scipy-yli/tests/test_beta_ratio.py

#   scipy-yli: Helpful SciPy utilities and recipes
#   Copyright © 2022  Lee Yingtong Li (RunasSudo)
#
#   This program is free software: you can redistribute it and/or modify
#   it under the terms of the GNU Affero General Public License as published by
#   the Free Software Foundation, either version 3 of the License, or
#   (at your option) any later version.
#
#   This program is distributed in the hope that it will be useful,
#   but WITHOUT ANY WARRANTY; without even the implied warranty of
#   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
#   GNU Affero General Public License for more details.
#
#   You should have received a copy of the GNU Affero General Public License
#   along with this program.  If not, see <https://www.gnu.org/licenses/>.

from pytest import approx

import numpy as np
from scipy import stats

import yli

def test_beta_ratio_vs_jsaffer_pdf():
	"""Compare beta_ratio.pdf with result from https://github.com/jsaffer/beta_quotient_distribution"""
	
	# Define the example beta distribution
	a1, b1, a2, b2 = 3, 6, 12, 7
	
	# Compute PDF
	x = np.linspace(0, 2, 100)
	dist = yli.beta_ratio(a1, b1, a2, b2)
	y = dist.pdf(x)
	
	# Compare with expected values from jsaffer implementation
	expected = np.load('tests/data/beta_ratio_vs_jsaffer.npy', allow_pickle=False)[0]
	assert y == approx(expected)

def test_beta_ratio_vs_jsaffer_cdf():
	"""Compare beta_ratio.cdf with result from https://github.com/jsaffer/beta_quotient_distribution"""
	
	# Define the example beta distribution
	a1, b1, a2, b2 = 3, 6, 12, 7
	
	# Compute PDF
	x = np.linspace(0, 2, 100)
	dist = yli.beta_ratio(a1, b1, a2, b2)
	y = dist.cdf(x)
	
	# Compare with expected values from jsaffer implementation
	expected = np.load('tests/data/beta_ratio_vs_jsaffer.npy', allow_pickle=False)[1]
	assert y == approx(expected)

def _gen_beta_ratio_vs_jsaffer():
	"""Generate beta_ratio_vs_jsaffer.npy for test_beta_ratio_vs_jsaffer_pdf/cdf"""
	
	import beta_quotient_distribution
	
	a1, b1, a2, b2 = 3, 6, 12, 7
	
	x = np.linspace(0, 2, 100)
	y1 = np.vectorize(lambda w: float(beta_quotient_distribution.pdf_bb_ratio(a1, a2, b1, b2, w)))(x)
	y2 = np.vectorize(lambda w: float(beta_quotient_distribution.cdf_bb_ratio(a1, a2, b1, b2, w)))(x)
	
	np.save('tests/data/beta_ratio_vs_jsaffer.npy', np.array([y1, y2]), allow_pickle=False)

def test_beta_ratio_mean_vs_empirical():
	"""Compare beta_ratio.mean (via beta_ratio._munp) with empirical mean"""
	
	# Define the example beta distribution
	beta1 = stats.beta(3, 6)
	beta2 = stats.beta(12, 7)
	dist = yli.beta_ratio.from_scipy(beta1, beta2)
	
	# Compute empirical distribution
	samples_p1 = beta1.rvs(10_000, random_state=31415)
	samples_p2 = beta2.rvs(10_000, random_state=92653)
	sample = samples_p1 / samples_p2
	
	# Allow 0.01 tolerance
	assert dist.mean() == approx(sample.mean(), abs=0.01)

def test_beta_ratio_var_vs_empirical():
	"""Compare beta_ratio.var (via beta_ratio._munp) with empirical variance"""
	
	# Define the example beta distribution
	beta1 = stats.beta(3, 6)
	beta2 = stats.beta(12, 7)
	dist = yli.beta_ratio.from_scipy(beta1, beta2)
	
	# Compute empirical distribution
	samples_p1 = beta1.rvs(10_000, random_state=31415)
	samples_p2 = beta2.rvs(10_000, random_state=92653)
	sample = samples_p1 / samples_p2
	
	# Allow 0.01 tolerance
	assert dist.var() == approx(sample.var(), abs=0.01)