DistributionLogNormal Class Reference

Description

Log-normal distribution.

Definition at line 217 of file Distributions.h.

Inheritance diagram for DistributionLogNormal:

Collaboration diagram for DistributionLogNormal:

Public Member Functions
	DistributionLogNormal (double median, double scale_param)
	DistributionLogNormal (std::vector< double > P)
void	checkNodeArgs () const
	Raises exception if a parameter value is invalid. More...
std::string	className () const final
	Returns the class name, to be hard-coded in each leaf class that inherits from INode. More...
DistributionLogNormal *	clone () const override
std::vector< double >	equidistantPoints (size_t nbr_samples, double sigma_factor, const RealLimits &limits=RealLimits()) const override
	generate list of sample values More...
virtual std::vector< double >	equidistantPointsInRange (size_t nbr_samples, double xmin, double xmax) const
	Returns equidistant interpolation points from xmin to xmax. More...
std::vector< ParameterSample >	equidistantSamples (size_t nbr_samples, double sigma_factor=0., const RealLimits &limits=RealLimits()) const
	Returns equidistant samples, using intrinsic parameters, weighted with probabilityDensity(). More...
std::vector< ParameterSample >	equidistantSamplesInRange (size_t nbr_samples, double xmin, double xmax) const
	Returns equidistant samples from xmin to xmax, weighted with probabilityDensity(). More...
double	getMedian () const
double	getScalePar () const
bool	isDelta () const override
	Returns true if the distribution is in the limit case of a Dirac delta distribution. More...
double	mean () const override
	Returns the distribution-specific mean. More...
virtual std::vector< const INode * >	nodeChildren () const
	Returns all children. More...
std::vector< const INode * >	nodeOffspring () const
	Returns all descendants. More...
std::vector< ParaMeta >	parDefs () const final
	Returns the parameter definitions, to be hard-coded in each leaf class. More...
double	probabilityDensity (double x) const override
	Returns the distribution-specific probability density for value x. More...
std::string	pythonConstructor (const std::string &units) const override
	Prints distribution with constructor parameters in given units. ba.DistributionGaussian(2.0*deg, 0.02*deg) More...
virtual void	transferToCPP ()
	Used for Python overriding of clone (see swig/tweaks.py) More...

Protected Member Functions
void	adjustMinMaxForLimits (double &xmin, double &xmax, const RealLimits &limits) const
	modifies xmin and xmax if they are outside of limits More...
std::vector< ParameterSample >	generateSamplesFromValues (const std::vector< double > &sample_values) const
	Returns weighted samples from given interpolation points and probabilityDensity(). More...

Protected Attributes
std::vector< double >	m_P

Private Attributes
const double &	m_median
const double &	m_scale_param

Friends
class	DistributionsTest_DistributionLogNormalParameters_Test

Constructor & Destructor Documentation

DistributionLogNormal() [1/2]

DistributionLogNormal::DistributionLogNormal

(

std::vector< double >

P

)

Definition at line 277 of file Distributions.cpp.

    : IDistribution1D(P)
    , m_median(m_P[0])
    , m_scale_param(m_P[1])
{
    checkNodeArgs();
    if (m_scale_param < 0.0)
        throw std::runtime_error("DistributionLogNormal: scale_param < 0");
    if (m_median <= 0.0)
        throw std::runtime_error("DistributionLogNormal: median < 0");
}

References INode::checkNodeArgs(), m_median, and m_scale_param.

Referenced by clone().

Here is the call graph for this function:

DistributionLogNormal() [2/2]

DistributionLogNormal::DistributionLogNormal	(	double	median,
		double	scale_param
	)

Definition at line 289 of file Distributions.cpp.

    : DistributionLogNormal(std::vector<double>{median, scale_param})
{
}

Member Function Documentation

adjustMinMaxForLimits()

void IDistribution1D::adjustMinMaxForLimits ( double &  xmin, double &  xmax, const RealLimits &  limits  ) const

protectedinherited

modifies xmin and xmax if they are outside of limits

Definition at line 81 of file Distributions.cpp.

{
    if (limits.hasLowerLimit() && xmin < limits.lowerLimit())
        xmin = limits.lowerLimit();
    if (limits.hasUpperLimit() && xmax > limits.upperLimit())
        xmax = limits.upperLimit();
    if (xmin > xmax) {
        std::ostringstream ostr;
        ostr << "IDistribution1D::adjustMinMaxForLimits() -> Error. Can't' adjust ";
        ostr << "xmin:" << xmin << " xmax:" << xmax << " for given limits " << limits << std::endl;
        throw std::runtime_error(ostr.str());
    }
}

References RealLimits::hasLowerLimit(), RealLimits::hasUpperLimit(), RealLimits::lowerLimit(), and RealLimits::upperLimit().

Referenced by DistributionGate::equidistantPoints(), DistributionLorentz::equidistantPoints(), DistributionGaussian::equidistantPoints(), equidistantPoints(), DistributionCosine::equidistantPoints(), and DistributionTrapezoid::equidistantPoints().

Here is the call graph for this function:

checkNodeArgs()

void INode::checkNodeArgs ( ) const

inherited

Raises exception if a parameter value is invalid.

Definition at line 27 of file INode.cpp.

{
    size_t nP = m_P.size();
    if (parDefs().size() != nP) {
        std::cerr << "BUG in class " << className() << std::endl;
        std::cerr << "#m_P = " << nP << std::endl;
        std::cerr << "#PDf = " << parDefs().size() << std::endl;
        for (const ParaMeta& pm : parDefs())
            std::cerr << "  PDf: " << pm.name << std::endl;
        ASSERT(0);
    }
    ASSERT(parDefs().size() == nP);
    for (size_t i = 0; i < nP; ++i) {
        const ParaMeta pm = parDefs()[i];
 
        RealLimits limits = RealLimits::limitless();
        if (pm.vMin == -INF) {
            ASSERT(pm.vMax == +INF);
            // nothing to do
        } else if (pm.vMax == +INF) {
            ASSERT(pm.vMin == 0);
            limits = RealLimits::nonnegative();
        } else {
            limits = RealLimits::limited(pm.vMin, pm.vMax);
        }
        limits.check(pm.name, m_P[i]);
    }
}

References ASSERT, RealLimits::check(), INode::className(), INF, RealLimits::limited(), RealLimits::limitless(), INode::m_P, ParaMeta::name, RealLimits::nonnegative(), INode::parDefs(), ParaMeta::vMax, and ParaMeta::vMin.

Referenced by BarGauss::BarGauss(), BarLorentz::BarLorentz(), Bipyramid4::Bipyramid4(), Box::Box(), CantellatedCube::CantellatedCube(), Cone::Cone(), ConstantBackground::ConstantBackground(), CosineRippleBox::CosineRippleBox(), CosineRippleGauss::CosineRippleGauss(), CosineRippleLorentz::CosineRippleLorentz(), Cylinder::Cylinder(), DistributionCosine::DistributionCosine(), DistributionGate::DistributionGate(), DistributionGaussian::DistributionGaussian(), DistributionLogNormal(), DistributionLorentz::DistributionLorentz(), DistributionTrapezoid::DistributionTrapezoid(), Dodecahedron::Dodecahedron(), EllipsoidalCylinder::EllipsoidalCylinder(), FootprintGauss::FootprintGauss(), FootprintSquare::FootprintSquare(), FuzzySphere::FuzzySphere(), GaussSphere::GaussSphere(), HemiEllipsoid::HemiEllipsoid(), HollowSphere::HollowSphere(), HorizontalCylinder::HorizontalCylinder(), Icosahedron::Icosahedron(), LongBoxGauss::LongBoxGauss(), LongBoxLorentz::LongBoxLorentz(), PlatonicOctahedron::PlatonicOctahedron(), PlatonicTetrahedron::PlatonicTetrahedron(), Prism3::Prism3(), Prism6::Prism6(), Profile1DCauchy::Profile1DCauchy(), Profile1DCosine::Profile1DCosine(), Profile1DGate::Profile1DGate(), Profile1DGauss::Profile1DGauss(), Profile1DTriangle::Profile1DTriangle(), Profile1DVoigt::Profile1DVoigt(), Profile2DCauchy::Profile2DCauchy(), Profile2DCone::Profile2DCone(), Profile2DGate::Profile2DGate(), Profile2DGauss::Profile2DGauss(), Profile2DVoigt::Profile2DVoigt(), Pyramid2::Pyramid2(), Pyramid3::Pyramid3(), Pyramid4::Pyramid4(), Pyramid6::Pyramid6(), RotationEuler::RotationEuler(), RotationX::RotationX(), RotationY::RotationY(), RotationZ::RotationZ(), SawtoothRippleBox::SawtoothRippleBox(), SawtoothRippleGauss::SawtoothRippleGauss(), SawtoothRippleLorentz::SawtoothRippleLorentz(), Sphere::Sphere(), Spheroid::Spheroid(), TruncatedCube::TruncatedCube(), TruncatedSphere::TruncatedSphere(), and TruncatedSpheroid::TruncatedSpheroid().

Here is the call graph for this function:

className()

std::string DistributionLogNormal::className ( ) const

inlinefinalvirtual

Returns the class name, to be hard-coded in each leaf class that inherits from INode.

Implements INode.

Definition at line 226 of file Distributions.h.

{ return "DistributionLogNormal"; }

Referenced by pythonConstructor().

clone()

DistributionLogNormal* DistributionLogNormal::clone ( ) const

inlineoverridevirtual

Implements IDistribution1D.

Definition at line 222 of file Distributions.h.

    {
        return new DistributionLogNormal(m_median, m_scale_param);
    }

References DistributionLogNormal(), m_median, and m_scale_param.

Here is the call graph for this function:

equidistantPoints()

std::vector< double > DistributionLogNormal::equidistantPoints ( size_t  nbr_samples, double  sigma_factor, const RealLimits &  limits = RealLimits()  ) const

overridevirtual

generate list of sample values

Implements IDistribution1D.

Definition at line 308 of file Distributions.cpp.

{
    if (nbr_samples < 2) {
        std::vector<double> result;
        result.push_back(m_median);
        return result;
    }
    if (sigma_factor <= 0.0)
        sigma_factor = 2.0;
    double xmin = m_median * std::exp(-sigma_factor * m_scale_param);
    double xmax = m_median * std::exp(sigma_factor * m_scale_param);
    adjustMinMaxForLimits(xmin, xmax, limits);
    return equidistantPointsInRange(nbr_samples, xmin, xmax);
}

References IDistribution1D::adjustMinMaxForLimits(), IDistribution1D::equidistantPointsInRange(), m_median, and m_scale_param.

Here is the call graph for this function:

equidistantPointsInRange()

std::vector< double > IDistribution1D::equidistantPointsInRange ( size_t  nbr_samples, double  xmin, double  xmax  ) const

virtualinherited

Returns equidistant interpolation points from xmin to xmax.

Definition at line 70 of file Distributions.cpp.

{
    if (nbr_samples < 2 || DoubleEqual(xmin, xmax))
        return {mean()};
    std::vector<double> result(nbr_samples);
    for (size_t i = 0; i < nbr_samples; ++i)
        result[i] = xmin + i * (xmax - xmin) / (nbr_samples - 1.0);
    return result;
}

References IDistribution1D::mean().

Referenced by DistributionGate::equidistantPoints(), DistributionLorentz::equidistantPoints(), DistributionGaussian::equidistantPoints(), equidistantPoints(), DistributionCosine::equidistantPoints(), DistributionTrapezoid::equidistantPoints(), and IDistribution1D::equidistantSamplesInRange().

Here is the call graph for this function:

equidistantSamples()

std::vector< ParameterSample > IDistribution1D::equidistantSamples ( size_t  nbr_samples, double  sigma_factor = 0., const RealLimits &  limits = RealLimits()  ) const

inherited

Returns equidistant samples, using intrinsic parameters, weighted with probabilityDensity().

