def train(board: Board,
striga: Striga,
witcher: Witcher,
initial_state: QState,
epochs=1000):
total_wins = 0
hits = []
win_p = []
lifetime = []
rewards = []
time = []
heatmap_start = board.to_numpy()
heatmap = board.to_numpy()
for i in range(epochs):
game = Env(board, initial_state, striga, witcher)
while game.game_res is None:
game.advance()
if game.game_res == GameResult.WITCHER:
total_wins += 1
if i == epochs // 100:
heatmap_start = heatmap.copy()
heatmap += game.witcher_positions
time.append(i + 1)
hits.append(float(game.hits))
win_p.append(total_wins / (i + 1))
lifetime.append(game.turn)
rewards.append(game.rewards)
return hits, win_p, lifetime, rewards, time, heatmap, heatmap_start
