Drunk: Weighted Random for Python

drunk is a simple module that lets you make simple random operations with customizable weights. It depends to built-in random module.

Installation

Just install it via pip. It's compatible with 2.7 and 3.x versions of Python.

pip install drunk

Use

drunk.static_picker(bundle, weight_key)

Creates a generator that chooses a random element from bundle Weight for each element is calculated by the weight_key. weight_key should provide a positive int or a positive float value.

drunk.choice(bundle, weight_key)

Choses a random element from bundle. Weight for each element is calculated by the weight_key. weight_key should provide a positive int or a positive float value.

drunk.shuffle(bundle, weight_key)

Creates a shuffled version of the bundle. weight_key calculates the weights. Assume A and B in bundle. If A's weight greater than B's, then you'll see A before B in the shuffled bundle, more frequently (that means weight).

drunk.sample(bundle, size, weight_key)

Gives a size sized sub-bundle of the bundle. weight_key calculates the weights. If weight_key returns a higher value for an element, The probability of chosing that element is also higher.

---

If you don't provide a weight_key;

drunk.choice will behave like random.choice

drunk.shuffle will behave like random.shuffle

Examples

static_picker example:

import drunk

my_pretty_bundle = ['Alpay', 'Gandalf', 'Kenobi', 'Amca']
weight = lambda x: len(x)  # Assume that weights are related with the length

picker = static_picker(my_pretty_bundle)

choices = [next(picker) for i in range(10)]
print(choices)  # returns a list filled with 10 picks.

choice example:

Here's a dummy class named ABasicClass, a list of some of its instances named my_pretty_bundle and a weight_key to calculate weights. drunk.choice choses an element according to the weights which are calculated by an inline function.

import drunk

class ABasicClass(object):
	def __init__(self, name, weight):
		self.name = name
		self.weight = weight
	def __repr__(self):
		return str(self.name)

my_pretty_bundle = []
my_pretty_bundle.append(ABasicClass("Amca", 40))
my_pretty_bundle.append(ABasicClass("Alpay", 30))
my_pretty_bundle.append(ABasicClass("Gandalf", 20))
my_pretty_bundle.append(ABasicClass("Kenobi", 10))

choosen_one = drunk.choice(my_pretty_bundle, lambda x: x.weight)
print("You picked ", choosen_one, "!!1!one!")

Shuffle example: Here's an list named my_pretty_list. The weight of an element is its length.

import drunk
my_pretty_list = ["Doctor Who", "Banana", "Meh", "Apple"]

print(drunk.shuffle(my_pretty_list, weight_key=lambda x: len(x)))

Sampling example: Let there be a bundle. We want random 5 elements of it. There is also a f function that calculates the weight of each element in bundle. Let the weight of an element is itself.

import drunk
my_sexy_bundle = [1, 2, 3, 4]

print(drunk.sample(my_pretty_list, weight_key=lambda x: x))

Or you can get some of that bundle which means a random sized sub-bundle.

import drunk
my_sexy_bundle = [1, 2, 3, 4]

print(drunk.sample(my_pretty_list, weight_key=lambda x: x))