def test_cookie_problem(self):
bowl_1 = distribution.Distribution(['vanilla'] * 30 +
['chocolate'] * 10,
name='bowl_1')
bowl_2 = distribution.Distribution(['vanilla'] * 20 +
['chocolate'] * 20,
name='bowl_2')
# given these two bowls, what is the chance that a vanilla cookie comes
# from bowl_1?
cookies = distribution.DistGroup([bowl_1, bowl_2])
data = 'vanilla'
cookies.normalize()
probabilities = dict([
(dist.name,
(cookies.P(dist) * dist.P(data)) / cookies.normalizer(data))
for dist in cookies
])
assert cookies.P('bowl_1') == 0.5, "prior probs are wrong"
assert cookies.normalizer(data) == 0.625, "overall prob is wrong"
assert len(probabilities) == 2, "not all distributions are present"
assert probabilities[
'bowl_1'] == 0.6, "probability for bowl_1 should be 0.6"
assert probabilities[
'bowl_2'] == 0.4, "probability for bowl_2 should be 0.4"
def test_variance(self):
text = get_books(3)
frequency = distribution.Distribution(
[s.strip('.,!?\'" ') for s in ' '.join(text).split()])
assert frequency.variance() == statistics.pvariance(frequency.values())
#!/usr/bin/env python
# -*- coding: utf-8 -*-
from distribution import distribution
from pathlib import Path
import statistics
d1 = distribution.Distribution({'chocolate': 10, 'vanilla': 30})
d2 = distribution.Distribution((('chocolate', 20), ('vanilla', 20)))
d3 = distribution.Distribution(['chocolate'] * 30 + ['vanilla'] * 10)
def get_books(n=0):
data_dir = Path('/home/niels/projects/python/bouquet/data')
books = [b for b in data_dir.glob('*/*') if b.suffix == '.txt']
text = []
if n > len(books):
print('{} books available, using all books'.format(len(books)))
for idx, book in enumerate(books):
if n > 0 and idx >= n:
break
else:
with book.open() as f:
text.append(f.read())
return text
class TestDistribution:
def test_itialization(self):
assert d1['vanilla'] is 30
assert d2['vanilla'] is 20