cli.py

""" Calculates Bayesian Probabilty that A - B > x

Usage:
  ab_test.py <a_population> <a_success> <b_population> <b_success>
             [--sample_size=<n>] [--absolute_effect_size=<x> | --relative_effect_size=<x>]
             [--alpha=<a>] [--beta=<b>]

Options:
  --sample_size=<n>   Select a sample size [default: 10000]
  --absolute_effect_size=<x>   Select the absolute size of effect to be detected [default: 0.0]
  --relative_effect_size=<x>   Select the relative size of effect to be detected
  --alpha=<a>         Select an alpha [default: 1.0]
  --beta=<b>          Select a beta [default: 1.0]

"""
from __future__ import print_function

from docopt import docopt
from ab_test import probability_relative_effect, probability_absolute_effect


def main():
    arguements = docopt(__doc__)

    a_population = int(arguements['<a_population>'])
    b_population = int(arguements['<b_population>'])
    a_success = int(arguements['<a_success>'])
    b_success = int(arguements['<b_success>'])
    sample_size = int(arguements['--sample_size'])
    absolute_effect_size = float(arguements['--absolute_effect_size'])
    if arguements['--relative_effect_size']:
        relative_effect_size = float(arguements['--relative_effect_size'])
    else:
        relative_effect_size = None

    alpha = float(arguements['--alpha'])
    beta = float(arguements['--beta'])

    if relative_effect_size:
        print(probability_relative_effect(
            a_population, a_success, b_population, b_success,
            alpha, beta, sample_size, relative_effect_size))
    else:
        print(probability_absolute_effect(
            a_population, a_success, b_population, b_success,
            alpha, beta, sample_size, absolute_effect_size))


if __name__ == "__main__":
    main()