| 1 |
//===----------------------------------------------------------------------===// |
| 2 |
// |
| 3 |
// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. |
| 4 |
// See https://llvm.org/LICENSE.txt for license information. |
| 5 |
// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception |
| 6 |
// |
| 7 |
//===----------------------------------------------------------------------===// |
| 8 |
|
| 9 |
#ifndef _LIBCPP___CXX03___RANDOM_POISSON_DISTRIBUTION_H |
| 10 |
#define _LIBCPP___CXX03___RANDOM_POISSON_DISTRIBUTION_H |
| 11 |
|
| 12 |
#include <__cxx03/__config> |
| 13 |
#include <__cxx03/__random/clamp_to_integral.h> |
| 14 |
#include <__cxx03/__random/exponential_distribution.h> |
| 15 |
#include <__cxx03/__random/is_valid.h> |
| 16 |
#include <__cxx03/__random/normal_distribution.h> |
| 17 |
#include <__cxx03/__random/uniform_real_distribution.h> |
| 18 |
#include <__cxx03/cmath> |
| 19 |
#include <__cxx03/iosfwd> |
| 20 |
#include <__cxx03/limits> |
| 21 |
|
| 22 |
#if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER) |
| 23 |
# pragma GCC system_header |
| 24 |
#endif |
| 25 |
|
| 26 |
_LIBCPP_PUSH_MACROS |
| 27 |
#include <__cxx03/__undef_macros> |
| 28 |
|
| 29 |
_LIBCPP_BEGIN_NAMESPACE_STD |
| 30 |
|
| 31 |
template <class _IntType = int> |
| 32 |
class _LIBCPP_TEMPLATE_VIS poisson_distribution { |
| 33 |
static_assert(__libcpp_random_is_valid_inttype<_IntType>::value, "IntType must be a supported integer type"); |
| 34 |
|
| 35 |
public: |
| 36 |
// types |
| 37 |
typedef _IntType result_type; |
| 38 |
|
| 39 |
class _LIBCPP_TEMPLATE_VIS param_type { |
| 40 |
double __mean_; |
| 41 |
double __s_; |
| 42 |
double __d_; |
| 43 |
double __l_; |
| 44 |
double __omega_; |
| 45 |
double __c0_; |
| 46 |
double __c1_; |
| 47 |
double __c2_; |
| 48 |
double __c3_; |
| 49 |
double __c_; |
| 50 |
|
| 51 |
public: |
| 52 |
typedef poisson_distribution distribution_type; |
| 53 |
|
| 54 |
_LIBCPP_HIDE_FROM_ABI explicit param_type(double __mean = 1.0); |
| 55 |
|
| 56 |
_LIBCPP_HIDE_FROM_ABI double mean() const { return __mean_; } |
| 57 |
|
| 58 |
friend _LIBCPP_HIDE_FROM_ABI bool operator==(const param_type& __x, const param_type& __y) { |
| 59 |
return __x.__mean_ == __y.__mean_; |
| 60 |
} |
| 61 |
friend _LIBCPP_HIDE_FROM_ABI bool operator!=(const param_type& __x, const param_type& __y) { return !(__x == __y); } |
| 62 |
|
| 63 |
friend class poisson_distribution; |
| 64 |
}; |
| 65 |
|
| 66 |
private: |
| 67 |
param_type __p_; |
| 68 |
|
| 69 |
public: |
| 70 |
// constructors and reset functions |
| 71 |
_LIBCPP_HIDE_FROM_ABI explicit poisson_distribution(double __mean = 1.0) : __p_(__mean) {} |
| 72 |
_LIBCPP_HIDE_FROM_ABI explicit poisson_distribution(const param_type& __p) : __p_(__p) {} |
| 73 |
_LIBCPP_HIDE_FROM_ABI void reset() {} |
| 74 |
|
| 75 |
// generating functions |
| 76 |
template <class _URNG> |
| 77 |
_LIBCPP_HIDE_FROM_ABI result_type operator()(_URNG& __g) { |
| 78 |
return (*this)(__g, __p_); |
| 79 |
} |
| 80 |
template <class _URNG> |
| 81 |
_LIBCPP_HIDE_FROM_ABI result_type operator()(_URNG& __g, const param_type& __p); |
| 82 |
|
| 83 |
