ME-Bot (DesertMind) 🌍🐫

ME-Bot is a specialized React and Flask-based Artificial Intelligence assistant using LangChain and Retrieval-Augmented Generation (RAG). This chatbot acts as an expert exclusively on Middle Eastern culture, history, politics, and geographic information. It strictly enforces a specialized focus on the Middle East, gracefully declining questions outside its scope.

---

🛠️ Architecture and Tech Stack

• Frontend: React.js
• Backend: Python / Flask
• LLM Orchestration: LangChain
• Vector Database: ChromaDB
• Embeddings: HuggingFace (sentence-transformers/all-mpnet-base-v2)
• LLM Provider: OpenRouter (OpenAI-compatible APIs)

---

🚀 Features

• Retrieval-Augmented Generation (RAG): Uses a local directory (middle_east_data/) populated with specialized .txt knowledge files to strictly base answers on verifiable context.
• Auto-Initialization: On initial boot, the backend automatically reads the text files, chunks them using the RecursiveCharacterTextSplitter, creates embeddings, and persists the data to a local chroma_db folder.
• Graceful Fallback Logic: The System Prompt is engineered to differentiate between correct facts, explicit false statements, and out-of-scope inquiries.
• RESTful API Endpoint: Handles standard backend connectivity via POST /api/chat.
• Static Hosting: The Flask application is built to serve the compiled frontend production static bundle directly.

---

📂 Project Structure

DesertMind/
├── chat_bot.py               # Main Flask API and LangChain configuration
├── requirement.txt           # Python dependencies (Flask, Langchain, etc.)
├── middle_east_data/         # Directory containing text documents for the knowledge base
│   └── me_members.txt        # Example facts and regional changes (2026 update)
├── chroma_db/                # (Generated) Local vector store initialized by LangChain
├── .env                      # Contains Environment variables (like OPENROUTER_API_KEY)
└── front_end/                # React.js application
    ├── src/                  # React components, pages, CSS styles, and UI logic
    ├── public/               # Static base HTML and assets
    └── package.json          # Node modules and build scripts

---

⚙️ Setup and Installation

1. Backend Setup (Flask & LangChain)

1. Navigate to the Root Directory

   cd DesertMind

2. Setup a Python Virtual Environment (Optional but Recommended)

   python3 -m venv venv
   source venv/bin/activate  # On Windows: venv\Scripts\activate

3. Install Dependencies

   pip install -r requirement.txt

4. Environment Variables Create a .env file in the root directory and add your OpenRouter key:

   OPENROUTER_API_KEY=your_api_key_here

5. Start the Backend Server

   python3 chat_bot.py

   Note: On your first run, the server will briefly pause to automatically chunk your middle_east_data/ files and create the chroma_db vector store before binding to http://localhost:5000.

2. Frontend Setup (React.js)

1. Open a new terminal and navigate to the frontend directory

   cd DesertMind/front_end

2. Install Node.js dependencies

   npm install

3. Start the Development Server

   npm start

   The React app will typically run on http://localhost:3000 and proxy backend requests to the running Flask server.

---

📦 Production Deployment

The Flask server (chat_bot.py) is written with production hosting in mind. It serves the static bundle directly from /api fallback routes:

1. Build the React Frontend

   cd front_end
   npm run build

   Ensure the output directory is configured as ../me-bot-frontend/build (or the respective script mapping) so Flask can serve index.html on the root / endpoint.
2. Run your Flask app under a WSGI server like gunicorn to expose and host the full stack seamlessly!

---

📝 Updating the Knowledge Base

To update what ME-Bot knows:

1. Add, modify, or remove .txt files inside the middle_east_data/ folder.
2. Delete the chroma_db/ folder from your root directory entirely.
3. Restart chat_bot.py. The backend will naturally rebuild the vector database with your new contextual data!