// test_nc_beta.cpp

// Copyright John Maddock 2008.

// Use, modification and distribution are subject to the
// Boost Software License, Version 1.0.
// (See accompanying file LICENSE_1_0.txt
// or copy at http://www.boost.org/LICENSE_1_0.txt)

//
// This must appear *before* any #includes, and precludes pch usage:
//
#define BOOST_MATH_ASSERT_UNDEFINED_POLICY false

#ifdef _MSC_VER
#pragma warning (disable:4127 4512)
#endif

#if !defined(TEST_FLOAT) && !defined(TEST_DOUBLE) && !defined(TEST_LDOUBLE) && !defined(TEST_REAL_CONCEPT)
#  define TEST_FLOAT
#  define TEST_DOUBLE
#  define TEST_LDOUBLE
#  define TEST_REAL_CONCEPT
#endif

#include <boost/math/concepts/real_concept.hpp> // for real_concept
#include <boost/math/distributions/non_central_beta.hpp> // for chi_squared_distribution
#include <boost/math/distributions/poisson.hpp> // for poisson_distribution
#include <boost/test/test_exec_monitor.hpp> // for test_main
#include <boost/test/results_collector.hpp>
#include <boost/test/unit_test.hpp>
#include <boost/test/floating_point_comparison.hpp> // for BOOST_CHECK_CLOSE

#include "functor.hpp"
#include "handle_test_result.hpp"
#include "test_ncbeta_hooks.hpp"

#include <iostream>
using std::cout;
using std::endl;
#include <limits>
using std::numeric_limits;

#define BOOST_CHECK_CLOSE_EX(a, b, prec, i) \
   {\
      unsigned int failures = boost::unit_test::results_collector.results( boost::unit_test::framework::current_test_case().p_id ).p_assertions_failed;\
      BOOST_CHECK_CLOSE(a, b, prec); \
      if(failures != boost::unit_test::results_collector.results( boost::unit_test::framework::current_test_case().p_id ).p_assertions_failed)\
      {\
         std::cerr << "Failure was at row " << i << std::endl;\
         std::cerr << std::setprecision(35); \
         std::cerr << "{ " << data[i][0] << " , " << data[i][1] << " , " << data[i][2];\
         std::cerr << " , " << data[i][3] << " , " << data[i][4] << " } " << std::endl;\
      }\
   }

#define BOOST_CHECK_EX(a, i) \
   {\
      unsigned int failures = boost::unit_test::results_collector.results( boost::unit_test::framework::current_test_case().p_id ).p_assertions_failed;\
      BOOST_CHECK(a); \
      if(failures != boost::unit_test::results_collector.results( boost::unit_test::framework::current_test_case().p_id ).p_assertions_failed)\
      {\
         std::cerr << "Failure was at row " << i << std::endl;\
         std::cerr << std::setprecision(35); \
         std::cerr << "{ " << data[i][0] << " , " << data[i][1] << " , " << data[i][2];\
         std::cerr << " , " << data[i][3] << " , " << data[i][4] << " } " << std::endl;\
      }\
   }

void expected_results()
{
   //
   // Define the max and mean errors expected for
   // various compilers and platforms.
   //
   const char* largest_type;
#ifndef BOOST_MATH_NO_LONG_DOUBLE_MATH_FUNCTIONS
   if(boost::math::policies::digits<double, boost::math::policies::policy<> >() == boost::math::policies::digits<long double, boost::math::policies::policy<> >())
   {
      largest_type = "(long\\s+)?double|real_concept";
   }
   else
   {
      largest_type = "long double|real_concept";
   }
#else
   largest_type = "(long\\s+)?double|real_concept";
#endif

#ifndef BOOST_MATH_NO_LONG_DOUBLE_MATH_FUNCTIONS
   if(boost::math::tools::digits<long double>() == 64)
   {
      //
      // Allow a small amount of error leakage from long double to double:
      //
      add_expected_result(
         "[^|]*",                          // compiler
         "[^|]*",                          // stdlib
         "[^|]*",                          // platform
         "double",                         // test type(s)
         "[^|]*large[^|]*",                // test data group
         "[^|]*", 5, 5);                   // test function
   }

   if(boost::math::tools::digits<long double>() == 64)
   {
      add_expected_result(
         "[^|]*",                          // compiler
         "[^|]*",                          // stdlib
         "[^|]*",                          // platform
         largest_type,                     // test type(s)
         "[^|]*medium[^|]*",               // test data group
         "[^|]*", 1200, 500);               // test function
      add_expected_result(
         "[^|]*",                          // compiler
         "[^|]*",                          // stdlib
         "[^|]*",                          // platform
         largest_type,                     // test type(s)
         "[^|]*large[^|]*",                // test data group
         "[^|]*", 40000, 6000);            // test function
   }
#endif
   //
   // Catch all cases come last:
   //
   add_expected_result(
      "[^|]*",                          // compiler
      "[^|]*",                          // stdlib
      "[^|]*",                          // platform
      largest_type,                     // test type(s)
      "[^|]*medium[^|]*",               // test data group
      "[^|]*", 700, 500);               // test function
   add_expected_result(
      "[^|]*",                          // compiler
      "[^|]*",                          // stdlib
      "[^|]*",                          // platform
      "real_concept",                   // test type(s)
      "[^|]*large[^|]*",                // test data group
      "[^|]*", 30000, 4000);             // test function
   add_expected_result(
      "[^|]*",                          // compiler
      "[^|]*",                          // stdlib
      "[^|]*",                          // platform
      largest_type,                     // test type(s)
      "[^|]*large[^|]*",                // test data group
      "[^|]*", 20000, 2000);             // test function
   //
   // Finish off by printing out the compiler/stdlib/platform names,
   // we do this to make it easier to mark up expected error rates.
   //
   std::cout << "Tests run with " << BOOST_COMPILER << ", " 
      << BOOST_STDLIB << ", " << BOOST_PLATFORM << std::endl;
}

template <class RealType>
RealType naive_pdf(RealType a, RealType b, RealType lam, RealType x)
{
   using namespace boost::math;

   RealType term = pdf(poisson_distribution<RealType>(lam/2), 0)
      * ibeta_derivative(a, b, x);
   RealType sum = term;

   int i = 1;
   while(term / sum > tools::epsilon<RealType>())
   {
      term = pdf(poisson_distribution<RealType>(lam/2), i)
      * ibeta_derivative(a + i, b, x);
      ++i;
      sum += term;
   }
   return sum;
}

template <class RealType>
void test_spot(
     RealType a,     // alpha
     RealType b,     // beta
     RealType ncp,   // non-centrality param
     RealType cs,    // Chi Square statistic
     RealType P,     // CDF
     RealType Q,     // Complement of CDF
     RealType D,     // PDF
     RealType tol)   // Test tolerance
{
   boost::math::non_central_beta_distribution<RealType> dist(a, b, ncp);
   BOOST_CHECK_CLOSE(
      cdf(dist, cs), P, tol);
   //
   // Sanity checking using the naive PDF calculation above fails at
   // float precision:
   //
   if(!boost::is_same<float, RealType>::value)
   {
      BOOST_CHECK_CLOSE(
         pdf(dist, cs), naive_pdf(dist.alpha(), dist.beta(), ncp, cs), tol);
   }
   BOOST_CHECK_CLOSE(
      pdf(dist, cs), D, tol);

   if((P < 0.99) && (Q < 0.99))
   {
      //
      // We can only check this if P is not too close to 1,
      // so that we can guarentee Q is reasonably free of error:
      //
      BOOST_CHECK_CLOSE(
         cdf(complement(dist, cs)), Q, tol);
      BOOST_CHECK_CLOSE(
            quantile(dist, P), cs, tol * 10);
      BOOST_CHECK_CLOSE(
            quantile(complement(dist, Q)), cs, tol * 10);
   }
}

template <class RealType> // Any floating-point type RealType.
void test_spots(RealType)
{
   RealType tolerance = (std::max)(
      boost::math::tools::epsilon<RealType>() * 100,
      (RealType)1e-6) * 100;
   RealType abs_tolerance = boost::math::tools::epsilon<RealType>() * 100;

   cout << "Tolerance = " << tolerance << "%." << endl;

   //
   // Spot tests use values computed by the R statistical
   // package and the pbeta and dbeta functions:
   //
   test_spot(
     RealType(2),                   // alpha
     RealType(5),                   // beta
     RealType(1),                   // non-centrality param
     RealType(0.25),                // Chi Square statistic
     RealType(0.3658349),           // CDF
     RealType(1-0.3658349),         // Complement of CDF
     RealType(2.184465),            // PDF
     RealType(tolerance));
   test_spot(
     RealType(20),                  // alpha
     RealType(15),                  // beta
     RealType(35),                  // non-centrality param
     RealType(0.75),                // Chi Square statistic
     RealType(0.6994175),           // CDF
     RealType(1-0.6994175),         // Complement of CDF
     RealType(5.576146),            // PDF
     RealType(tolerance));
   test_spot(
     RealType(100),                 // alpha
     RealType(3),                   // beta
     RealType(63),                  // non-centrality param
     RealType(0.95),                // Chi Square statistic
     RealType(0.03529306),          // CDF
     RealType(1-0.03529306),        // Complement of CDF
     RealType(3.637894),            // PDF
     RealType(tolerance));
   test_spot(
     RealType(0.25),                // alpha
     RealType(0.75),                // beta
     RealType(150),                 // non-centrality param
     RealType(0.975),               // Chi Square statistic
     RealType(0.09752216),          // CDF
     RealType(1-0.09752216),        // Complement of CDF
     RealType(8.020935),            // PDF
     RealType(tolerance));

