IRangedDistribution Class Reference

Interface for one-dimensional ranged distributions. More...

Inheritance diagram for IRangedDistribution:

Collaboration diagram for IRangedDistribution:

Public Member Functions
	IRangedDistribution ()
	IRangedDistribution (size_t n_samples, double sigma_factor, const RealLimits &limits=RealLimits::limitless())
	IRangedDistribution (size_t n_samples, double sigma_factor, double min, double max)
	Initializes Ranged distribution with given number of samples, sigma factor (range in standard deviations to take into account during sample generation) and limits (either RealLimits object or just min and max limits). More...
	~IRangedDistribution () override
IRangedDistribution *	clone () const override=0
std::unique_ptr< IDistribution1D >	distribution (double mean, double stddev) const
	Public interface function to underlying IDistribution1D object. More...
std::vector< std::vector< ParameterSample > >	generateSamples (const std::vector< double > &mean, const std::vector< double > &stddev) const
	Generates list of sampled values with their weights from given means and standard deviations. More...
std::vector< ParameterSample >	generateSamples (double mean, double stddev) const
RealLimits	limits () const
	Returns current limits of the distribution. More...
virtual std::string	name () const =0
	Returns distribution name for python-formatted text. More...
size_t	nSamples () const
	Returns number of samples to generate. More...
void	setLimits (const RealLimits &limits)
double	sigmaFactor () const
	Returns sigma factor to use during sampling. More...
virtual void	transferToCPP ()
	Used for Python overriding of clone (see swig/tweaks.py) More...

Protected Member Functions
virtual std::unique_ptr< IDistribution1D >	distribution_impl (double mean, double stddev) const =0
	Returns underlying IDistribution1D object. More...

Private Member Functions
void	checkInitialization ()

Private Attributes
RealLimits	m_limits
size_t	m_n_samples
double	m_sigma_factor

Detailed Description

Interface for one-dimensional ranged distributions.

All derived distributions allow for generating samples in-place for known mean and standard deviation (except for RangedDistributionLorentz which uses median and hwhm).

Definition at line 38 of file RangedDistributions.h.

Constructor & Destructor Documentation

IRangedDistribution() [1/3]

IRangedDistribution::IRangedDistribution

(

)

Definition at line 35 of file RangedDistributions.cpp.

    : m_n_samples(5), m_sigma_factor(2.0), m_limits(RealLimits::limitless())
{
    checkInitialization();
}

References checkInitialization().

Here is the call graph for this function:

IRangedDistribution() [2/3]

IRangedDistribution::IRangedDistribution	(	size_t	n_samples,
		double	sigma_factor,
		const RealLimits &	limits = RealLimits::limitless()
	)

Definition at line 41 of file RangedDistributions.cpp.

    : m_n_samples(n_samples), m_sigma_factor(sigma_factor), m_limits(limits)
{
    checkInitialization();
}

References checkInitialization().

Here is the call graph for this function:

IRangedDistribution() [3/3]

IRangedDistribution::IRangedDistribution	(	size_t	n_samples,
		double	sigma_factor,
		double	min,
		double	max
	)

Initializes Ranged distribution with given number of samples, sigma factor (range in standard deviations to take into account during sample generation) and limits (either RealLimits object or just min and max limits).

By default n_samples = 5, sigma_factor = 2.0, while the limits are (-inf, +inf).

Definition at line 48 of file RangedDistributions.cpp.

    : m_n_samples(n_samples), m_sigma_factor(sigma_factor), m_limits(RealLimits::limited(min, max))
{
    checkInitialization();
}

References checkInitialization().

Here is the call graph for this function:

~IRangedDistribution()

IRangedDistribution::~IRangedDistribution ( )

overridedefault

Member Function Documentation

checkInitialization()

void IRangedDistribution::checkInitialization ( )

private

Definition at line 95 of file RangedDistributions.cpp.

{
    if (m_n_samples < 1u)
        throw std::runtime_error("Error in IRangedDistribution::checkInitialization: number of "
                                 "samples shall be positive");
 
    if (m_sigma_factor < 0.0)
        throw std::runtime_error("Error in IRangedDistribution::checkInitialization: sigma factor "
                                 "shall be non-negative.");
 
    if (!m_limits.hasLowerAndUpperLimits())
        return;
 
    if (m_limits.lowerLimit() >= m_limits.upperLimit())
        throw std::runtime_error("Error in IRangedDistribution::checkInitialization: lower limit "
                                 "shall not exceed the upper one.");
}

References RealLimits::hasLowerAndUpperLimits(), RealLimits::lowerLimit(), m_limits, m_n_samples, m_sigma_factor, and RealLimits::upperLimit().

Referenced by IRangedDistribution().

Here is the call graph for this function:

clone()

IRangedDistribution* IRangedDistribution::clone ( ) const

overridepure virtual

Implements ICloneable.

