Skip to content

Instantly share code, notes, and snippets.

@kmandreza
Last active December 24, 2015 04:59
Show Gist options
  • Star 0 You must be signed in to star a gist
  • Fork 0 You must be signed in to fork a gist
  • Save kmandreza/6747393 to your computer and use it in GitHub Desktop.
Save kmandreza/6747393 to your computer and use it in GitHub Desktop.
Sudoku
PART 1
By the end of the next two challenges you'll have a fully-functioning Sudoku solver that you can run from the command line.
Sudoku is a logic-based, combinatorial number-placement puzzle. The objective is to fill a 9×9 grid with digits so that each column, each row, and each of the nine 3×3 sub-grids that compose the grid (also called "boxes") contains all of the digits from 1 to 9.
The person who created the puzzle provides a partial solution so that some squares already have numbers. Typically, there are enough initial numbers to guarantee a unique solution.
Unsolved Solved Sudoku
For the first iteration, we're just going build a solver that fills in "logically necessary" squares and requires no guessing.
Learning Goals
Breathe: we know this is the hardest challenge you've hit yet. When in doubt, take a step back and reflect on how you or someone else plays Sudoku and don't get "stuck" in the code.
Your goal here is to learn how to model a relatively complex real-world system — a person solving Sudoku. You'll learn the importance of well-encapsulated code and use some basic object-oriented design.
Objectives
Modeling: Write down the nouns and verbs of the game
Think carefully about all the nouns and verbs in a Sudoku game. There's the person who created the puzzle (the setter). There's the person who is solving the puzzle (the player). What are the important parts of the board called? How do the player and setting interact with them?
A computer program that solves Sudoku is simulating the player, which means the better you can empathize with the player the more likely you'll understand how to write a Sudoky solver. You'll be tempted to focus on the board first — is it some kind of array, a hash, something else? — but don't! Understanding the person playing the game is key; the code to "power" the board is a detail.
Modeling: Strategies for Humans
Get out an actual Sudoku puzzle, printed on a piece of paper. Play it with your pair. Pay attention to yourself and to each other.
What strategies are you adopting and why?
How do you choose where to start?
How do you know when to really put a number in a cell?
Did you adopt the same notation/board markings while playing Sudoku? Why? If not, why did you choose differently?
Are you avoiding any strategies because they're too tedious or require you to remember too much?
It's important to see that sometimes the strategies that work for us (humans) would be really hard to implement on a computer, and vice versa: strategies we avoid because we'd have to write too much, use too many sheets of paper, or remember too much are a cakewalk for a computer.
Modeling: Pseudocode for First Iteration
Remember, for the first iteration, we're just going build a solver that fills in "logically necessary" squares and requires no guessing. This might not solve every Sudoku board, although it often solves the easiest. How can you tell when you've filled in all the "logically necessary" squares?
Write out pseudocode for this version, separately, and compare it to each other. How does it differ? Which approach seems more sound? Are there some core operations or methods you need to support?
For example, given a cell/square, you'll probably need at least three methods:
Give me the other cells in that cell's row.
Give me the other cells in that cell's column.
Give me the other cells in that cell's box.
Ruby: Now code it!
Your Sudoku solver should take a string like this as its input:
619030040270061008000047621486302079000014580031009060005720806320106057160400030
Each consecutive 9 digits represents a row of the Sudoku board, and a 0 represents an empty cell. It'd work like this:
game = Sudoku.new('619030040270061008000047621486302079000014580031009060005720806320106057160400030')
game.solve!
puts game.board
This would print something like this:
---------------------
4 8 3 | 9 2 1 | 6 5 7
9 6 7 | 3 4 5 | 8 2 1
2 5 1 | 8 7 6 | 4 9 3
---------------------
5 4 8 | 1 3 2 | 9 7 6
7 2 9 | 5 6 4 | 1 3 8
1 3 6 | 7 9 8 | 2 4 5
---------------------
3 7 2 | 6 8 9 | 5 1 4
8 1 4 | 2 5 3 | 7 6 9
6 9 5 | 4 1 7 | 3 8 2
---------------------
Again, note, this first iteration might not solve every possible Sudoku board. This means it would finish when it can no longer make a choice and "give up." We'll create the fully-featured version in the next challenge.
Don't worry about the particular format of the board when printed. The key thing is getting the logic around solving/guessing correctly.
What about performance?
Do not worry about performance yet! Optimizations can come later. Clean, logical code is more important and will be easier to refactor.
PART 2
Starting with your attempt from the previous iteration of the Sudoku solver, we now want to add the ability to make guessing.
What happens if a square can contain multiple possible numbers and you don't have enough information to conclude right then and there which number it is?
Most people who play Sudoku "guess," usually by writing possibilities in the corners of the square.
Objectives
Implement a guessing-friendly version of Sudoku
Once we've filled in all cells which have only one possible value, we have to guess. Write out pseudocode for how that might work before you dive in!
This solver should now be able to solve any Sudoku puzzle, although some will take a long time. Try it with the sample.unsolved.txt file in the attached gist.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment