Meta information about dataset, always sit in memory. More...

#include <data.h>

Collaboration diagram for xgboost::MetaInfo:

Public Member Functions
	MetaInfo ()=default
	default constructor More...

	MetaInfo (MetaInfo &&that)=default

MetaInfo &	operator= (MetaInfo &&that)=default

MetaInfo &	operator= (MetaInfo const &that)=delete

void	Validate (DeviceOrd device) const
	Validate all metainfo. More...

MetaInfo	Slice (common::Span< int32_t const > ridxs) const

MetaInfo	Copy () const

bst_float	GetWeight (size_t i) const
	Get weight of each instances. More...

const std::vector< size_t > &	LabelAbsSort (Context const *ctx) const
	get sorted indexes (argsort) of labels by absolute value (used by cox loss) More...

void	Clear ()
	clear all the information More...

void	LoadBinary (dmlc::Stream *fi)
	Load the Meta info from binary stream. More...

void	SaveBinary (dmlc::Stream *fo) const
	Save the Meta info to binary stream. More...

void	SetInfo (Context const &ctx, StringView key, StringView interface_str)
	Set information in the meta info with array interface. More...

void	GetInfo (char const key, bst_ulong out_len, DataType dtype, const void **out_dptr) const

void	SetFeatureInfo (const char key, const char *info, const bst_ulong size)

void	GetFeatureInfo (const char field, std::vector< std::string > out_str_vecs) const

void	Extend (MetaInfo const &that, bool accumulate_rows, bool check_column)
	Extend with other MetaInfo. More...

void	SynchronizeNumberOfColumns (Context const *ctx)
	Synchronize the number of columns across all workers. More...

bool	IsRowSplit () const
	Whether the data is split row-wise. More...

bool	IsColumnSplit () const
	Whether the data is split column-wise. More...

bool	IsRanking () const
	Whether this is a learning to rank data. More...

bool	IsVerticalFederated () const
	A convenient method to check if we are doing vertical federated learning, which requires some special processing. More...

bool	ShouldHaveLabels () const
	A convenient method to check if the MetaInfo should contain labels. More...

bool	HasCategorical () const
	Flag for whether the DMatrix has categorical features. More...

Public Attributes
uint64_t	num_row_ {0}
	number of rows in the data More...

uint64_t	num_col_ {0}
	number of columns in the data More...

uint64_t	num_nonzero_ {0}
	number of nonzero entries in the data More...

linalg::Tensor< float, 2 >	labels
	label of each instance More...

DataSplitMode	data_split_mode {DataSplitMode::kRow}
	data split mode More...

std::vector< bst_group_t >	group_ptr_
	the index of begin and end of a group needed when the learning task is ranking. More...

HostDeviceVector< bst_float >	weights_
	weights of each instance, optional More...

linalg::Tensor< float, 2 >	base_margin_
	initialized margins, if specified, xgboost will start from this init margin can be used to specify initial prediction to boost from. More...

HostDeviceVector< bst_float >	labels_lower_bound_
	lower bound of the label, to be used for survival analysis (censored regression) More...

HostDeviceVector< bst_float >	labels_upper_bound_
	upper bound of the label, to be used for survival analysis (censored regression) More...

std::vector< std::string >	feature_type_names
	Name of type for each feature provided by users. Eg. "int"/"float"/"i"/"q". More...

std::vector< std::string >	feature_names
	Name for each feature. More...

HostDeviceVector< FeatureType >	feature_types

HostDeviceVector< float >	feature_weights

Static Public Attributes
static constexpr uint64_t	kNumField = 12
	number of data fields in MetaInfo More...

Detailed Description

Meta information about dataset, always sit in memory.

Constructor & Destructor Documentation

◆ MetaInfo() [1/2]

xgboost::MetaInfo::MetaInfo ( )

default

default constructor

◆ MetaInfo() [2/2]

xgboost::MetaInfo::MetaInfo ( MetaInfo && that )

default

Member Function Documentation

◆ Clear()

void xgboost::MetaInfo::Clear ( )

clear all the information

◆ Copy()

MetaInfo xgboost::MetaInfo::Copy ( ) const

◆ Extend()

void xgboost::MetaInfo::Extend	(	MetaInfo const &	that,
		bool	accumulate_rows,
		bool	check_column
	)

Extend with other MetaInfo.

Parameters

that	The other MetaInfo object.
accumulate_rows	Whether rows need to be accumulated in this function. If client code knows number of rows in advance, set this parameter to false.
check_column	Whether the extend method should check the consistency of columns.

◆ GetFeatureInfo()

void xgboost::MetaInfo::GetFeatureInfo	(	const char *	field,
		std::vector< std::string > *	out_str_vecs
	)		const

◆ GetInfo()

void xgboost::MetaInfo::GetInfo	(	char const *	key,
		bst_ulong *	out_len,
		DataType	dtype,
		const void **	out_dptr
	)		const

◆ GetWeight()

bst_float xgboost::MetaInfo::GetWeight ( size_t i ) const

inline

Get weight of each instances.

Parameters

i	Instance index.

Returns: The weight.

◆ HasCategorical()

bool xgboost::MetaInfo::HasCategorical ( ) const

inline

Flag for whether the DMatrix has categorical features.

◆ IsColumnSplit()

bool xgboost::MetaInfo::IsColumnSplit ( ) const

inline

Whether the data is split column-wise.

◆ IsRanking()

bool xgboost::MetaInfo::IsRanking ( ) const

inline

Whether this is a learning to rank data.

◆ IsRowSplit()

bool xgboost::MetaInfo::IsRowSplit ( ) const

inline

Whether the data is split row-wise.