   BOOST_MATH_STD_USING
   boost::math::non_central_beta_distribution<RealType> dist(100, 3, 63);
   BOOST_CHECK_CLOSE(mean(dist), RealType(4.82280451915522329944315287538684030781836554279474240490936e13L) * exp(-RealType(31.5)) * 100 / 103, tolerance);
   // Variance only guarentees small absolute error:
   BOOST_CHECK_SMALL(variance(dist) 
      - static_cast<RealType>(RealType(4.85592267707818899235900237275021938334418424134218087127572e13L)
      * exp(RealType(-31.5)) * 100 * 101 / (103 * 104) - 
      RealType(4.82280451915522329944315287538684030781836554279474240490936e13L) * RealType(4.82280451915522329944315287538684030781836554279474240490936e13L) 
      * exp(RealType(-63)) * 10000 / (103 * 103)), abs_tolerance);
   BOOST_CHECK_THROW(skewness(dist), boost::math::evaluation_error);
   BOOST_CHECK_THROW(kurtosis(dist), boost::math::evaluation_error);
   BOOST_CHECK_THROW(kurtosis_excess(dist), boost::math::evaluation_error);
} // template <class RealType>void test_spots(RealType)

template <class T>
T nc_beta_cdf(T a, T b, T nc, T x)
{
   return cdf(boost::math::non_central_beta_distribution<T>(a, b, nc), x);
}

template <class T>
T nc_beta_ccdf(T a, T b, T nc, T x)
{
   return cdf(complement(boost::math::non_central_beta_distribution<T>(a, b, nc), x));
}

template <typename T>
void do_test_nc_chi_squared(T& data, const char* type_name, const char* test)
{
   typedef typename T::value_type row_type;
   typedef typename row_type::value_type value_type;

   std::cout << "Testing: " << test << std::endl;

   value_type (*fp1)(value_type, value_type, value_type, value_type) = nc_beta_cdf;
   boost::math::tools::test_result<value_type> result;

   result = boost::math::tools::test(
      data,
      bind_func(fp1, 0, 1, 2, 3),
      extract_result(4));
   handle_test_result(result, data[result.worst()], result.worst(),
      type_name, "CDF", test);

   fp1 = nc_beta_ccdf;
   result = boost::math::tools::test(
      data,
      bind_func(fp1, 0, 1, 2, 3),
      extract_result(5));
   handle_test_result(result, data[result.worst()], result.worst(),
      type_name, "CCDF", test);

#ifdef TEST_OTHER
   fp1 = other::ncbeta_cdf;
   result = boost::math::tools::test(
      data,
      bind_func(fp1, 0, 1, 2, 3),
      extract_result(4));
   handle_test_result(result, data[result.worst()], result.worst(),
      type_name, "Other::CDF", test);
#endif
   std::cout << std::endl;

}

template <typename T>
void quantile_sanity_check(T& data, const char* type_name, const char* test)
{
   typedef typename T::value_type row_type;
   typedef typename row_type::value_type value_type;

   //
   // Tests with type real_concept take rather too long to run, so
   // for now we'll disable them:
   //
   if(!boost::is_floating_point<value_type>::value)
      return;

   std::cout << "Testing: " << type_name << " quantile sanity check, with tests " << test << std::endl;

   //
   // These sanity checks test for a round trip accuracy of one half
   // of the bits in T, unless T is type float, in which case we check
   // for just one decimal digit.  The problem here is the sensitivity
   // of the functions, not their accuracy.  This test data was generated
   // for the forward functions, which means that when it is used as
   // the input to the inverses then it is necessarily inexact.  This rounding
   // of the input is what makes the data unsuitable for use as an accuracy check,
   // and also demonstrates that you can't in general round-trip these functions.
   // It is however a useful sanity check.
   //
   value_type precision = static_cast<value_type>(ldexp(1.0, 1-boost::math::policies::digits<value_type, boost::math::policies::policy<> >()/2)) * 100;
   if(boost::math::policies::digits<value_type, boost::math::policies::policy<> >() < 50)
      precision = 1;   // 1% or two decimal digits, all we can hope for when the input is truncated to float

   for(unsigned i = 0; i < data.size(); ++i)
   {
      //
      // Test case 493 fails at float precision: not enough bits to get
      // us back where we started:
      //
      if((i == 493) && boost::is_same<float, value_type>::value)
         continue;

