sam::Experiment¶

Experiment encapsulates data and methods needed by the Researcher to conduct its research. More...

#include <Experiment.h>

Public Functions¶

	Name
DependentVariable &	operator[](std::size_t idx)
const DependentVariable &	operator[](std::size_t idx) const
auto	begin()
auto	end()
auto	begin() const
auto	end() const
	Experiment() =default Default constructor.
	Experiment(json & experiment_config) Constructs an Experiment object using the given JSON configuration.
	Experiment(ExperimentSetup & e) Constructs an Experiment using an already initialized ExperimentSetup.
	Experiment(ExperimentSetup & e, std::shared_ptr< DataStrategy > & ds, std::shared_ptr< TestStrategy > & ts, std::shared_ptr< EffectStrategy > & es) Directly constructs an Experiment from its components.
void	addNewCondition()
void	addNewCondition(const std::vector< DependentVariable > & dvs)
void	addNewDependentVariable(const DependentVariable & dv) Adds a new dependent variable to the list.
void	addNewCandidates(const std::vector< Submission > & subs) Adds new submissions to the list of submissions.
void	setHackedStatus(const bool status) Sets the hacked status of the experiment.
void	setPublishedStatus(const bool status) Sets the published status of the experiment.
void	setHackedStatusOf(const std::vector< size_t > & idxs, const bool status) Sets the hacked status of a group of dvs.
void	setCandidateStatusOf(const std::vector< size_t > & idxs, const bool status) Sets the candidate status of a group of dvs.
bool	isHacked() const Returns true if the experiment is hacked.
bool	isModified() const Returns true if the experiment has been modified in anyway.
bool	hasCandidates() const Returns true if there is an candidate in the experiment.
bool	hasCovariants() const Returns true if there is an candidate in the experiment.
bool	isPublished() const Returns true if the experiment has been published by the Journal.
size_t	nCandidates() const Returns the number of candidate DVs.
int	nCovariants() const
void	reset() Clears and re-initializes the dependent variables.
void	generateData() Uses the DataStrategy to populate every DVs with raw the data.
void	generateCovariants() Generates covariants data.
void	calculateStatistics() Asks each DependentVariable to update its general statistics, e.g., mean, var.
void	calculateTests() Uses the TestStrategy to run the statistical test.
void	calculateEffects() Uses the EffectStrategy to calculates the effect sizes.
void	recalculateEverything() Runs calculateStatistics(), calculateTests(), and calculateEffects() in order.
void	clear() Clears the content of the experiment.
void	setTestStrategy(std::shared_ptr< TestStrategy > & ts)
void	setDataStrategy(std::shared_ptr< DataStrategy > & ds)
void	setEffectSizeEstimator(std::shared_ptr< EffectStrategy > & es)

Public Attributes¶

	Name
int	simid
int	exprid
int	repid
int	n_covariants Indicates the number of covariants.
arma::Mat< int >	covariants
ExperimentSetup	setup
std::shared_ptr< DataStrategy >	data_strategy
std::shared_ptr< TestStrategy >	test_strategy
std::shared_ptr< EffectStrategy >	effect_strategy
std::vector< DependentVariable >	dvs_
std::optional< std::vector< Submission > >	candidates List of all possible candidates from this experiment so far!

Detailed Description¶

class sam::Experiment;

Experiment encapsulates data and methods needed by the Researcher to conduct its research.

Experiment has a full access to all data-related strategies, e.g., DataStrategy, TestStrategy, and EffectStrategy. It also contains the raw research data in the form of list of DependentVariable(s). Moreover, the Experiment stores a copy of the ExperimentSetup where most study parameters are stored in.

Public Functions Documentation¶

function operator[]¶

DependentVariable & operator[](
    std::size_t idx
)

Todo: I'm not the fan of this find_if here, and I think I can do better.

I think I can avoid this search if I use std::reference_wrapper

These operators are returning the correct group, even if the group list is not sorted.

function operator[]¶

const DependentVariable & operator[](
    std::size_t idx
) const

function begin¶

inline auto begin()

function end¶

inline auto end()

function begin¶

inline auto begin() const

function end¶

inline auto end() const

function Experiment¶

Experiment() =default

Default constructor.

function Experiment¶

Experiment(
    json & experiment_config
)

Constructs an Experiment object using the given JSON configuration.

This constructs an Experiment, and allocates the resources necessary for the experiment to setup its dependent variables, etc. Using the config parameters of the Data, Test, and Effect strategies, it also initializes those to be used by the Researcher or the Experiment itself

function Experiment¶

Experiment(
    ExperimentSetup & e
)

Constructs an Experiment using an already initialized ExperimentSetup.

Note: This can be used in cases where the underlying strategies should be preserved while some experiment parameters needs to be modified.

Since ExperimentSetup already contains the definition of Data, Test, and Effect strategies, this method accepts the ExperimentSetup as it is, and only initialize those strategies.

function Experiment¶

Experiment(
    ExperimentSetup & e,
    std::shared_ptr< DataStrategy > & ds,
    std::shared_ptr< TestStrategy > & ts,
    std::shared_ptr< EffectStrategy > & es
)

Directly constructs an Experiment from its components.