// property functions |
| 84 |
_LIBCPP_HIDE_FROM_ABI double mean() const { return __p_.mean(); } |
| 85 |
|
| 86 |
_LIBCPP_HIDE_FROM_ABI param_type param() const { return __p_; } |
| 87 |
_LIBCPP_HIDE_FROM_ABI void param(const param_type& __p) { __p_ = __p; } |
| 88 |
|
| 89 |
_LIBCPP_HIDE_FROM_ABI result_type min() const { return 0; } |
| 90 |
_LIBCPP_HIDE_FROM_ABI result_type max() const { return numeric_limits<result_type>::max(); } |
| 91 |
|
| 92 |
friend _LIBCPP_HIDE_FROM_ABI bool operator==(const poisson_distribution& __x, const poisson_distribution& __y) { |
| 93 |
return __x.__p_ == __y.__p_; |
| 94 |
} |
| 95 |
friend _LIBCPP_HIDE_FROM_ABI bool operator!=(const poisson_distribution& __x, const poisson_distribution& __y) { |
| 96 |
return !(__x == __y); |
| 97 |
} |
| 98 |
}; |
| 99 |
|
| 100 |
template <class _IntType> |
| 101 |
poisson_distribution<_IntType>::param_type::param_type(double __mean) |
| 102 |
// According to the standard `inf` is a valid input, but it causes the |
| 103 |
// distribution to hang, so we replace it with the maximum representable |
| 104 |
// mean. |
| 105 |
: __mean_(isinf(__mean) ? numeric_limits<double>::max() : __mean) { |
| 106 |
if (__mean_ < 10) { |
| 107 |
__s_ = 0; |
| 108 |
__d_ = 0; |
| 109 |
__l_ = std::exp(-__mean_); |
| 110 |
__omega_ = 0; |
| 111 |
__c3_ = 0; |
| 112 |
__c2_ = 0; |
| 113 |
__c1_ = 0; |
| 114 |
__c0_ = 0; |
| 115 |
__c_ = 0; |
| 116 |
} else { |
| 117 |
__s_ = std::sqrt(__mean_); |
| 118 |
__d_ = 6 * __mean_ * __mean_; |
| 119 |
__l_ = std::trunc(__mean_ - 1.1484); |
| 120 |
__omega_ = .3989423 / __s_; |
| 121 |
double __b1 = .4166667E-1 / __mean_; |
| 122 |
double __b2 = .3 * __b1 * __b1; |
| 123 |
__c3_ = .1428571 * __b1 * __b2; |
| 124 |
__c2_ = __b2 - 15. * __c3_; |
| 125 |
__c1_ = __b1 - 6. * __b2 + 45. * __c3_; |
| 126 |
__c0_ = 1. - __b1 + 3. * __b2 - 15. * __c3_; |
| 127 |
__c_ = .1069 / __mean_; |
| 128 |
} |
| 129 |
} |
| 130 |
|
| 131 |
template <class _IntType> |
| 132 |
template <class _URNG> |
| 133 |
_IntType poisson_distribution<_IntType>::operator()(_URNG& __urng, const param_type& __pr) { |
| 134 |
static_assert(__libcpp_random_is_valid_urng<_URNG>::value, ""); |
| 135 |
double __tx; |
| 136 |
uniform_real_distribution<double> __urd; |
| 137 |
if (__pr.__mean_ < 10) { |
| 138 |
__tx = 0; |
| 139 |
for (double __p = __urd(__urng); __p > __pr.__l_; ++__tx) |
| 140 |
__p *= __urd(__urng); |
| 141 |
} else { |
| 142 |
double __difmuk; |
| 143 |
double __g = __pr.__mean_ + __pr.__s_ * normal_distribution<double>()(__urng); |
| 144 |
double __u; |
| 145 |
if (__g > 0) { |
| 146 |
__tx = std::trunc(__g); |
| 147 |
if (__tx >= __pr.__l_) |
| 148 |
return std::__clamp_to_integral<result_type>(__tx); |
| 149 |
__difmuk = __pr.__mean_ - __tx; |
| 150 |
__u = __urd(__urng); |
| 151 |
if (__pr.__d_ * __u >= __difmuk * __difmuk * __difmuk) |
| 152 |
return std::__clamp_to_integral<result_type>(__tx); |
| 153 |
} |
| 154 |
exponential_distribution<double> __edist; |
| 155 |
for (bool __using_exp_dist = false; true; __using_exp_dist = true) { |
| 156 |
double __e; |
| 157 |
if (__using_exp_dist || __g <= 0) { |
| 158 |
