#! /usr/bin/env python

import openturns as ot

ot.TESTPREAMBLE()
ot.PlatformInfo.SetNumericalPrecision(6)
ot.ResourceMap.SetAsUnsignedInteger("RandomMixture-DefaultMaxSize", 4000000)


# 2D test using FFT
# Collection of distributions
coll = ot.DistributionCollection(0)
coll.add(ot.Uniform(0, 1))
coll.add(ot.Uniform(0, 1))
coll.add(ot.Uniform(0, 1))
# Set weights
weights = ot.Matrix([[1.0, -2.0, 1.0], [1.0, 1.0, -3.0]])
# Build the RandomMixture
dist_2D = ot.RandomMixture(coll, weights)
# Defining RandomMixture
N = 16
points = ot.Indices(2, N)
mean = dist_2D.getMean()
sigma = dist_2D.getStandardDeviation()
xMin = mean - 2.9 * sigma
xMax = mean + 2.9 * sigma
grid = ot.Sample()
print("distribution = ", repr(dist_2D))
print("range = ", dist_2D.getRange())
print("mean = ", dist_2D.getMean())
print("cov = ", dist_2D.getCovariance())
print("sigma = ", dist_2D.getStandardDeviation())
print("xMin = ", xMin)
print("xMax = ", xMax)
result, grid = dist_2D.computePDF(xMin, xMax, points)
for i in range(grid.getSize()):
    print("%.6g;%.6g;%.6g" % (grid[i][0], grid[i][1], result[i][0]))
# Defining new case, involving Normal distributions
collection = ot.DistributionCollection(0)
collection.add(ot.Normal(2.0, 3.0))
collection.add(ot.Normal(1.0, 4.0))
weights = ot.Matrix([[4.0, 1.0], [2.0, 1.4]])
# Build the RandomMixture
distribution2D = ot.RandomMixture(collection, weights)
print("distribution = ", repr(distribution2D))
print("range = ", distribution2D.getRange())
print("mean = ", distribution2D.getMean())
print("cov = ", distribution2D.getCovariance())
print("sigma = ", distribution2D.getStandardDeviation())
xMin = distribution2D.getMean() - 2.9 * distribution2D.getStandardDeviation()
xMax = distribution2D.getMean() + 2.9 * distribution2D.getStandardDeviation()
result, grid = distribution2D.computePDF(xMin, xMax, points)
print("x;y;PDF")
for i in range(grid.getSize()):
    print("%.6g;%.6g;%.6g" % (grid[i][0], grid[i][1], result[i][0]))