NotchNet Backend

NotchNet is an AI-powered Minecraft knowledge companion that uses RAG (Retrieval Augmented Generation) to answer questions about Minecraft and its mods. Check it out: https://github.com/aaravchour/notchnet-mod

🌟 Features

• Local RAG Pipeline: Runs entirely on your machine using Ollama.
• Dynamic Wiki Fetching: Can fetch and index any MediaWiki-based wiki (e.g., RLCraft, Feed The Beast).
• Auto Mod Detection: Automatically finds and learns about installed mods when the game launches.
• Cloud Mode: Support for offloading AI inference to a remote server for low-end machines.
• Mod Awareness: Context-aware answers based on the loaded wikis.

🚀 Getting Started (Local)

Prerequisites

1. Python 3.10+
2. Ollama installed and running.
3. (For local model) 8GB of VRAM and 16GB of RAM for a smooth experience.

Quick Start

1. Clone the repository.

2. Run the startup script:

   Windows:

   start_local.bat

   Mac/Linux:

   chmod +x start_local.sh
   ./start_local.sh

   This script will:

   • Create a virtual environment and automatically install dependencies.
   • Ask if you want to run Locally or use Cloud/Remote AI.
   • Local: Pulls the necessary Ollama model (default: llama3:8b).
   • Cloud: Configures connection to your remote Ollama instance.
   • Start the API server.

3. Interact with the API: The server runs at http://localhost:8000.

   Ask a Question:

   curl -X POST http://localhost:8000/ask \
        -H "Content-Type: application/json" \
        -d '{"question": "How do I make a shield?"}'

📚 Adding Mod Wikis (still in development)

You can teach NotchNet about new mods by fetching their wikis.

Endpoint: POST /admin/add-wiki

Example (Teaching it RLCraft):

curl -X POST http://localhost:8000/admin/add-wiki \
     -H "Content-Type: application/json" \
     -d '{
           "api_url": "https://rlcraft.fandom.com/api.php", 
           "categories": ["Crafting", "Items", "Mobs"] 
         }'

Note: This process runs in the background. It will fetch pages, clean them, rebuild the index, and reload the bot's memory.

🛠 Configuration

See config.py for default settings. You can override them using environment variables or a .env file.

Variable	Description	Default
LOCAL_MODE	Bypass API key checks for local use	true (in start script)
LLM_MODEL	Ollama model to use	llama3
OLLAMA_HOST	URL of Ollama server	http://127.0.0.1:11434

📜 License

GPLv3 License. See LICENSE for details.