Scaling down a column based on a condition

Question

I have a data frame looking like this.

interval = np.arange(1,11)
length = 0.5*np.random.randint(low=4, high=7, size=10)
grade = np.random.randint(low=1, high=10,