      if(data[i][4] == 0)
      {
         BOOST_CHECK(0 == quantile(boost::math::non_central_beta_distribution<value_type>(data[i][0], data[i][1], data[i][2]), data[i][4]));
      }
      else if(data[i][4] < 0.9999f)
      {
         value_type p = quantile(boost::math::non_central_beta_distribution<value_type>(data[i][0], data[i][1], data[i][2]), data[i][4]);
         value_type pt = data[i][3];
         BOOST_CHECK_CLOSE_EX(pt, p, precision, i);
      }
      if(data[i][5] == 0)
      {
         BOOST_CHECK(1 == quantile(complement(boost::math::non_central_beta_distribution<value_type>(data[i][0], data[i][1], data[i][2]), data[i][5])));
      }
      else if(data[i][5] < 0.9999f)
      {
         value_type p = quantile(complement(boost::math::non_central_beta_distribution<value_type>(data[i][0], data[i][1], data[i][2]), data[i][5]));
         value_type pt = data[i][3];
         BOOST_CHECK_CLOSE_EX(pt, p, precision, i);
      }
      if(boost::math::tools::digits<value_type>() > 50)
      {
         //
         // Sanity check mode, accuracy of
         // the mode is at *best* the square root of the accuracy of the PDF:
         //
         value_type m = mode(boost::math::non_central_beta_distribution<value_type>(data[i][0], data[i][1], data[i][2]));
         if((m == 1) || (m == 0))
            break;
         value_type p = pdf(boost::math::non_central_beta_distribution<value_type>(data[i][0], data[i][1], data[i][2]), m);
         if(m * (1 + sqrt(precision) * 10) < 1)
         {
            BOOST_CHECK_EX(pdf(boost::math::non_central_beta_distribution<value_type>(data[i][0], data[i][1], data[i][2]), m * (1 + sqrt(precision) * 10)) <= p, i);
         }
         if(m * (1 - sqrt(precision)) * 10 > boost::math::tools::min_value<value_type>())
         {
            BOOST_CHECK_EX(pdf(boost::math::non_central_beta_distribution<value_type>(data[i][0], data[i][1], data[i][2]), m * (1 - sqrt(precision)) * 10) <= p, i);
         }
      }
   }
}

template <typename T>
void test_accuracy(T, const char* type_name)
{
#if !defined(TEST_DATA) || (TEST_DATA == 1)
#include "ncbeta.ipp"
    do_test_nc_chi_squared(ncbeta, type_name, "Non Central Beta, medium parameters");
    quantile_sanity_check(ncbeta, type_name, "Non Central Beta, medium parameters");
#endif
#if !defined(TEST_DATA) || (TEST_DATA == 2)
#include "ncbeta_big.ipp"
    do_test_nc_chi_squared(ncbeta_big, type_name, "Non Central Beta, large parameters");
    // Takes too long to run:
    // quantile_sanity_check(ncbeta_big, type_name, "Non Central Beta, large parameters");
#endif
}

int test_main(int, char* [])
{
   BOOST_MATH_CONTROL_FP;
   // Basic sanity-check spot values.
    expected_results();
   // (Parameter value, arbitrarily zero, only communicates the floating point type).
#ifdef TEST_FLOAT
   test_spots(0.0F); // Test float.
#endif
#ifdef TEST_DOUBLE
   test_spots(0.0); // Test double.
#endif
#ifndef BOOST_MATH_NO_LONG_DOUBLE_MATH_FUNCTIONS
#ifdef TEST_LDOUBLE
   test_spots(0.0L); // Test long double.
#endif
#if !BOOST_WORKAROUND(__BORLANDC__, BOOST_TESTED_AT(0x582))
#ifdef TEST_REAL_CONCEPT
   test_spots(boost::math::concepts::real_concept(0.)); // Test real concept.
#endif
#endif
#endif

#ifdef TEST_FLOAT
   test_accuracy(0.0F, "float"); // Test float.
#endif
#ifdef TEST_DOUBLE
   test_accuracy(0.0, "double"); // Test double.
#endif
#ifndef BOOST_MATH_NO_LONG_DOUBLE_MATH_FUNCTIONS
#ifdef TEST_LDOUBLE
   test_accuracy(0.0L, "long double"); // Test long double.
#endif
#if !BOOST_WORKAROUND(__BORLANDC__, BOOST_TESTED_AT(0x582))
#ifdef TEST_REAL_CONCEPT
   test_accuracy(boost::math::concepts::real_concept(0.), "real_concept"); // Test real concept.
#endif
#endif
#endif
   return 0;
} // int test_main(int, char* [])