Thanks to visit codestin.com
Credit goes to github.com

Skip to content

bpo-40541: Add optional *counts* parameter to random.sample() #19970

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 10 commits into from
May 8, 2020
21 changes: 14 additions & 7 deletions Doc/library/random.rst
Original file line number Diff line number Diff line change
Expand Up @@ -217,7 +217,7 @@ Functions for sequences
The optional parameter *random*.


.. function:: sample(population, k)
.. function:: sample(population, k, *, counts=None)

Return a *k* length list of unique elements chosen from the population sequence
or set. Used for random sampling without replacement.
Expand All @@ -231,13 +231,21 @@ Functions for sequences
Members of the population need not be :term:`hashable` or unique. If the population
contains repeats, then each occurrence is a possible selection in the sample.

Repeated elements can be specified one at a time or with the optional
keyword-only *counts* parameter. For example, ``sample(['red', 'blue'],
counts=[4, 2], k=5)`` is equivalent to ``sample(['red', 'red', 'red', 'red',
'blue', 'blue'], k=5)``.

To choose a sample from a range of integers, use a :func:`range` object as an
argument. This is especially fast and space efficient for sampling from a large
population: ``sample(range(10000000), k=60)``.

If the sample size is larger than the population size, a :exc:`ValueError`
is raised.

.. versionchanged:: 3.9
Added the *counts* parameter.

.. deprecated:: 3.9
In the future, the *population* must be a sequence. Instances of
:class:`set` are no longer supported. The set must first be converted
Expand Down Expand Up @@ -420,12 +428,11 @@ Simulations::
>>> choices(['red', 'black', 'green'], [18, 18, 2], k=6)
['red', 'green', 'black', 'black', 'red', 'black']

>>> # Deal 20 cards without replacement from a deck of 52 playing cards
>>> # and determine the proportion of cards with a ten-value
>>> # (a ten, jack, queen, or king).
>>> deck = collections.Counter(tens=16, low_cards=36)
>>> seen = sample(list(deck.elements()), k=20)
>>> seen.count('tens') / 20
>>> # Deal 20 cards without replacement from a deck
>>> # of 52 playing cards, and determine the proportion of cards
>>> # with a ten-value: ten, jack, queen, or king.
>>> dealt = sample(['tens', 'low cards'], counts=[16, 36], k=20)
>>> dealt.count('tens') / 20
0.15

>>> # Estimate the probability of getting 5 or more heads from 7 spins
Expand Down
34 changes: 29 additions & 5 deletions Lib/random.py
Original file line number Diff line number Diff line change
Expand Up @@ -331,7 +331,7 @@ def shuffle(self, x, random=None):
j = _int(random() * (i+1))
x[i], x[j] = x[j], x[i]

def sample(self, population, k):
def sample(self, population, k, *, counts=None):
"""Chooses k unique random elements from a population sequence or set.

Returns a new list containing elements from the population while
Expand All @@ -344,9 +344,21 @@ def sample(self, population, k):
population contains repeats, then each occurrence is a possible
selection in the sample.

To choose a sample in a range of integers, use range as an argument.
This is especially fast and space efficient for sampling from a
large population: sample(range(10000000), 60)
Repeated elements can be specified one at a time or with the optional
counts parameter. For example:

sample(['red', 'blue'], counts=[4, 2], k=5)

is equivalent to:

sample(['red', 'red', 'red', 'red', 'blue', 'blue'], k=5)

To choose a sample from a range of integers, use range() for the
population argument. This is especially fast and space efficient
for sampling from a large population:

sample(range(10000000), 60)

"""

