/* boost random/lognormal_distribution.hpp header file
 *
 * Copyright Jens Maurer 2000-2001
 * Distributed under 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)
 *
 * See http://www.boost.org for most recent version including documentation.
 *
 * $Id: lognormal_distribution.hpp 60755 2010-03-22 00:45:06Z steven_watanabe $
 *
 * Revision history
 *  2001-02-18  moved to individual header files
 */

#ifndef BOOST_RANDOM_LOGNORMAL_DISTRIBUTION_HPP
#define BOOST_RANDOM_LOGNORMAL_DISTRIBUTION_HPP

#include <boost/config/no_tr1/cmath.hpp>      // std::exp, std::sqrt
#include <cassert>
#include <iostream>
#include <boost/limits.hpp>
#include <boost/static_assert.hpp>
#include <boost/random/detail/config.hpp>
#include <boost/random/normal_distribution.hpp>

#ifdef BOOST_NO_STDC_NAMESPACE
namespace std {
  using ::log;
  using ::sqrt;
}
#endif

namespace boost {

#if defined(__GNUC__) && (__GNUC__ < 3)
// Special gcc workaround: gcc 2.95.x ignores using-declarations
// in template classes (confirmed by gcc author Martin v. Loewis)
  using std::sqrt;
  using std::exp;
#endif

/**
 * Instantiations of class template lognormal_distribution model a
 * \random_distribution. Such a distribution produces random numbers
 * with \f$p(x) = \frac{1}{x \sigma_N \sqrt{2\pi}} e^{\frac{-\left(\log(x)-\mu_N\right)^2}{2\sigma_N^2}}\f$
 * for x > 0, where \f$\mu_N = \log\left(\frac{\mu^2}{\sqrt{\sigma^2 + \mu^2}}\right)\f$ and
 * \f$\sigma_N = \sqrt{\log\left(1 + \frac{\sigma^2}{\mu^2}\right)}\f$.
 */
template<class RealType = double>
class lognormal_distribution
{
public:
  typedef typename normal_distribution<RealType>::input_type input_type;
  typedef RealType result_type;

#ifndef BOOST_NO_LIMITS_COMPILE_TIME_CONSTANTS
    BOOST_STATIC_ASSERT(!std::numeric_limits<RealType>::is_integer);
#endif

  /**
   * Constructs a lognormal_distribution. @c mean and @c sigma are the
   * mean and standard deviation of the lognormal distribution.
   */
  explicit lognormal_distribution(result_type mean_arg = result_type(1),
                                  result_type sigma_arg = result_type(1))
    : _mean(mean_arg), _sigma(sigma_arg)
  { 
    assert(_mean > result_type(0));
    init();
  }

  // compiler-generated copy ctor and assignment operator are fine

  RealType mean() const { return _mean; }
  RealType sigma() const { return _sigma; }
  void reset() { _normal.reset(); }

  template<class Engine>
  result_type operator()(Engine& eng)
  {
#ifndef BOOST_NO_STDC_NAMESPACE
    // allow for Koenig lookup
    using std::exp;
#endif
    return exp(_normal(eng) * _nsigma + _nmean);
  }

#ifndef BOOST_RANDOM_NO_STREAM_OPERATORS
  template<class CharT, class Traits>
  friend std::basic_ostream<CharT,Traits>&
  operator<<(std::basic_ostream<CharT,Traits>& os, const lognormal_distribution& ld)
  {
    os << ld._normal << " " << ld._mean << " " << ld._sigma;
    return os;
  }

  template<class CharT, class Traits>
  friend std::basic_istream<CharT,Traits>&
  operator>>(std::basic_istream<CharT,Traits>& is, lognormal_distribution& ld)
  {
    is >> std::ws >> ld._normal >> std::ws >> ld._mean >> std::ws >> ld._sigma;
    ld.init();
    return is;
  }
#endif

private:

  /// \cond hide_private_members
  void init()
  {
#ifndef BOOST_NO_STDC_NAMESPACE
    // allow for Koenig lookup
    using std::exp; using std::log; using std::sqrt;
#endif
    _nmean = log(_mean*_mean/sqrt(_sigma*_sigma + _mean*_mean));
    _nsigma = sqrt(log(_sigma*_sigma/_mean/_mean+result_type(1)));
  }
  /// \endcond

  RealType _mean, _sigma;
  RealType _nmean, _nsigma;
  normal_distribution<result_type> _normal;
};

} // namespace boost

#endif // BOOST_RANDOM_LOGNORMAL_DISTRIBUTION_HPP