I’m now officially done with this project. You can see a preview on https://asciinema.org/a/vIZ6F7bfI0mkdDf5. I really happy with what it turned out to be. I learned alot of json formatting, database data flow and code structure. This is the final product:
The Wandering Modder
A fast, AI-driven search engine for Minecraft modding. It fetches data via the Modrinth API and uses a local ChromaDB vector database, allowing you to search for mods, shaders, datapacks, resource packs and plugins using natural language.
Installation
Windows:
git clone [https://github.com/Camiel13/The-Wandering-Modder.git](https://github.com/Camiel13/The-Wandering-Modder.git) the_wandering_modder
cd the_wandering_modder
pip install -r requirements.txt
Linux:
git clone [https://github.com/Camiel13/The-Wandering-Modder.git](https://github.com/Camiel13/The-Wandering-Modder.git) the_wandering_modder
cd the_wandering_modder
python3 -m venv venv
source venv/bin/activate
pip install -r requirements.txt
Usage
To start execute this command:
python3 main.py
The Wandering Modder Commands
You have to replace {project_type} with one of the following project types: mod, datapack, resourcepack, shader or plugin. Make sure to describe your project in keywords (e.g. redstone, technical, components, wires)!
{project_type} init - Sets up the searching of a specific project type. !! REQUIRED !!
{project_type} query - Search the built database for specific mods with keywords.
help - Shows all the commands available.
clear - Clears the terminal.
exit / quit - Shuts down the program.
Hardware Acceleration
You can make the process of building the vector database up to 10-50x faster by using a dedicated Nvidia GPU with the CUDA toolkit. This is done by replacing the normal package with a GPU-supported variant and installing the CUDA toolkit onto your system. Note that this can only be done when using a Nvidia GPU with the proper installation of the CUDA toolkit and using recent drivers!
pip uninstall onnxruntime
pip install onnxruntime-gpu
