Rubik’s Cube Model

Rubik's Cube Notebook

I've been thinking for a while about developing a model to simulate a Rubik's cube in Python. I wanted to make use of object-oriented programming, and this project was the perfect excuse to get down to work.

View and download the code here
Go to the repo

Objectives

Code a Rubiks Cube model using object-oriented programming. The model should provide a user interface to interact with the cube and a render engine to visualize its process.

Object-oriented programming model

I started by developing a centric architecture pattern with an orchestrator class. Then I added a class to generate the mapped cube as a data frame and a 2D visualizer with the unwrapped cube to render the work in progress. Finally, I developed a "Drive" class to apply permutations in the cube.

Although it was not enough with a 2D unwrapped model, I started thinking about a 3D render engine using the Matplotlib library. I choose this library for its ease of use. It was hard to figure out how to update the model in real-time but way worth it.

To develop the model, I defined these four Python classes:

1. Class App: This class initiates the whole program and acts as an orchestrator calling other classes
2. Class Cube: This class represents the Rubik's cube model as a Dataframe.
3. Class Drive: This class maps the permutations and applies moves to the cube.
4. Class Viz: This class generates 2D and 3D models and renders each representation.

As a graphic support I drew a pattern with the classes, their methods and the calls between the different components in the infographic below.

OOP Rubik's Cube

Cube notation model

Before developing the code, I needed to define how to represent and serialize the data. It was so fun to notate and map the cube.

Naming convention

First, I established a naming convention naming each face by color. I used the same color scheme I have at my physical Rubik's cube, the BOY scheme (blue-orange-yellow).

• W -> white
• G -> green
• O -> Orange
• B -> Blue
• R -> Red
• Y -> Yellow

Faces notation

I am used to solving my Rubik's cube starting on the white face. Therefore, I decided to take this face as the main reference, becoming the front and the center face when unwrapping the cube.

    | R |
| B | W | G | Y |
    | O |

Faces parametrization

Once I assigned the colors to the faces, I named each face square by the initial of its color and the position occupied in the matrix. After this step, I had my model parametrized and ready to roll.

         |R1|R2|R3|
         |R4|R5|R6|
         |R7|R8|R9|
|B1|B2|B3|W1|W2|W3|G1|G2|G3|Y1|Y2|Y3|
|B4|B5|B6|W4|W5|W6|G4|G5|G6|Y4|Y5|Y6|
|B7|B8|B9|W7|W8|W9|G7|G8|G9|Y7|Y8|Y9|
         |O1|O2|O3|
         |O4|O5|O6|
         |O7|O8|O9|

Serialization structure

Last but not least, it was necessary to save the state of the data. To accomplish this, I used a JSON format structure defining keys by their face color and the list of squares of each face as values.

{
  "w": ["w1", "w2", "w3", "w4", "w5", "w6", "w7", "w8", "w9"],
  "r": ["r1", "r2", "r3", "r4", "r5", "r6", "r7", "r8", "r9"],
  "g": ["g1", "g2", "g3", "g4", "g5", "g6", "g7", "g8", "g9"],
  "o": ["o1", "o2", "o3", "o4", "o5", "o6", "o7", "o8", "o9"],
  "b": ["b1", "b2", "b3", "b4", "b5", "b6", "b7", "b8", "b9"],
  "y": ["y1", "y2", "y3", "y4", "y5", "y6", "y7", "y8", "y9"]
}

Moves notation

I used the Singmaster notation developed by David Singmaster. Its relative nature allows algorithms to be written in such a way that they can be applied regardless of which side is designated the top or how the colors are organized on a particular cube.

• t: Top side move 1 clockwise
• f: Front side move 1 clockwise
• d: Down side move 1 clockwise
• r: Right side move 1 clockwise
• l: Left side move 1 clockwise
• b: Back side move 1 clockwise

So for example we could apply a combined premutation by doing:

t2l3b1r2

• t2: Top side move 2 clockwise
• l3: Left side move 3 clockwise (same as left side move 1 anticlockwise)
• b1: Back side move 1 clockwise
• r2: Right side move 2 clockwise

Result

The final output displayed was way faster than I expected. The trick of regenerating the whole geometry for each move works like a charm is capable of applying over a hundred permutations in less than 30 seconds. Here is a capture of the final output applying twenty random moves to the cube.

Rubik's Cube RT rendered model

References and links

• Github repository: github.com/charlstown/RubiksCube
• Rubik's cube color schemes
• Singmaster notation
• David Singmaster
• More Projects like this here