Metadata-Version: 2.1
Name: pymindcore
Version: 0.2.1
Summary: A Python library for AI, easy to use, freely available for any use
Home-page: https://github.com/Dogz-R-Godz/Pymind
Author: Dogz R Godz
Author-email: doggocam01@gmail.com
License: CC0 1.0 Universal (CC0 1.0) Public Domain Dedication
Classifier: Programming Language :: Python :: 3
Classifier: License :: CC0 1.0 Universal (CC0 1.0) Public Domain Dedication
Description-Content-Type: text/markdown
License-File: LICENSE
Requires-Dist: numpy
Requires-Dist: Pillow
Requires-Dist: numexpr
Requires-Dist: numba

# Pymindcore

Pymind is a Python library for AI, supporting Numpy.

## Installation

You can install Pymind using pip: pip install pymindcore.

## Usage

You import it, you call the class with your neural network size, and you're away! There will be actual test files in the future, but for the meantime,

## License

This project is licensed under the CCO license.

## Examples

VideoGameAI.py: Currently set up to play the game Bedwars on the Minecraft server, Hypixel. It is set up to work on a 3440x1440 display, on specific game settings. 

### Changing resolutions

To make it work on a different resolution: 
    1: Go into a game, lose your bed once, die once, lose the game once. Take a screenshot of the whole screen each time you do one of those.
    2: Go into the screenshot in an app like ms paint, and find a pixel that is a certain colour, when you die, lose your bed, and lose the game. (must be unique, and not trigger when either of the others happen)
    3: Go into line 199 of VideoGameAI.py, and add in your coordinates and the pixel colour of that coordinate (in RGB). I'd recommend using Gimp for it. It should extend down to line 207. Only change the rewards if you know what you are doing.
    4: Run it, and it should work!

### Different game

To make it work on a different game:
    1: Go into the game you want to train an AI for, and think what points you would punish the AI for getting wrong. eg: dying, losing, etc.
    2: Play a game, and do one of each of those things yourself, screenshotting the whole screen each time you do one.
    3: Go into the screenshot in an app like ms paint, and find a pixel that is a certain colour when you did each of the bad things. (must be unique, and not trigger when either of the others happen)
    4: Go into line 199 of VideoGameAI.py, and add in your coordinates and the pixel colour of that coordinate (in RGB). I'd recommend using Gimp for it. It should extend down to line 207. Only change the rewards if you know what you are doing.
    4.5: If you have more than 3 punishment options, then you should copy one of them, and paste it below it, copy the __Coords, __RGB, and __Punishment, and paste them below as well.
    5: Figure out what keybinds your AI will need to have for the game. The ones set up for Minecraft should cover most bases, but to be safe, double check. If you want to, remove all the keybinds that your AI will not need, so it can learn easier.
    6: Run it and it should work!

## Discord server

Join the discord server https://discord.gg/MSMzmJ6ffj!
