Python: argument of a class in Think Bayes by Allen B. Downey

Hi Derek,

I don't know anything about Bayes' Theorem but I can try explaining how the code is being processed when you create a Dice object.

You were able to trace it up to the _DictWrapper class which does contain an __init__ method. This is the method that will be called when you initialize your Dice object. The list will be stored in the local parameter values

Here's the code that is doing something with the argument that is passed in to initialize it:

init_methods = [
    self.InitPmf,
    self.InitMapping,
    self.InitSequence,
    self.InitFailure,
    ]

for method in init_methods:
    try:
        method(values)
        break
    except AttributeError:
        continue

A list of 4 methods in the _DictWrapper class are set up. These are then tried in turn with the for loop. If a method can't process the values passed in then a AttributeError will be raised and it will continue to the next method.

InitPmf will fail because it's expecting a Pmf object and a list was passed in. InitMapping will fail as well because it's expecting a mapping of values to probabilities.

InitSequence is the one that should work out since it's expecting a sequence and a list is a sequence.

InitFailure raises a ValueError if none of the first 3 methods work out.