# Sampling without replacement entails tracking either potential
Expand Down Expand Up @@ -379,8 +391,20 @@ def sample(self, population, k):
population = tuple(population)
if not isinstance(population, _Sequence):
raise TypeError("Population must be a sequence. For dicts or sets, use sorted(d).")
randbelow = self._randbelow
n = len(population)
if counts is not None:
cum_counts = list(_accumulate(counts))
if len(cum_counts) != n:
raise ValueError('The number of counts does not match the population')
total = cum_counts.pop()
if not isinstance(total, int):
raise TypeError('Counts must be integers')
if total <= 0:
raise ValueError('Total of counts must be greater than zero')
selections = sample(range(total), k=k)
bisect = _bisect
return [population[bisect(cum_counts, s)] for s in selections]
randbelow = self._randbelow
if not 0 <= k <= n:
raise ValueError("Sample larger than population or is negative")
result = [None] * k
Expand Down
73 changes: 72 additions & 1 deletion Lib/test/test_random.py
Original file line number Diff line number Diff line change
Expand Up @@ -9,7 +9,7 @@
from math import log, exp, pi, fsum, sin, factorial
from test import support
from fractions import Fraction

from collections import Counter

class TestBasicOps:
# Superclass with tests common to all generators.
Expand Down Expand Up @@ -161,6 +161,77 @@ def test_sample_on_sets(self):
population = {10, 20, 30, 40, 50, 60, 70}
self.gen.sample(population, k=5)

def test_sample_with_counts(self):
sample = self.gen.sample

# General case
colors = ['red', 'green', 'blue', 'orange', 'black', 'brown', 'amber']
counts = [500, 200, 20, 10, 5, 0, 1 ]
k = 700
summary = Counter(sample(colors, counts=counts, k=k))
self.assertEqual(sum(summary.values()), k)
for color, weight in zip(colors, counts):
self.assertLessEqual(summary[color], weight)
self.assertNotIn('brown', summary)

# Case that exhausts the population
k = sum(counts)
summary = Counter(sample(colors, counts=counts, k=k))
self.assertEqual(sum(summary.values()), k)
for color, weight in zip(colors, counts):
self.assertLessEqual(summary[color], weight)
self.assertNotIn('brown', summary)

# Case with population size of 1
summary = Counter(sample(['x'], counts=[10], k=8))
self.assertEqual(summary, Counter(x=8))

# Case with all counts equal.
nc = len(colors)
summary = Counter(sample(colors, counts=[10]*nc, k=10*nc))
self.assertEqual(summary, Counter(10*colors))

# Test error handling
with self.assertRaises(TypeError):
sample(['red', 'green', 'blue'], counts=10, k=10) # counts not iterable
with self.assertRaises(ValueError):
sample(['red', 'green', 'blue'], counts=[-3, -7, -8], k=2) # counts are negative
with self.assertRaises(ValueError):
sample(['red', 'green', 'blue'], counts=[0, 0, 0], k=2) # counts are zero
with self.assertRaises(ValueError):
sample(['red', 'green'], counts=[10, 10], k=21) # population too small
with self.assertRaises(ValueError):
sample(['red', 'green', 'blue'], counts=[1, 2], k=2) # too few counts
with self.assertRaises(ValueError):
sample(['red', 'green', 'blue'], counts=[1, 2, 3, 4], k=2) # too many counts

def test_sample_counts_equivalence(self):
# Test the documented strong equivalence to a sample with repeated elements.
# We run this test on random.Random() which makes deterministic selections
# for a given seed value.
sample = random.sample
seed = random.seed

colors = ['red', 'green', 'blue', 'orange', 'black', 'amber']
counts = [500, 200, 20, 10, 5, 1 ]
k = 700
seed(8675309)
s1 = sample(colors, counts=counts, k=k)
seed(8675309)
expanded = [color for (color, count) in zip(colors, counts) for i in range(count)]
self.assertEqual(len(expanded), sum(counts))
s2 = sample(expanded, k=k)
self.assertEqual(s1, s2)

pop = 'abcdefghi'
counts = [10, 9, 8, 7, 6, 5, 4, 3, 2]
seed(8675309)
s1 = ''.join(sample(pop, counts=counts, k=30))
expanded = ''.join([letter for (letter, count) in zip(pop, counts) for i in range(count)])
seed(8675309)
s2 = ''.join(sample(expanded, k=30))
self.assertEqual(s1, s2)

def test_choices(self):
choices = self.gen.choices
data = ['red', 'green', 'blue', 'yellow']
Expand Down
Original file line number Diff line number Diff line change
@@ -0,0 +1 @@
Added an optional *counts* parameter to random.sample().