Definition at line 43 of file Distributions.cpp.

{
    if (nbr_samples == 0)
        throw std::runtime_error("IDistribution1D::generateSamples: "
                                 "number of generated samples must be bigger than zero");
    if (isDelta())
        return {ParameterSample(mean())};
    return generateSamplesFromValues(equidistantPoints(nbr_samples, sigma_factor, limits));
}

References IDistribution1D::equidistantPoints(), IDistribution1D::generateSamplesFromValues(), IDistribution1D::isDelta(), and IDistribution1D::mean().

Here is the call graph for this function:

equidistantSamplesInRange()

std::vector< ParameterSample > IDistribution1D::equidistantSamplesInRange ( size_t  nbr_samples, double  xmin, double  xmax  ) const

inherited

Returns equidistant samples from xmin to xmax, weighted with probabilityDensity().

Definition at line 58 of file Distributions.cpp.

{
    if (nbr_samples == 0)
        throw std::runtime_error("IDistribution1D::generateSamples: "
                                 "number of generated samples must be bigger than zero");
    if (isDelta())
        return {ParameterSample(mean())};
    return generateSamplesFromValues(equidistantPointsInRange(nbr_samples, xmin, xmax));
}

References IDistribution1D::equidistantPointsInRange(), IDistribution1D::generateSamplesFromValues(), IDistribution1D::isDelta(), and IDistribution1D::mean().

Here is the call graph for this function:

generateSamplesFromValues()

std::vector< ParameterSample > IDistribution1D::generateSamplesFromValues ( const std::vector< double > &  sample_values ) const

protectedinherited

Returns weighted samples from given interpolation points and probabilityDensity().

Definition at line 99 of file Distributions.cpp.

{
    std::vector<ParameterSample> result;
    double norm_factor = 0.0;
    for (double value : sample_values) {
        double pdf = probabilityDensity(value);
        result.emplace_back(value, pdf);
        norm_factor += pdf;
    }
    if (norm_factor <= 0.0)
        throw std::runtime_error("IDistribution1D::generateSamples: "
                                 "total probability must be bigger than zero");
    for (ParameterSample& sample : result)
        sample.weight /= norm_factor;
    return result;
}

References IDistribution1D::probabilityDensity().

Referenced by IDistribution1D::equidistantSamples(), and IDistribution1D::equidistantSamplesInRange().

Here is the call graph for this function:

getMedian()

double DistributionLogNormal::getMedian ( ) const

inline

Definition at line 236 of file Distributions.h.

{ return m_median; }

References m_median.

getScalePar()

double DistributionLogNormal::getScalePar ( ) const

inline

Definition at line 237 of file Distributions.h.

{ return m_scale_param; }

References m_scale_param.

isDelta()

bool DistributionLogNormal::isDelta ( ) const

overridevirtual

Returns true if the distribution is in the limit case of a Dirac delta distribution.

Implements IDistribution1D.

Definition at line 325 of file Distributions.cpp.

{
    return m_scale_param == 0.0;
}

References m_scale_param.

mean()

double DistributionLogNormal::mean ( ) const

overridevirtual

Returns the distribution-specific mean.

Implements IDistribution1D.

Definition at line 302 of file Distributions.cpp.

{
    double exponent = m_scale_param * m_scale_param / 2.0;
    return m_median * std::exp(exponent);
}

References m_median, and m_scale_param.

nodeChildren()

std::vector< const INode * > INode::nodeChildren ( ) const

virtualinherited

Returns all children.

Reimplemented in ISimulation2D, ISimulation, ParticleCoreShell, ParticleComposition, Particle, MesoCrystal, IParticle, Crystal, MultiLayer, Layer, LayerInterface, ParticleLayout, InterferenceRadialParaCrystal, InterferenceFinite3DLattice, InterferenceFinite2DLattice, Interference3DLattice, Interference2DSuperLattice, Interference2DParaCrystal, Interference2DLattice, Interference1DLattice, ConvolutionDetectorResolution, IDetector, and Beam.

