Tape Transformation Game

This repository contains an interactive game designed to solve abstraction and reasoning tasks. The game presents users with input-output string pairs, and the goal is to determine the transformation logic to predict the output for a given input.

Table of Contents

• Installation
• Usage
• Game Interface
• System Prompt
• Development and Testing
• Contributing
• License

Installation

1. Clone the repository:

   git clone https://github.com/rsinghal757/babyARC.git
   cd babyARC

2. Set up a virtual environment (optional but recommended):

   python -m venv venv
   source venv/bin/activate  # On Windows use `venv\Scripts\activate`

3. Install the required packages:

   pip install -r requirements.txt

4. Set up environment variables: Create a .env file in the root directory with the following content:

   OPENAI_ORGANIZATION=your_organization_id
   OPENAI_PROJECT=your_project_id
   OPENAI_API_KEY=your_api_key
   

Usage

1. Run the game:

   jupyter notebook

   Open the TapeTransformationGame.ipynb notebook and execute the cells to start the game interface.

2. Playing the game: The game consists of multiple tasks, each with 4 example input-output pairs and a fifth input for which you need to predict the output. Use the provided color buttons to construct your predicted output tape.

Game Interface

The game interface is implemented using Jupyter widgets to create an interactive experience. The main components of the interface include:

• Task Display: Shows the current task number and example input-output pairs.

• Input and Output Tapes: Displays the input tape for the current task and allows the user to construct the output tape.

• Color Buttons: Buttons representing different colors that users can click to add to the output tape.

• Delete Button: Allows users to remove the last added color from the output tape.

• Check Button: Validates the constructed output tape against the correct output for the current task.

• Navigation Buttons: Previous and Next buttons to navigate through different tasks.

System Prompt

The system prompt guides the logic and reasoning process for solving the tasks. It consists of the following steps:

1. Understand the Problem: Identify the logic or transformation guiding the input-output mapping.

2. Generate a Program: Write a Python function named transform that implements the identified logic. The function should take an input tape as a string and return the transformed output tape as a string.

3. Evaluate the Program: Test the generated function against the example pairs to ensure correctness.

Development and Testing

The repository includes a test framework to evaluate the generated programs. The main components include:

• test_program_and_get_feedback Function: Compiles and executes the generated Python code, then evaluates it against example pairs and provides feedback.

• decode_task Function: Converts a task dictionary into a human-readable format for the system prompt.

• solve Function: Iteratively attempts to solve the task by generating reasoning and programs, testing them, and refining based on feedback.

Contributing

Contributions are welcome! If you have any suggestions, improvements, or bug fixes, please open an issue or submit a pull request.

License

This project is licensed under the MIT License. See the LICENSE file for details.