我如何将变量从if语句分配给外部源？

Question

我基本上一直在尝试一个简单的q学习方法，以尝试更好地使用ML，但是我有一个非常简单的问题，我无法解决，

for i in range(runs * amt_per_step):

    done = None
    didgood = None
    newstate = None
    lastq = None
    results = None

    starter = env.start()[0]
    render = False
    if i % 50 == 0:
        render = True
    if i == 0:
        pass

    if i == 1:
        action = np.argmax(get_discrete_state(starter)) + 1
        a, b, c, d, e = env.step(starter, action)
        done = a
        didgood = b
        newstate = c
        lastq = d
        results = e
        if didgood == False:
            reward = -1
        else:
            reward = 0

        new_q =  (1 - learning_rate) * lastq + learning_rate * (reward + discount * 3)

应该是step（）函数为其分配的内容，而不是newstate和所有其他变量为None