from Player import ComputerPlayer, HumanPlayer
from State import State

BOARD_ROWS = 3
BOARD_COLS = 3


def tic_tac_toe():
    # train with 2 computer player
    p1 = ComputerPlayer("p1", board_cols=BOARD_COLS, board_rows=BOARD_ROWS)
    p2 = ComputerPlayer("p2", board_cols=BOARD_COLS, board_rows=BOARD_ROWS)

    st = State(p1, p2, board_cols=BOARD_COLS, board_rows=BOARD_ROWS)
    print("training...")
    st.play(50000)
    p1.save_policy()

    # play with human
    p1 = ComputerPlayer("computer", board_cols=BOARD_COLS, board_rows=BOARD_ROWS, exp_rate=0)
    p1.load_policy("policy_p1")

    p2 = HumanPlayer("human")

    st = State(p1, p2, board_cols=BOARD_COLS, board_rows=BOARD_ROWS)
    st.play2()


if __name__ == "__main__":
    tic_tac_toe()