ASCII Maze Solver with LLMs

Overview:

Welcome to the ASCII maze solver - an innovative platform designed to challenge the problem-solving abilities of Large Language Models (LLMs). With this solver, users can manually navigate through ASCII-based mazes while leveraging LLMs' proficiency in dissecting problems into smaller, more manageable components. This aids in logical navigation and highlights the potential of LLMs in complex reasoning tasks.

Key Features:

• Universal Compatibility: Accommodate ASCII mazes of various sizes.
• Maze Visualization for LLMs: After every move, the solver produces a comprehensive map displaying the absolute position. This visualization aids in providing precise feedback to Large Language Models and enhances their problem-solving capabilities.
• Interactive Navigation: Seamlessly input move directions, with the solver offering real-time updates on your current position in the maze.
• Move Validations: Receive immediate feedback on the legitimacy of your move, ensuring a guided experience.

Getting Started:

Clone the repository: git clone https://github.com/rfwarn/ASCII_LLM_Maze.git
Start the solver: python ASCIImaze.py
Dive into the Challenge: Provide your move direction. Inputs are flexible and not case-sensitive for user convenience.

Initial Prompt Directions:

You have your starting position 'S', the end position 'E' and your current position will be indicated after every move with '*'. Walls are labeled as '#' and are impenetrable. This is done one turn at a time giving me the direction you would like to go (up 'U', down 'D', left 'L', right 'R'). You can also request a 3x3 grid of the immediate area around you with '3'.

Example:

>>> Enter a move direction (U, D, L, R, or 3 (for a 3x3 of the current position)): d
>>> Good, here you are:
...
>>> Enter a move direction (U, D, L, R, or 3 (for a 3x3 of the current position)): l
>>> That move is invalid since there is a wall there.
>>> Enter a move direction (U, D, L, R, or 3 (for a 3x3 of the current position)): 3
>>> Here is a 3x3 of your current position:
>>> # #
>>> #*#
>>> # #

Contributing:

If you would like to contribute to this project, please feel free to fork the repository and create a pull request. We welcome any contributions that improve the solver or make it more user-friendly.