doc.md

Minesweeper API summary

Game

The game is implemented in minesweeper.py. Game state is instance of class Board and is defined by 2-dimensional list (list of lists) of Tiles. In this case you won't get access to all information about the state — some tiles are covered, so you will be getting only view of the board, that won't contain any information about covered tiles.

class Tile

Represents one tile of the board as a dataclass containing this fields:

• mine — Whether there is a mine on this tile (None for unknown).
• mines_around — How many mines are in total on neighboring (maximal 8) tiles. Has value -1 if tile is not uncovered.
• visible — Whether the tile is visible to the agent — whether it is uncovered.
• flag — Whether the tile is flagged.

class Board

Here is the list of methods you might want to use in your agent implementation:

• tile — Return tile at a given position.
• is_possible — Whether action is possible in given state.

The rest are methods for updating game state, but since you will be working only with its view representation you won't be able to use them properly.

class ActionFactory

There are only 3 actions which you can make: ask for advice, uncover a tile and toggle flag on the tile. Since they all are represented by single class Action you can use methods of this class to create instance you need.

Artificial agent

ArtificialAgent is a basic agent interface implemented in script artificial_agent.py that implements core agent methods for interacting with minesweeper game. You can use random, verbose (note that in this assignment verbose can be 0, 1, 2 and 3 specifying verbose level) and unknown (as read-only) that holds all unknown tiles' positions.

Agent will automatically use hints — uncover advised tile. If there is no such tile it will call think_impl.