minimax_with_alpha_beta_en.py

def minimax(node, depth, min, max): #we add min and max - these are thresholds for pruning 
    if node.isLeaf() or depth == 0: #if node appears to be leaf or recursion limit exceeded
        return evaluate_board(node) #some evaluation function for node in a tree (returns some value)
    children = getChildrenConfigurations(node) #some function to get all possible children nodes (i.e. possible configurations) from current node
    if node.type() == max_node: #if we define current node as max-node (i.e. we play with black stones and try to maximize cost)
        best_evaluation = MIN_VALUE #initialize with minimum possible value
        for child_node in children:
            cur_evalutaion = minimax(child_node, depth - 1, best_evaluation, max) #just like we did before, but also update tresholds for pruning
            if cur_evaluation > best_evaluation: #evaluate "cost" of a possible move but recursively decrease depth
                best_evaluation = cur_evaluation
            if cur_evaluation > max: #if we exceed treshold than we can stop computing this branch of a tree (make pruning)
                return max
        return best_evaluation #return best cost found
    if node.type() == min_node: #if we define current node as min-node (i.e. we play with white stones and try to minimize cost)
        best_evaluation = MAX_VALUE #initialize with maximum possible value
        for child_node in children:
            cur_evalutaion = minimax(child_node, depth - 1, min, best_evaluation) #just like we did before, but also update tresholds for pruning
            if cur_evaluation < best_evaluation: #if we define current node as min-node (i.e. we play with white stones and try to minimize cost)
                best_evaluation = cur_evaluation
            if cur_evaluation < min: #if we exceed treshold than we can stop computing this branch of a tree (make pruning)
                return min
        return best_evaluation #return best cost found