BLM.py 文件源码

def __init__(self, maxUmbralAstral):
        # Print debug
        self.debug = False

        # Outer bound for Astral Fire and Umbral Ice
        BLM.MAXUMBRALASTRAL = maxUmbralAstral

        # Available buffs
        self.BUFFS = []

        # Maximum time available
        self.MAXTIME = 45

        self.HELPER = BLM.Helper()

        # Available abilities
        self.ABILITIES = [
            BLM.Ability("Blizzard 1", 180, 6,  2.5,  2.49, self.HELPER.UmbralIceIncrease, BLM.DamageType.Ice, self.HELPER), #480
            BLM.Ability("Fire 1",     180, 15, 2.5,  2.49, self.HELPER.AstralFireIncrease, BLM.DamageType.Fire, self.HELPER), #1200
            BLM.Ability("Transpose",  0,   0,  0.75, 12.9, self.HELPER.SwapAstralUmbral, BLM.DamageType.Neither, self.HELPER),
            BLM.Ability("Fire 3",     240, 30, 3.5,  2.5, self.HELPER.AstralFireMax, BLM.DamageType.Fire, self.HELPER), #2400
            BLM.Ability("Blizzard 3", 240, 18, 3.5,  2.5, self.HELPER.UmbralIceMax, BLM.DamageType.Ice, self.HELPER), #2400
            BLM.Ability("Fire 4",     260, 15, 2.8,  2.5, None, BLM.DamageType.Fire, self.HELPER)] #2400

        # State including ability cooldowns, buff time remaining, mana, and Astral/Umbral
        self.initialState = np.array([0] * (len(self.ABILITIES) + len(self.BUFFS)) + [BLM.MAXMANA] + [0])

        self.state = self._reset()

        # What the learner can pick between
        self.action_space = spaces.Discrete(len(self.ABILITIES))

        # What the learner can see to make a choice (cooldowns and buffs)
        self.observation_space = spaces.MultiDiscrete([[0,180]] * (len(self.ABILITIES) + len(self.BUFFS)) + [[0, BLM.MAXMANA]] + [[-3,3]])