How can I correctly implement this example of using classes?

Question

In the following example, one can choose constants to depend upon the context of a future situtation.

class Constants:
    SPEEDLIGHT = 3 * 10**8
    GRAVITY = 9.81

C = Constants()
print(C.GRAVITY)
>> 9.81

That was not too difficult because each quantity is a fixed constant. But suppose I want to do something similar for functions. In this first block of code below, I specify two distributions of integrable variable x and fixed parameters a and b.

class IntegrableDistribution:
    def Gaussian(x,a,b):
        cnorm = 1 / ( b * (2 * pi)**(1/2) )
        return cnorm * np.exp( (-1) * (x-a)**2 / (2 * b**2) )
    # Gaussian = Gaussian(x,a,b)

    def Lognormal(x,a,b):
        cnorm = 1 / ( b * (2 * pi)**(1/2) )
        return cnorm * exp( (-1) * (np.log(x)-a)**2 / (2 * b**2) ) / x
    # Lognormal = Lognormal(x,a,b)

I was trying to name the distributions so that they could be callable. That resulted in an error message, hence the commented out code above. In this next block of code, I am trying to use an input to select a distribution for integration (though I feel it is extremely inefficient).

Integrable = IntegrableDistribution()

class CallIntegrableDistribution:

    def Model():

        def Pick():
            """
            1 :   Gaussian Distribution
            2 :   Lognormal Distribution
            """
            self.cmnd = cmnd
            cmnd = int(input("Pick a Distribution Model:    "))
            return cmnd

        self.cmnd = cmnd

        if cmnd == 1:
            Distribution = Integrable.Gaussian
        if cmnd == 2:
            Distribution = Integrable.Lognormal

        return Distribution

OR ALTERNATIVELY

    cmnd = {
        1: Gaussian,
        2: Lognormal,
    }

I'm not really concerned with the problem of distributions; I'm only applying it to showcase my knowns and unknowns. What are some ways to properly do this or something similar/simpler using classes or dictionaries?

1. What was the error? 2. Don't give functions uppercase names. That's reserved for classes. — Carcigenicate, Mar 23 '17 at 13:17
`TypeError: unsupported operand type(s) for *: 'property' and 'float'` — , Mar 23 '17 at 13:19

score 2 · Accepted Answer · answered Mar 23 '17 at 13:23

2

Use static methods:

class IntegrableDistribution:
    @staticmethod
    def Gaussian(x,a,b):
        cnorm = 1 / ( b * (2 * pi)**(1/2) )
        return cnorm * np.exp( (-1) * (x-a)**2 / (2 * b**2) )

    @staticmethod
    def Lognormal(x,a,b):
        cnorm = 1 / ( b * (2 * pi)**(1/2) )
        return cnorm * exp( (-1) * (np.log(x)-a)**2 / (2 * b**2) ) / x

And usage:

some_result = IntegrableDistribution.Gaussian(1, 2, 3)
another_result = IntegrableDistribution.Lognormal(1, 2, 3)

answered Mar 23 '17 at 13:23

DeepSpace

78,697
11
109
154

That makes more sense and gives me something to play around with, thanks. Per the linked doc, "A static method does not receive an implicit first argument." Does this mean I cannot pass x as a function variable instead of a fixed parameter? – Mar 23 '17 at 13:28
1

@mikey. No. It means that unlike instance methods, when calling a static method the first argument is not the instance the method was called with (which makes sense since there *is no* instance). See this answer: http://stackoverflow.com/a/1669524/1453822 – DeepSpace Mar 23 '17 at 13:31

How can I correctly implement this example of using classes?

1 Answers1