◆ IsVerticalFederated()

bool xgboost::MetaInfo::IsVerticalFederated ( ) const

A convenient method to check if we are doing vertical federated learning, which requires some special processing.

◆ LabelAbsSort()

const std::vector<size_t>& xgboost::MetaInfo::LabelAbsSort ( Context const * ctx ) const

get sorted indexes (argsort) of labels by absolute value (used by cox loss)

◆ LoadBinary()

void xgboost::MetaInfo::LoadBinary ( dmlc::Stream * fi )

Load the Meta info from binary stream.

Parameters

fi	The input stream

◆ operator=() [1/2]

MetaInfo& xgboost::MetaInfo::operator= ( MetaInfo && that )

default

◆ operator=() [2/2]

MetaInfo& xgboost::MetaInfo::operator= ( MetaInfo const & that )

delete

◆ SaveBinary()

void xgboost::MetaInfo::SaveBinary ( dmlc::Stream * fo ) const

Save the Meta info to binary stream.

Parameters

fo	The output stream.

◆ SetFeatureInfo()

void xgboost::MetaInfo::SetFeatureInfo	(	const char *	key,
		const char **	info,
		const bst_ulong	size
	)

◆ SetInfo()

void xgboost::MetaInfo::SetInfo	(	Context const &	ctx,
		StringView	key,
		StringView	interface_str
	)

Set information in the meta info with array interface.

Parameters

key	The key of the information.
interface_str	String representation of json format array interface.

◆ ShouldHaveLabels()

bool xgboost::MetaInfo::ShouldHaveLabels ( ) const

A convenient method to check if the MetaInfo should contain labels.

Normally we assume labels are available everywhere. The only exception is in vertical federated learning where labels are only available on worker 0.

◆ Slice()

MetaInfo xgboost::MetaInfo::Slice ( common::Span< int32_t const > ridxs ) const

◆ SynchronizeNumberOfColumns()

void xgboost::MetaInfo::SynchronizeNumberOfColumns ( Context const * ctx )

Synchronize the number of columns across all workers.

Normally we just need to find the maximum number of columns across all workers, but in vertical federated learning, since each worker loads its own list of columns, we need to sum them.

◆ Validate()

void xgboost::MetaInfo::Validate ( DeviceOrd device ) const

Validate all metainfo.

Member Data Documentation

◆ base_margin_

linalg::Tensor<float, 2> xgboost::MetaInfo::base_margin_

initialized margins, if specified, xgboost will start from this init margin can be used to specify initial prediction to boost from.

◆ data_split_mode

DataSplitMode xgboost::MetaInfo::data_split_mode {DataSplitMode::kRow}

data split mode

◆ feature_names

std::vector<std::string> xgboost::MetaInfo::feature_names

Name for each feature.

◆ feature_type_names

std::vector<std::string> xgboost::MetaInfo::feature_type_names

Name of type for each feature provided by users. Eg. "int"/"float"/"i"/"q".

◆ feature_types

HostDeviceVector<FeatureType> xgboost::MetaInfo::feature_types

◆ feature_weights

HostDeviceVector<float> xgboost::MetaInfo::feature_weights

◆ group_ptr_

std::vector<bst_group_t> xgboost::MetaInfo::group_ptr_

the index of begin and end of a group needed when the learning task is ranking.

◆ kNumField

constexpr uint64_t xgboost::MetaInfo::kNumField = 12

staticconstexpr

number of data fields in MetaInfo

◆ labels

linalg::Tensor<float, 2> xgboost::MetaInfo::labels

label of each instance

◆ labels_lower_bound_

HostDeviceVector<bst_float> xgboost::MetaInfo::labels_lower_bound_

lower bound of the label, to be used for survival analysis (censored regression)

◆ labels_upper_bound_

HostDeviceVector<bst_float> xgboost::MetaInfo::labels_upper_bound_

upper bound of the label, to be used for survival analysis (censored regression)

◆ num_col_

uint64_t xgboost::MetaInfo::num_col_ {0}

number of columns in the data

◆ num_nonzero_

uint64_t xgboost::MetaInfo::num_nonzero_ {0}

number of nonzero entries in the data

◆ num_row_

uint64_t xgboost::MetaInfo::num_row_ {0}

number of rows in the data

◆ weights_

HostDeviceVector<bst_float> xgboost::MetaInfo::weights_

weights of each instance, optional

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

/home/docs/checkouts/readthedocs.org/user_builds/xgboost/checkouts/latest/include/xgboost/data.h

Public Member Functions

Public Attributes

Static Public Attributes

Detailed Description

Constructor & Destructor Documentation

◆ MetaInfo() [1/2]

◆ MetaInfo() [2/2]

Member Function Documentation

◆ Clear()

◆ Copy()

◆ Extend()

◆ GetFeatureInfo()

◆ GetInfo()

◆ GetWeight()

◆ HasCategorical()

◆ IsColumnSplit()

◆ IsRanking()

◆ IsRowSplit()

◆ IsVerticalFederated()

◆ LabelAbsSort()

◆ LoadBinary()

◆ operator=() [1/2]

◆ operator=() [2/2]

◆ SaveBinary()

◆ SetFeatureInfo()

◆ SetInfo()

◆ ShouldHaveLabels()

◆ Slice()

◆ SynchronizeNumberOfColumns()

◆ Validate()

Member Data Documentation

◆ base_margin_

◆ data_split_mode

◆ feature_names

◆ feature_type_names

◆ feature_types

◆ feature_weights

◆ group_ptr_

◆ kNumField

◆ labels

◆ labels_lower_bound_

◆ labels_upper_bound_

◆ num_col_

◆ num_nonzero_

◆ num_row_

◆ weights_