This class converts raw simulation data into multiple data rows, with piracy/policy, SR/LR as treatment effects on the base (pre-piracy, pre-policy) row. More...

#include <creativity/data/Treatment.hpp>

Public Member Functions
	Treatment ()=default
	Default constructor.

	Treatment (CSVParser &&csv)
	Takes over the given CSVParser and reads the data from it.

	Treatment (const std::string &filename)
	Creates a new CSVParser to read the given filename.

void	readCSV (CSVParser &&csv)
	Reads the data from the given CSVParser.

void	requirePiracy (bool require=true)
	Requires that the parsed data contain piracy observations. More...

void	requirePolicy (bool require=true)
	Like requirePiracy(), but for policy data.

void	requirePre (bool require=true)
	Like requirePiracy(), but for pre-piracy data.

void	requireSR (bool require=true)
	Requires short-run data for long-run piracy/policy data that exists in the source data. More...

const bool &	hasPre () const
	True if this treatment data pre-piracy observations.

const bool &	hasPiracy () const
	True if the source data has piracy data.

const bool &	hasPolicy () const
	True if the source data has policy data.

const bool &	hasPiracySR () const
	True if the source data has short-run piracy data.

const bool &	hasPolicySR () const
	True if the source data has short-run policy data.

bool	hasShortrun () const
	True if the source data has short-run data for each type of associated long-run data. More...

const unsigned int &	rowsPerSimulation () const
	The number of data rows per simulation input. This equals 1 plus the number of treatments.

unsigned int	simulations () const
	The number of simulation files contributing to this data: this is simply an alias for `obj.data().rows() / obj.rowsPerSimulation()`

const Eigen::MatrixXd &	data () const
	The full set of data. More...

const std::vector< std::string > &	sourceFiles () const
	The simulation source values. More...

const std::string &	sourceFile (unsigned r) const
	Return the source file associated with row r. More...

const std::unordered_map< std::string, unsigned > &	columns () const
	Accesses the map of field name to column indices. More...

std::shared_ptr< const SimpleVariable >	variable (const std::string &name) const
	Returns a shared_ptr to a SimpleVariable containing a copy of the requested column. More...

std::shared_ptr< const SimpleVariable >	operator[] (const std::string &name) const
	Alias for variable()

Protected Member Functions
void	generateRow (const CSVParser &csv, Eigen::Ref< Eigen::RowVectorXd, 0, Eigen::InnerStride<>> newrow, const std::string &prefix) const
	Generates a row of data from the CSVParser at its current position. More...

Protected Attributes
Eigen::MatrixXd	data_
	The data matrix.

std::vector< std::string >	source_
	In-order source values for data.

bool	have_data_ {false}
	True once we've parsed a data file.

bool	has_pre_ {false}
	True if the data contains pre-piracy, non-treatment rows.

bool	has_piracy_ {false}
	True if the data contains LR piracy treatment rows.

bool	has_piracy_sr_ {false}
	True if the data contains SR piracy treatment rows.

bool	has_policy_ {false}
	True if the data contains LR policy treatment rows.

bool	has_policy_sr_ {false}
	True if the data contains SR policy treatment rows.

bool	require_pre_ {false}
	True if pre data is required.

bool	require_piracy_ {false}
	True if piracy data is required.

bool	require_policy_ {false}
	True if policy data is required.

bool	require_sr_ {false}
	True if short-run data is required for each type of long-run data.

unsigned int	rows_per_sim_ {0}
	The number of treatment row observations per source data rows (i.e. per simulation)

std::unordered_map< std::string, unsigned >	data_column_
	Field name to column map.

std::unordered_map< std::string, std::shared_ptr< SimpleVariable > >	var_cache_
	Cache of weak pointers to returned SimpleVariables.

Detailed Description

This class converts raw simulation data into multiple data rows, with piracy/policy, SR/LR as treatment effects on the base (pre-piracy, pre-policy) row.

It additionally supports filtering the resulting data–for instance, only using observations that still exhibit some writing during piracy.

Member Function Documentation

§ columns()

const std::unordered_map<std::string, unsigned>& creativity::data::Treatment::columns ( ) const

Accesses the map of field name to column indices.

Data fields (e.g. "books_written") are not prefixed with "pre.", "piracy.", etc.; parameters (e.g. "param.readers") are prefixed; there are also dummies "piracy", "policy", "SR", and "LR".

§ data()

const Eigen::MatrixXd& creativity::data::Treatment::data ( ) const

inline

The full set of data.

See also: TreatmentFilter for extracting conditional data.

§ generateRow()

void creativity::data::Treatment::generateRow	(	const CSVParser &	csv,
		Eigen::Ref< Eigen::RowVectorXd, 0, Eigen::InnerStride<>>	newrow,
		const std::string &	prefix
	)		const

protected

Generates a row of data from the CSVParser at its current position.

Parameters

csv	the CSVParser positioned at the desired row.
newrow	the matrix row in which values will be set
prefix	the prefix (such as "pre." or "policy.SR.") to add to non-param. data columns.

§ hasShortrun()

bool creativity::data::Treatment::hasShortrun ( ) const

inline

True if the source data has short-run data for each type of associated long-run data.

That is, if the data has piracy data, it must also have short-run piracy data; likewise for policy data.

§ requirePiracy()

void creativity::data::Treatment::requirePiracy ( bool require = true )

Requires that the parsed data contain piracy observations.

If the data has already been parsed, this throws an exception immediately if the data does not contain piracy; if not yet parsed, this causes the next readCSV() to throw an exception if piracy data is not present.

Parameters

require defaults to true of omitted; can be explicitly specified as false to cancel a previous requirePiracy() call.

§ requireSR()

void creativity::data::Treatment::requireSR ( bool require = true )

Requires short-run data for long-run piracy/policy data that exists in the source data.

In particular, if this option is enabled, the data must contain short-run observations for each category with equivalent long-run observations: that is, if there is long-run piracy data, there must also be short-run piracy data, and likewise for policy data. If the source data does not contain long-run piracy data, this option will not require short-run piracy data.

Like requirePiracy(), this throws immediately if data has already been parsed; if not, the exception will be raised if attempting to read data that doesn't contain the required short-run data.

§ sourceFile()

const std::string& creativity::data::Treatment::sourceFile ( unsigned r ) const

inline

Return the source file associated with row r.

This is simply an alias for sourceFiles().at(r / rowsPerSimulation()).

§ sourceFiles()

const std::vector<std::string>& creativity::data::Treatment::sourceFiles ( ) const

inline

The simulation source values.

Data in row x is from source file at index x / rowsPerSimulation().

§ variable()

std::shared_ptr<const SimpleVariable> creativity::data::Treatment::variable ( const std::string & name ) const

Returns a shared_ptr to a SimpleVariable containing a copy of the requested column.

The column naming values are the same as those accepted by column().

Values are internally cached–multiple calls to variable with the same value return the same SimpleVariable.

Exceptions

std::out_of_range if the requested variable does not exist.

Inheritance diagram for creativity::data::Treatment:

[legend]

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

creativity/data/Treatment.hpp

Public Member Functions

Protected Member Functions

Protected Attributes

Detailed Description

Member Function Documentation

§ columns()

§ data()

§ generateRow()

§ hasShortrun()

§ requirePiracy()

§ requireSR()

§ sourceFile()

§ sourceFiles()

§ variable()