DistributionsInterfaceBase Class Reference

Rcpp interface that serves as the parent class for Rcpp distribution interfaces. This type should be inherited and not called from R directly. More...

#include <rcpp_distribution.hpp>

Inheritance diagram for DistributionsInterfaceBase:

Public Member Functions
	DistributionsInterfaceBase ()
	The constructor.
	DistributionsInterfaceBase (const DistributionsInterfaceBase &other)
	Construct a new Distributions Interface Base object.
virtual	~DistributionsInterfaceBase ()
	The destructor.
virtual uint32_t	get_id ()=0
	Get the ID for the child distribution interface objects to inherit.
virtual bool	set_distribution_links (std::string input_type, Rcpp::IntegerVector ids)
	Sets pointers for data observations, random effects, or priors.
virtual bool	set_distribution_mean (double input_value)
	Set the expected mean value for the distribution.
virtual bool	set_observed_data (int observed_data_id)
	Set the unique ID for the observed data object.
virtual double	evaluate ()=0
	A method for each child distribution interface object to inherit so each distribution can have an evaluate() function.
std::string	parameters_to_json (const std::vector< std::pair< std::string, VariableVector * > > &parameters)
	Writes the parameters of a distribution as a JSON "parameters" member, in the same shape the modules use, so each element keeps its parameter id in the model output.
std::string	linked_ids_to_json ()
	Writes the ids of the vectors this distribution is linked to as a JSON "linked_ids" member.
Public Member Functions inherited from FIMSRcppInterfaceBase
virtual bool	add_to_fims_tmb ()
	A virtual method to inherit to add objects to the TMB model.
virtual void	finalize ()
	Extracts derived quantities back to the Rcpp interface object from the Information object.
virtual std::string	to_json ()
	Convert the data to json representation for the output.
std::string	value_to_string (double value)
	Report the variable value as a string.
std::string	make_dimensions (uint32_t start, uint32_t end, uint32_t rep=1)
	Make a string of dimensions for the model.

Public Attributes
uint32_t	id_m
	The local ID of the DistributionsInterfaceBase object.
std::shared_ptr< std::vector< uint32_t > >	key_m
	The unique ID for the variable map that points to a fims::Vector.
SharedString	input_type_m
	The type of density input. The options are prior, re, or data.
SharedString	use_mean_m = fims::to_string("no")
	Control flag indicating whether to use the expected mean in the distribution calculations.
SharedInt	interface_observed_data_id_m = -999
	The ID of the observed data object, which is set to -999.
double	lpdf_value = 0
	The log probability density function value.
Public Attributes inherited from FIMSRcppInterfaceBase
bool	finalized = false
	Is the object already finalized? The default is false.

Static Public Attributes
static uint32_t	id_g = 1
	The static ID of the DistributionsInterfaceBase object.
static std::map< uint32_t, std::shared_ptr< DistributionsInterfaceBase > >	live_objects
	The map associating the ID of the DistributionsInterfaceBase to the DistributionsInterfaceBase objects. This is a live object, which is an object that has been created and lives in memory.
Static Public Attributes inherited from FIMSRcppInterfaceBase
static std::vector< std::shared_ptr< FIMSRcppInterfaceBase > >	fims_interface_objects
	FIMS interface object vectors.

Detailed Description

Rcpp interface that serves as the parent class for Rcpp distribution interfaces. This type should be inherited and not called from R directly.

Constructor & Destructor Documentation

DistributionsInterfaceBase()

DistributionsInterfaceBase::DistributionsInterfaceBase ( const DistributionsInterfaceBase &  other )

inline

Construct a new Distributions Interface Base object.

Parameters

other

Member Function Documentation

evaluate()

virtual double DistributionsInterfaceBase::evaluate ( )

pure virtual

A method for each child distribution interface object to inherit so each distribution can have an evaluate() function.

Implemented in DnormDistributionsInterface, DlnormDistributionsInterface, and DmultinomDistributionsInterface.

get_id()

virtual uint32_t DistributionsInterfaceBase::get_id ( )

pure virtual

Get the ID for the child distribution interface objects to inherit.

