- Connect 4 written in Python 3.7
- No external dependencies beyond the Python standard library
- Starter code is included
Please do not change any of the methods that are already implemented, any of the method signatures, and please do not change any of the unit tests. Add your code into the incomplete methods and add any additional utility methods/functions. Incomplete methods are explicitly marked "Your Code Here."
Run python3 play.py when in the same directory as the source files. Your Python command may be defined as python instead of python3.
Run python3 connectfour_tests.py when in the source's directory to run all of the unit tests. Your Python command may be defined as python instead of python3.
You will be creating an artificial opponent that plays the game of Connect Four using the minimax algorithm. Your Connect Four AI will likely be good enough to beat most human players.
In Connect Four, two players alternate dropping different colored (red or black traditionally) pieces in a seven-column, six-row grid. Pieces fall from the top of the grid to the bottom until they hit the bottom or another piece. In essence, the player’s only decision each turn is which of the seven columns to drop a piece into. The player may not drop it into a full column. The first player that has four pieces of their color next to one another with no breaks in a row, column, or diagonal wins. If no player achieves this, and the grid is completely filled, the game is a draw.
Your goal is to build a program in which the user can play a game of Connect Four against the computer. In other words, you will have a main game loop that takes a move from the user, updates the state of the game, gets a response from the computer AI, and updates the state again. All along your program should be printing out the contents of the board to the user so they can see the game's progression.
Your game must use the included implementation of the minimax/alphabeta algorithm to figure out the computer's moves. You must use the provided starter code as part of your program. Please do not change the included function/method signatures, the Board interface, nor the properties of the C4Board class (although you may add additional methods). Your program must pass the included unit tests. You should also test your program beyond the unit tests since they do not cover every aspect of the program. Feel free to add additional utility functions. You are not limited to the function signatures included in the starter code. They are just a minimum that must be included.
-
Use the contents of the repository as starter code, copying it into a private repository of your own that you create.
-
Add us (@davecom, @brianrhall, @mkgungor) as collaborators on the repository on GitHub.
-
Fill in the functions with the missing code and add any additional utility functions that you see fit.
-
Test your program on its own and also test the included unit tests.
-
Submit the URL of your repository.
Before you begin, take the time to carefully read through the included files from the starter code and try to understand how they fit together. There are in-line comments explaining all aspects of the starter code, including comments at the top of each file explaining that file's purpose. You will almost certainly need to add additional utility functions/methods to make your code readable.
The included MiniMax/AlphaBeta algorithm in minimax.py should be called # of legal moves times for a given starting position from your version of find_best_move(). Each call to alphabeta() will return an evaluation of a move. You must look at how high that evaluation is relative to the evaluation of the other legal moves and return the highest evaluated move (the "best move").
Note that C4Board implements the Board abstract base class and maintains state. position is the grid at any given time. What pieces are on what squares? It is maintained as a list of columns, and using them we can quickly tell if a column is full and therefore a move cannot be played there. Initially, every column has a count of 0. turn is which player's turn it is. Note that the Piece type is used both to designate pieces on the grid and to designate players via the turn property.
One slightly nuanced method on C4Board is move(). It is meant to return a new C4Board instance with the given move made. It is not meant to modify the existing C4Board instance.
The hardest part of doing this will likely be writing the evaluate() method on C4Board. The purpose of evaluate() is to tell alphabeta() "how good" a given position is for the provided player. For instance, a position in which a player is about to win because they have two segments of the board where they have 3 pieces in a row and there is no way to block them at both simultaneously should be evaluated as "very good" for that player. In other words it should get a high number returned. On the other hand, a position in which both players do not seem to have any winning opportunities should be evaluated "lower" than the first example. And a position in which the opponent looks to be near winning, should be evaluated even lower. The evaluation numbers only matter in relation to one another. You make up the scale. Note that there is more than one way to evaluate a position, and it is very possible for you to come up with your own solution that will be different from other peoples and still work well.
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent