connect4: Implement connect4 game

This was a joint effort of ChatGPT and me. I'm not super interested
in plain algorithms. Therefore, I got myself some help from ChatGPT.
Other than that, everything is implemented by myself.

userspace/Cargo.toml

@@ -18,6 +18,11 @@ common = { path = "../common" }
 test = false
 bench = false
 
+[[bin]]
+name = "connect4"
+test = false
+bench = false
+
 [[bin]]
 name = "init"
 test = false

userspace/src/bin/connect4/game_board.rs

@@ -0,0 +1,306 @@
+use userspace::{print, println};
+
+const COLUMNS: u8 = 7;
+const ROWS: u8 = 6;
+
+#[derive(Debug, Clone, Copy)]
+pub enum Player {
+    C,
+    H,
+}
+
+impl Player {
+    fn opponent(&self) -> Self {
+        match self {
+            Self::C => Self::H,
+            Self::H => Self::C,
+        }
+    }
+
+    pub fn switch(&mut self) {
+        *self = self.opponent();
+    }
+}
+
+#[derive(Clone, Copy, PartialEq, Eq)]
+enum Position {
+    Empty,
+    C,
+    H,
+}
+
+impl From<Player> for Position {
+    fn from(value: Player) -> Self {
+        match value {
+            Player::C => Position::C,
+            Player::H => Position::H,
+        }
+    }
+}
+
+#[derive(Clone)]
+pub struct GameBoard {
+    board: [[Position; COLUMNS as usize]; ROWS as usize],
+}
+
+impl GameBoard {
+    pub fn new() -> Self {
+        Self {
+            board: [[Position::Empty; COLUMNS as usize]; ROWS as usize],
+        }
+    }
+
+    pub fn print(&self) {
+        for row in 0..ROWS {
+            for column in 0..COLUMNS {
+                match self.board[row as usize][column as usize] {
+                    Position::Empty => print!(" - "),
+                    Position::C => print!(" C "),
+                    Position::H => print!(" H "),
+                }
+            }
+            println!("");
+        }
+        println!(" 1  2  3  4  5  6  7");
+        println!("");
+    }
+
+    pub fn put(&mut self, player: Player, column: u8) -> Result<(), ()> {
+        for row in (0..ROWS).rev() {
+            if self.board[row as usize][column as usize] == Position::Empty {
+                self.board[row as usize][column as usize] = player.into();
+                return Ok(());
+            }
+        }
+        Err(())
+    }
+
+    fn calculate_score(&self, player: Player) -> i64 {
+        let opponent = player.opponent();
+
+        let mut score = 0;
+
+        // Evaluate all possible directions for scoring
+        let directions = [(1, 0), (0, 1), (1, 1), (1, -1)]; // Right, Down, Diagonal-right-down, Diagonal-left-down
+
+        for row in 0..ROWS {
+            for col in 0..COLUMNS {
+                for &(dr, dc) in &directions {
+                    score += self.evaluate_line(row, col, dr, dc, player, opponent);
+                }
+            }
+        }
+
+        score
+    }
+
+    fn evaluate_line(
+        &self,
+        start_row: u8,
+        start_col: u8,
+        dr: isize,
+        dc: isize,
+        player: Player,
+        opponent: Player,
+    ) -> i64 {
+        let mut player_count = 0;
+        let mut opponent_count = 0;
+        let mut empty_count = 0;
+
+        // Iterate through up to 4 positions in the specified direction
+        for i in 0..4 {
+            let r = start_row as isize + i * dr;
+            let c = start_col as isize + i * dc;
+
+            // Check bounds
+            if r < 0 || r >= ROWS as isize || c < 0 || c >= COLUMNS as isize {
+                return 0;
+            }
+
+            match self.board[r as usize][c as usize] {
+                pos if pos == player.into() => player_count += 1,
+                pos if pos == opponent.into() => opponent_count += 1,
+                Position::Empty => empty_count += 1,
+                _ => {}
+            }
+        }
+
+        // Scoring rules
+        match (player_count, opponent_count, empty_count) {
+            (4, 0, _) => 100000,  // Winning line for the player
+            (3, 0, 1) => 100,     // Strong threat for the player
+            (2, 0, 2) => 10,      // Moderate threat for the player
+            (1, 0, 3) => 1,       // Weak threat for the player
+            (0, 4, _) => -100000, // Opponent's winning line
+            (0, 3, 1) => -100,    // Strong threat for the opponent
+            (0, 2, 2) => -10,     // Moderate threat for the opponent
+            (0, 1, 3) => -1,      // Weak threat for the opponent
+            _ => 0,               // Neutral
+        }
+    }
+
+    // Checks if the game is over
+    pub fn is_game_over(&self) -> Option<Player> {
+        // Check if all columns are full
+        if self
+            .board
+            .iter()
+            .all(|row| row.iter().all(|&pos| pos != Position::Empty))
+        {
+            return None; // Draw
+        }
+
+        self.check_winner()
+    }
+
+    // Checks if there is a winner
+    fn check_winner(&self) -> Option<Player> {
+        for row in 0..ROWS {
+            for col in 0..COLUMNS {
+                if let Some(player) = self.check_four_in_a_row(row, col) {
+                    return Some(player);
+                }
+            }
+        }
+        None
+    }
+
+    // Checks for four in a row starting from a specific position
+    fn check_four_in_a_row(&self, row: u8, col: u8) -> Option<Player> {
+        let directions = [
+            (0, 1),  // Horizontal
+            (1, 0),  // Vertical
+            (1, 1),  // Diagonal down-right
+            (1, -1), // Diagonal down-left
+        ];
+
+        if let Position::C | Position::H = self.board[row as usize][col as usize] {
+            let current_position = self.board[row as usize][col as usize];
+            for (dr, dc) in directions {
+                let mut count = 1;
+
+                for step in 1..4 {
+                    let new_row = row as isize + dr * step;
+                    let new_col = col as isize + dc * step;
+
+                    if new_row < 0
+                        || new_row >= ROWS as isize
+                        || new_col < 0
+                        || new_col >= COLUMNS as isize
+                    {
+                        break;
+                    }
+
+                    if self.board[new_row as usize][new_col as usize] == current_position {
+                        count += 1;
+                    } else {
+                        break;
+                    }
+                }
+
+                if count == 4 {
+                    return match current_position {
+                        Position::C => Some(Player::C),
+                        Position::H => Some(Player::H),
+                        _ => None,
+                    };
+                }
+            }
+        }
+        None
+    }
+
+    fn for_valid_moves(&self, mut f: impl FnMut(u8) -> bool) {
+        for column in 0..COLUMNS {
+            if self.board[0][column as usize] == Position::Empty {
+                if !f(column) {
+                    break;
+                }
+            }
+        }
+    }
+
+    /// Perform the minimax algorithm with alpha-beta pruning.
+    fn minimax(
+        &self,
+        depth: u8,
+        alpha: i64,
+        beta: i64,
+        maximizing_player: bool,
+        player: Player,
+        counter: &mut usize,
+    ) -> i64 {
+        *counter += 1;
+
+        // Check for terminal states or maximum depth
+        if depth == 0 || self.is_game_over().is_some() {
+            return self.calculate_score(player);
+        }
+
+        let mut alpha = alpha;
+        let mut beta = beta;
+
+        if maximizing_player {
+            let mut max_eval = i64::MIN;
+
+            self.for_valid_moves(|column| {
+                let mut new_state = self.clone();
+                new_state.put(player, column).unwrap();
+
+                let eval = new_state.minimax(depth - 1, alpha, beta, false, player, counter);
+                max_eval = max_eval.max(eval);
+                alpha = alpha.max(eval);
+
+                // Alpha-beta pruning
+                if beta <= alpha {
+                    return false;
+                }
+                true
+            });
+
+            max_eval
+        } else {
+            let opponent = player.opponent();
+            let mut min_eval = i64::MAX;
+
+            self.for_valid_moves(|column| {
+                let mut new_state = self.clone();
+                new_state.put(opponent, column).unwrap();
+
+                let eval = new_state.minimax(depth - 1, alpha, beta, true, player, counter);
+                min_eval = min_eval.min(eval);
+                beta = beta.min(eval);
+
+                // Alpha-beta pruning
+                if beta <= alpha {
+                    return false;
+                }
+                true
+            });
+
+            min_eval
+        }
+    }
+
+    /// Get the best move using minimax with alpha-beta pruning.
+    pub fn find_best_move(&self, depth: u8, player: Player, counter: &mut usize) -> Option<u8> {
+        let mut best_move = None;
+        let mut best_score = i64::MIN;
+
+        self.for_valid_moves(|column| {
+            let mut new_state = self.clone();
+            new_state.put(player, column).unwrap();
+
+            let score = new_state.minimax(depth - 1, i64::MIN, i64::MAX, false, player, counter);
+
+            if score > best_score {
+                best_score = score;
+                best_move = Some(column);
+            }
+
+            true
+        });
+
+        best_move
+    }
+}