Edit .env and add your Mem0 API key

5. Run the chatbot

   python main.py

💡 Usage Examples

Basic Conversation

You: Hello, how are you?
AI: Hello! I'm doing well and ready to help you with any questions or tasks you have.

Document Loading

You: load document /path/to/your/document.pdf
AI: 📄 Document '/path/to/your/document.pdf' loaded successfully.

Setting User Context

You: set_context I'm a software developer interested in Python and AI
AI: ✅ Your user context has been updated.

🏗️ Architecture

Core Components

🧮 State Management (graph_setup.py)

• LangGraph Integration: Manages conversation flow and routing
• Conditional Routing: Automatically handles different input types
• State Persistence: Maintains conversation state across interactions

🤖 Chatbot Engine (chatbot.py)

• Context Integration: Combines memory, documents, and conversation history
• Intelligent Response: Uses retrieved context to provide relevant answers
• Memory Management: Automatically stores conversations for future reference

📄 Document Processing (document.py)

• PDF Loading: Extracts and processes PDF content
• Smart Chunking: Optimally splits documents for retrieval
• Type Classification: Automatically categorizes document content

🔍 Memory & Retrieval (memory_db.py)

• Hybrid Search: Combines BM25 and semantic similarity
• Relevance Filtering: Multi-layer relevance scoring
• User Context Storage: Persistent user preference management

⚙️ Configuration (config.py)

• Model Setup: Configures LLM and embedding models
• API Management: Handles external service connections
• Device Optimization: Automatic GPU/CPU selection

🔧 Configuration

Environment Variables

Create a .env file with:

MEM0_API_KEY=your_mem0_api_key_here

Model Configuration

The system uses:

• LLM: Ollama Llama3 (configurable in config.py)
• Embeddings: HuggingFace sentence-transformers/all-MiniLM-L6-v2
• Vector Store: ChromaDB for document storage
• Memory: Mem0 for conversation memory

📁 Project Structure

ChatBot_memory/
├── main.py              # Application entry point
├── config.py            # Configuration and model setup
├── chatbot.py           # Main conversation logic
├── memory_db.py         # Memory and document management
├── document.py          # Document processing
├── graph_setup.py       # State graph configuration
├── utils.py             # Utility functions
├── requirements.txt     # Python dependencies
├── .env.example         # Environment template
├── .gitignore          # Git ignore rules
├── chroma_db/          # Document vector storage (created automatically)
└── user_contexts/      # User context files (created automatically)