Implemented in DnormDistributionsInterface, DlnormDistributionsInterface, and DmultinomDistributionsInterface.

linked_ids_to_json()

std::string DistributionsInterfaceBase::linked_ids_to_json ( )

inline

Writes the ids of the vectors this distribution is linked to as a JSON "linked_ids" member.

The first id is the vector the distribution applies to, e.g., recruitment deviations or a fleet's expected index, so its parameters can be reported under the module that owns that vector.

Returns

The JSON member followed by a comma and a newline.

parameters_to_json()

std::string DistributionsInterfaceBase::parameters_to_json ( const std::vector< std::pair< std::string, VariableVector * > > &  parameters )

inline

Writes the parameters of a distribution as a JSON "parameters" member, in the same shape the modules use, so each element keeps its parameter id in the model output.

Each distribution passes the parameter vectors it owns, e.g., {{"log_sd", &this->log_sd}}, so any process that uses the distribution reports its parameters, and a distribution with other parameters only needs to add them to its list. Empty vectors are skipped because the VariableVector stream operator does not handle a size of zero.

Parameters

parameters

Pairs of parameter names and pointers to the vectors.

Returns

The JSON member followed by a comma and a newline, or an empty string when no vector has elements.

set_distribution_links()

virtual bool DistributionsInterfaceBase::set_distribution_links ( std::string  input_type, Rcpp::IntegerVector  ids  )

inlinevirtual

Sets pointers for data observations, random effects, or priors.

Parameters

input_type	String that sets whether the distribution type is for priors, random effects, or data.
ids	Vector of unique ids for each linked parameter(s), derived value(s), or observed data vector.

Reimplemented in DnormDistributionsInterface, DlnormDistributionsInterface, and DmultinomDistributionsInterface.

set_distribution_mean()

virtual bool DistributionsInterfaceBase::set_distribution_mean ( double  input_value )

inlinevirtual

Set the expected mean value for the distribution.

This virtual function provides an interface for setting a fixed mean value for distribution objects. When overridden in derived classes, this method typically stores the provided mean value as a fixed effect parameter and marks the distribution to use the mean in its calculations.

The base class implementation returns false to indicate the operation is not supported. Derived classes that support mean specification should override this method to implement the actual functionality.

Parameters

input_value

The numeric value to set as the distribution's expected mean. This value will be treated as a fixed effect parameter (not estimated) in derived class implementations.

Returns

bool Returns true if the mean was successfully set, false otherwise. The base class implementation always returns false to indicate the operation is not supported by default.

See also

DnormDistributionsInterface::set_distribution_mean for an example implementation that sets the mean as a fixed effect parameter.

Reimplemented in DnormDistributionsInterface.

set_observed_data()

virtual bool DistributionsInterfaceBase::set_observed_data ( int  observed_data_id )

inlinevirtual

Set the unique ID for the observed data object.

Parameters

observed_data_id

Unique ID for the Observed Age Comp Data object

Reimplemented in DnormDistributionsInterface, DlnormDistributionsInterface, and DmultinomDistributionsInterface.

Member Data Documentation

use_mean_m

SharedString DistributionsInterfaceBase::use_mean_m = fims::to_string("no")

Control flag indicating whether to use the expected mean in the distribution calculations.

This shared string member serves as a boolean flag (i.e., "yes" or "no") that determines whether the distribution should use the expected_mean vector or other expected values (e.g., from data or random effects) when computing the expected value in the likelihood calculations.

When set to "no" (default), the distribution uses expected values based on the input_type setting (data expected values for "data", random effects expected values for "random_effects", or standard expected values otherwise).

When set to "yes" (typically by calling set_distribution_mean()), the distribution overrides the default expected value source and uses the expected_mean vector instead. This is useful for setting a fixed mean value for the distribution that doesn't depend on other model components.

See also

set_distribution_mean() for the method that sets this flag to "yes".

DensityComponentBase::get_expected() in density_components_base.hpp for the implementation that checks this flag.

---

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

• inst/include/interface/rcpp/rcpp_objects/rcpp_distribution.hpp
• src/rcpp_distribution.cpp