double __t; |
| 159 |
do { |
| 160 |
__e = __edist(__urng); |
| 161 |
__u = __urd(__urng); |
| 162 |
__u += __u - 1; |
| 163 |
__t = 1.8 + (__u < 0 ? -__e : __e); |
| 164 |
} while (__t <= -.6744); |
| 165 |
__tx = std::trunc(__pr.__mean_ + __pr.__s_ * __t); |
| 166 |
__difmuk = __pr.__mean_ - __tx; |
| 167 |
__using_exp_dist = true; |
| 168 |
} |
| 169 |
double __px; |
| 170 |
double __py; |
| 171 |
if (__tx < 10 && __tx >= 0) { |
| 172 |
const double __fac[] = {1, 1, 2, 6, 24, 120, 720, 5040, 40320, 362880}; |
| 173 |
__px = -__pr.__mean_; |
| 174 |
__py = std::pow(__pr.__mean_, (double)__tx) / __fac[static_cast<int>(__tx)]; |
| 175 |
} else { |
| 176 |
double __del = .8333333E-1 / __tx; |
| 177 |
__del -= 4.8 * __del * __del * __del; |
| 178 |
double __v = __difmuk / __tx; |
| 179 |
if (std::abs(__v) > 0.25) |
| 180 |
__px = __tx * std::log(1 + __v) - __difmuk - __del; |
| 181 |
else |
| 182 |
__px = __tx * __v * __v * |
| 183 |
(((((((.1250060 * __v + -.1384794) * __v + .1421878) * __v + -.1661269) * __v + .2000118) * __v + |
| 184 |
-.2500068) * |
| 185 |
__v + |
| 186 |
.3333333) * |
| 187 |
__v + |
| 188 |
-.5) - |
| 189 |
__del; |
| 190 |
__py = .3989423 / std::sqrt(__tx); |
| 191 |
} |
| 192 |
double __r = (0.5 - __difmuk) / __pr.__s_; |
| 193 |
double __r2 = __r * __r; |
| 194 |
double __fx = -0.5 * __r2; |
| 195 |
double __fy = __pr.__omega_ * (((__pr.__c3_ * __r2 + __pr.__c2_) * __r2 + __pr.__c1_) * __r2 + __pr.__c0_); |
| 196 |
if (__using_exp_dist) { |
| 197 |
if (__pr.__c_ * std::abs(__u) <= __py * std::exp(__px + __e) - __fy * std::exp(__fx + __e)) |
| 198 |
break; |
| 199 |
} else { |
| 200 |
if (__fy - __u * __fy <= __py * std::exp(__px - __fx)) |
| 201 |
break; |
| 202 |
} |
| 203 |
} |
| 204 |
} |
| 205 |
return std::__clamp_to_integral<result_type>(__tx); |
| 206 |
} |
| 207 |
|
| 208 |
template <class _CharT, class _Traits, class _IntType> |
| 209 |
_LIBCPP_HIDE_FROM_ABI basic_ostream<_CharT, _Traits>& |
| 210 |
operator<<(basic_ostream<_CharT, _Traits>& __os, const poisson_distribution<_IntType>& __x) { |
| 211 |
__save_flags<_CharT, _Traits> __lx(__os); |
| 212 |
typedef basic_ostream<_CharT, _Traits> _OStream; |
| 213 |
__os.flags(_OStream::dec | _OStream::left | _OStream::fixed | _OStream::scientific); |
| 214 |
return __os << __x.mean(); |
| 215 |
} |
| 216 |
|
| 217 |
template <class _CharT, class _Traits, class _IntType> |
| 218 |
_LIBCPP_HIDE_FROM_ABI basic_istream<_CharT, _Traits>& |
| 219 |
operator>>(basic_istream<_CharT, _Traits>& __is, poisson_distribution<_IntType>& __x) { |
| 220 |
typedef poisson_distribution<_IntType> _Eng; |
| 221 |
typedef typename _Eng::param_type param_type; |
| 222 |
__save_flags<_CharT, _Traits> __lx(__is); |
| 223 |
typedef basic_istream<_CharT, _Traits> _Istream; |
| 224 |
__is.flags(_Istream::dec | _Istream::skipws); |
| 225 |
double __mean; |
| 226 |
__is >> __mean; |
| 227 |
if (!__is.fail()) |
| 228 |
__x.param(param_type(__mean)); |
| 229 |
return __is; |
| 230 |
} |
| 231 |
|
| 232 |
_LIBCPP_END_NAMESPACE_STD |
| 233 |
|
| 234 |
_LIBCPP_POP_MACROS |
| 235 |
|
| 236 |
#endif // _LIBCPP___CXX03___RANDOM_POISSON_DISTRIBUTION_H |