Development area of coded solutions for the Sussex attempt at the AGI challenge

$ git clone https://github.com/shepai/arc_sussex
$ cd arc_sussex 
$ python -m venv .venv && source .venv/bin/activate 
$ pip install -r requirements.txt 

Most of the solutions so far make use of the utility functions provided by the arc_utils pacakge. In order to get started with development, you may need to build this package manually. Simply run pip install . inside the arc_utils directory from within your virtual environment.

Conventions for coded solutions is to have a generate and a solve function, with the name of the problem. For exmaple this would be the structure for problem 1:

def solve_1(pattern): #solve given a pattern
    # @param pattern --> matrix
    # @returns matrix
    pass

def generate_1(): #generate a new pattern and the answer
    pass

THe filename will be given he number of the problem. So for the above example 1.py. This will make everything easier to bind later on.

Data type conventions are to use numpy arrays to represent patterns, and colours be represented by these colours.

import numpy as np

colours={'black':0,'blue':1,'red':2,'green':3,'yellow':4,'gray':5,'pink':6,'orange':7,'light-blue':8,'maroon':9}
example_matrix = np.zeros((3,3))