choices 0.1

Making choices from a probability distribution

The PDist is designed to make it easy to make it easy to make things happen in a probablistic manner. Provide keys with an associated probability attached, and then call the resulting object when you need a new value.

>>> mood = PDist({'happy': 0.3, 'neutral': 0.6, 'sad': 0.1})
>>> mood()
'happy'
>>> mood()
'neutral'

You can retrieve the distribution of how those values are applied by accessing the PDist.distribution attribute:

>>> mood.distribution
[('happy', 0.3), ('neutral', 0.6), ('sad', 0.1)]

As well as providing a dict to assign probabilities, you can send in a list of lists/tuples. This allows for for non-hashable types to be used. This means that you provide objects such as functions to be used as keys:

>>> def grumpy(news):
...    return ':/'
>>> def happy(news):
...    return ':)'
>>> react = PDist([(grumpy, 0.7), (happy, 0.3)])
>>> reaction = react()
>>> reaction("We're getting married!")
':/'

You are not restricted to adding your input variables up to 1. For example, if you only have a tally chart, and wish to calculate a probability distribution from that sample, you can provide that too:

>>> spotted = {'geese': 0, 'ducks': 12, 'sparrows': 4, 'other': 39}
>>> bird_pdist = PDist(spotted)
>>> bird_pdist.distribution
[('geese', 0.0), ('sparrows', 0.07272727272727272), ('ducks', 0.21818181818181817), ('other', 0.7090909090909091)]

Retrieving the probability of a particular key is supported. However, the search strategy is very inefficient and probably shouldn’t be used outside of the interactive Python shell.

>>> bird_pdist['geese']
0.0