Definition at line 56 of file INode.cpp.

{
    return {};
}

Referenced by NodeUtils::AllDescendantsOfType(), NodeUtils::ChildNodesOfType(), ISampleNode::containedMaterials(), and INode::nodeOffspring().

nodeOffspring()

std::vector< const INode * > INode::nodeOffspring ( ) const

inherited

Returns all descendants.

Definition at line 61 of file INode.cpp.

{
    std::vector<const INode*> result;
    result.push_back(this);
    for (const auto* child : nodeChildren()) {
        for (const auto* p : child->nodeOffspring())
            result.push_back(p);
    }
    return result;
}

References INode::nodeChildren().

Here is the call graph for this function:

parDefs()

std::vector<ParaMeta> DistributionLogNormal::parDefs ( ) const

inlinefinalvirtual

Returns the parameter definitions, to be hard-coded in each leaf class.

Reimplemented from INode.

Definition at line 228 of file Distributions.h.

    {
        return {{"Median", "", "para_tooltip", -INF, +INF, 0},
                {"ScaleParameter", "", "para_tooltip", -INF, +INF, 0}};
    }

References INF.

probabilityDensity()

double DistributionLogNormal::probabilityDensity ( double  x ) const

overridevirtual

Returns the distribution-specific probability density for value x.

Implements IDistribution1D.

Definition at line 294 of file Distributions.cpp.

{
    if (m_scale_param == 0.0)
        return DoubleEqual(x, m_median) ? 1.0 : 0.0;
    double t = std::log(x / m_median) / m_scale_param;
    return std::exp(-t * t / 2.0) / (x * m_scale_param * std::sqrt(M_TWOPI));
}

References m_median, m_scale_param, and M_TWOPI.

pythonConstructor()

std::string DistributionLogNormal::pythonConstructor ( const std::string &  units ) const

overridevirtual

Prints distribution with constructor parameters in given units. ba.DistributionGaussian(2.0*deg, 0.02*deg)

Implements IDistribution1D.

Definition at line 330 of file Distributions.cpp.

{
    // scale parameter remains unitless
    return Py::Fmt::printFunction(className(), m_median, units, m_scale_param, "");
}

References className(), m_median, m_scale_param, and Py::Fmt::printFunction().

Here is the call graph for this function:

transferToCPP()

virtual void ICloneable::transferToCPP ( )

inlinevirtualinherited

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

Definition at line 32 of file ICloneable.h.

Friends And Related Function Documentation

DistributionsTest_DistributionLogNormalParameters_Test

friend class DistributionsTest_DistributionLogNormalParameters_Test

friend

Definition at line 253 of file Distributions.h.

Member Data Documentation

m_median

const double& DistributionLogNormal::m_median

private

Definition at line 250 of file Distributions.h.

Referenced by DistributionLogNormal(), clone(), equidistantPoints(), getMedian(), mean(), probabilityDensity(), and pythonConstructor().

m_P

std::vector<double> INode::m_P

protectedinherited

Definition at line 63 of file INode.h.

Referenced by IFootprintFactor::IFootprintFactor(), INode::checkNodeArgs(), IProfile1D::pythonConstructor(), IProfile2D::pythonConstructor(), IFormFactor::pythonConstructor(), Profile1DVoigt::pythonConstructor(), and Profile2DVoigt::pythonConstructor().

m_scale_param

const double& DistributionLogNormal::m_scale_param

private

Definition at line 251 of file Distributions.h.

Referenced by DistributionLogNormal(), clone(), equidistantPoints(), getScalePar(), isDelta(), mean(), probabilityDensity(), and pythonConstructor().

---

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

• /home/build/builds/o5h8MZZm/0/mlz/bornagain/Param/Distrib/Distributions.h
• /home/build/builds/o5h8MZZm/0/mlz/bornagain/Param/Distrib/Distributions.cpp