Note: This is mainly used by the ExperimentBuilder

This directly constructs the experiment by directly constructing its internal based on the set of given parameters.

function addNewCondition¶

void addNewCondition()

Adds a new condition @TODO TO BE IMPLEMENTED!

function addNewCondition¶

void addNewCondition(
    const std::vector< DependentVariable > & dvs
)

Adds a new condition @TODO TO BE IMPLEMENTED!

function addNewDependentVariable¶

void addNewDependentVariable(
    const DependentVariable & dv
)

Adds a new dependent variable to the list.

function addNewCandidates¶

void addNewCandidates(
    const std::vector< Submission > & subs
)

Adds new submissions to the list of submissions.

Adds new candidates to the list of selected candidates

function setHackedStatus¶

void setHackedStatus(
    const bool status
)

Sets the hacked status of the experiment.

Attention: Setting this to true will make the [isHacked()](/doxygen/Classes/classsam_1_1_experiment/#function-ishacked) to return true. Basically, this overrules the status of the dependent variables, but it doesn't overwrite them!

It sets the overall hacked status of the experiment to status.

function setPublishedStatus¶

void setPublishedStatus(
    const bool status
)

Sets the published status of the experiment.

function setHackedStatusOf¶

void setHackedStatusOf(
    const std::vector< size_t > & idxs,
    const bool status
)

Sets the hacked status of a group of dvs.

function setCandidateStatusOf¶

void setCandidateStatusOf(
    const std::vector< size_t > & idxs,
    const bool status
)

Sets the candidate status of a group of dvs.

function isHacked¶

bool isHacked() const

Returns true if the experiment is hacked.

An experiment is hacked if its hacked status has been set to true, or one of its dependent variables has been flagged as hacked

function isModified¶

bool isModified() const

Returns true if the experiment has been modified in anyway.

function hasCandidates¶

bool hasCandidates() const

Returns true if there is an candidate in the experiment.

function hasCovariants¶

bool hasCovariants() const

Returns true if there is an candidate in the experiment.

function isPublished¶

bool isPublished() const

Returns true if the experiment has been published by the Journal.

function nCandidates¶

size_t nCandidates() const

Returns the number of candidate DVs.

function nCovariants¶

int nCovariants() const

function reset¶

void reset()

Clears and re-initializes the dependent variables.

This cleanup the Experiment, and reallocate its memory again. Technically, preparing the experiment for a new run.

function generateData¶

void generateData()

Uses the DataStrategy to populate every DVs with raw the data.

function generateCovariants¶

void generateCovariants()

Generates covariants data.

@Todo Check if all groups are the same size

function calculateStatistics¶

void calculateStatistics()

Asks each DependentVariable to update its general statistics, e.g., mean, var.

function calculateTests¶

void calculateTests()

Uses the TestStrategy to run the statistical test.

function calculateEffects¶

void calculateEffects()

Uses the EffectStrategy to calculates the effect sizes.

function recalculateEverything¶

void recalculateEverything()

Runs calculateStatistics(), calculateTests(), and calculateEffects() in order.

function clear¶

void clear()

Clears the content of the experiment.

Todo: I think this is a bad implementation, and I should probably just discard the list of DVs and recreate them for the new Experiment, which is probably safer! Something like dvs_.clear()

It clears every DVs individually, and also sort them back into the correct order.

function setTestStrategy¶

inline void setTestStrategy(
    std::shared_ptr< TestStrategy > & ts
)

Parameters:

t A reference to a Test Strategy instance

Set or re-set the Test Strategy

function setDataStrategy¶

inline void setDataStrategy(
    std::shared_ptr< DataStrategy > & ds
)

Parameters:

d A reference to a Data Strategy instance

Set or re-set the Data Strategy

function setEffectSizeEstimator¶

inline void setEffectSizeEstimator(
    std::shared_ptr< EffectStrategy > & es
)

Parameters:

es A reference to an Effect Strategy instance.

Set or re-set the Effect Strategy

Public Attributes Documentation¶

variable simid¶

int simid {0};

variable exprid¶

int exprid {0};

variable repid¶

int repid {0};

variable n_covariants¶

int n_covariants {0};

Indicates the number of covariants.

variable covariants¶

arma::Mat< int > covariants;

variable setup¶

ExperimentSetup setup;

An instance of the ExperimentSetup. All other strategies can access and query it for meta information about the Experiment

variable data_strategy¶

std::shared_ptr< DataStrategy > data_strategy;

variable test_strategy¶

std::shared_ptr< TestStrategy > test_strategy;

variable effect_strategy¶

std::shared_ptr< EffectStrategy > effect_strategy;

variable dvs_¶

std::vector< DependentVariable > dvs_;

variable candidates¶

std::optional< std::vector< Submission > > candidates;

List of all possible candidates from this experiment so far!

Note: This is not yet in use!

This is a list of any dvs that has been flagged as submissions during the lifespan of this experiment

Updated on 29 June 2021 at 16:13:47 CEST

Last update: 2021-09-18