Implemented in RangedDistributionCosine, RangedDistributionLogNormal, RangedDistributionGaussian, RangedDistributionLorentz, and RangedDistributionGate.

distribution()

std::unique_ptr< IDistribution1D > IRangedDistribution::distribution	(	double	mean,
		double	stddev
	)		const

Public interface function to underlying IDistribution1D object.

Definition at line 87 of file RangedDistributions.cpp.

{
    if (stddev < 0.0)
        throw std::runtime_error(
            "Error in IRangedDistribution::distribution: standard deviation is less than zero");
    return distribution_impl(mean, stddev);
}

References distribution_impl().

Referenced by generateSamples().

Here is the call graph for this function:

distribution_impl()

virtual std::unique_ptr<IDistribution1D> IRangedDistribution::distribution_impl ( double  mean, double  stddev  ) const

protectedpure virtual

Returns underlying IDistribution1D object.

Implemented in RangedDistributionLorentz, RangedDistributionCosine, RangedDistributionLogNormal, RangedDistributionGaussian, and RangedDistributionGate.

Referenced by distribution().

generateSamples() [1/2]

std::vector< std::vector< ParameterSample > > IRangedDistribution::generateSamples	(	const std::vector< double > &	mean,
		const std::vector< double > &	stddev
	)		const

Generates list of sampled values with their weights from given means and standard deviations.

Definition at line 71 of file RangedDistributions.cpp.

{
    if (mean.size() != stddev.size())
        throw std::runtime_error("Error in IRangedDistribution::generateSamples: mean and variance "
                                 "vectors shall be of the same size");
 
    const size_t size = mean.size();
 
    std::vector<std::vector<ParameterSample>> result;
    result.resize(size);
    for (size_t i = 0; i < size; ++i)
        result[i] = generateSamples(mean[i], stddev[i]);
    return result;
}

References generateSamples().

Here is the call graph for this function:

generateSamples() [2/2]

std::vector< ParameterSample > IRangedDistribution::generateSamples	(	double	mean,
		double	stddev
	)		const

Definition at line 57 of file RangedDistributions.cpp.

{
    auto generator = distribution(mean, stddev);
    if (!generator->isDelta())
        return generator->equidistantSamples(m_n_samples, m_sigma_factor, m_limits);
 
    // handling the case of delta distributions
    auto samples = generator->equidistantSamples(m_n_samples, m_sigma_factor, m_limits);
    ParameterSample& sample = samples[0]; // there is only one element in the vector
    sample.weight = 1.0 / m_n_samples;
    return std::vector<ParameterSample>(m_n_samples, sample);
}

References distribution(), m_limits, m_n_samples, m_sigma_factor, and ParameterSample::weight.

Referenced by generateSamples().

Here is the call graph for this function:

limits()

RealLimits IRangedDistribution::limits ( ) const

inline

Returns current limits of the distribution.

Definition at line 64 of file RangedDistributions.h.

{ return m_limits; }

References m_limits.

Referenced by pyfmt2::printRangedDistribution(), and setLimits().

name()

virtual std::string IRangedDistribution::name ( ) const

pure virtual

Returns distribution name for python-formatted text.

Implemented in RangedDistributionCosine, RangedDistributionLogNormal, RangedDistributionGaussian, RangedDistributionLorentz, and RangedDistributionGate.

Referenced by pyfmt2::printRangedDistribution().

nSamples()

size_t IRangedDistribution::nSamples ( ) const

inline

Returns number of samples to generate.

Definition at line 68 of file RangedDistributions.h.

{ return m_n_samples; }

References m_n_samples.

Referenced by pyfmt2::printRangedDistribution().

setLimits()

void IRangedDistribution::setLimits ( const RealLimits &  limits )

inline

Definition at line 72 of file RangedDistributions.h.

{ m_limits = limits; }

References limits(), and m_limits.

Here is the call graph for this function:

sigmaFactor()

double IRangedDistribution::sigmaFactor ( ) const

inline

Returns sigma factor to use during sampling.

Definition at line 66 of file RangedDistributions.h.

{ return m_sigma_factor; }

References m_sigma_factor.

Referenced by pyfmt2::printRangedDistribution().

transferToCPP()

virtual void ICloneable::transferToCPP ( )

inlinevirtualinherited

Used for Python overriding of clone (see swig/tweaks.py)

Definition at line 34 of file ICloneable.h.

Member Data Documentation

m_limits

RealLimits IRangedDistribution::m_limits

private

Definition at line 87 of file RangedDistributions.h.

Referenced by checkInitialization(), generateSamples(), limits(), and setLimits().

m_n_samples

size_t IRangedDistribution::m_n_samples

private

Definition at line 85 of file RangedDistributions.h.

Referenced by checkInitialization(), generateSamples(), and nSamples().

m_sigma_factor

double IRangedDistribution::m_sigma_factor

private

Definition at line 86 of file RangedDistributions.h.

Referenced by checkInitialization(), generateSamples(), and sigmaFactor().

---

The documentation for this class was generated from the following files:

• /home/www/ba/Param/Distrib/RangedDistributions.h
• /home/www/ba/Param/Distrib/RangedDistributions.cpp