ultra::src::columns_info Class Reference

#include <columns_info.h>

Classes
class	column_info
	Information about a single column of the dataset. More...

Public Types
using	const_iterator = std::vector<column_info>::const_iterator
using	iterator = std::vector<column_info>::iterator
using	size_type = std::size_t
using	value_type = column_info

Public Member Functions
	columns_info (const columns_info &)
	columns_info (const std::vector< std::pair< std::string, domain_t > > &)
columns_info &	operator= (const columns_info &)
const column_info &	operator[] (size_type) const
column_info &	operator[] (size_type)
const column_info &	operator[] (const std::string &) const
const column_info &	front () const
column_info &	front ()
const column_info &	back () const
column_info &	back ()
size_type	size () const noexcept
bool	empty () const noexcept
const_iterator	begin () const noexcept
iterator	begin () noexcept
const_iterator	end () const noexcept
iterator	end () noexcept
const_iterator	cbegin () const noexcept
const_iterator	cend () const noexcept
void	pop_back ()
void	push_back (const column_info &)
void	push_front (const column_info &)
void	data_typing (typing) noexcept
template<DataframeMatrix R>
void	build (const R &, std::optional< std::size_t >)
std::set< symbol::category_t >	used_categories () const
domain_t	domain_of_category (symbol::category_t) const
task_t	task () const noexcept
bool	is_valid () const
template<DataframeMatrix R>
void	build (const R &exs, std::optional< std::size_t > output_index)

Detailed Description

Metadata container describing the columns of a dataframe.

This class stores and manages column-level information such as:

• column names;
• inferred domains;
• input/output roles;
• typing constraints.

It also provides facilities to infer such metadata from a dataframe-like range and to query the resulting schema.

Constructor & Destructor Documentation

columns_info() [1/2]

ultra::src::columns_info::columns_info

(

const columns_info &

other

)

Copy constructor.

Parameters

[in]

other

another container to be used as source to initialize the elements of the container with

Constructs the container with the copy of the contents of other.

columns_info() [2/2]

ultra::src::columns_info::columns_info

(

const std::vector< std::pair< std::string, domain_t > > &

schema

)

Constructs the object from a user provided-schema.

Parameters

[in]

schema

user-provided schema for the dataframe

Member Function Documentation

back()

const columns_info::column_info & ultra::src::columns_info::back ( ) const

nodiscard

Returns a reference to the last element in the container.

Warning

Calling back on an empty container causes undefined behavior.

build() [1/2]

template<DataframeMatrix R>

void ultra::src::columns_info::build	(	const R &	exs,
		std::optional< std::size_t >	output_index )

Infer column metadata from a dataframe-like range.

Template Parameters

R	a DataframeMatrix type

Parameters

[in]	exs	a non-empty range of rows representing the dataset. The first row must contain column headers
[in]	output_index	optional index of the output column. If not provided, a surrogate output column may be inserted

Precondition

The range must not be empty

Rows must have non-zero size.

The first row of exs is interpreted as the header row. An optional output column may be designated via output_index; when present, the corresponding column is treated as the output and is normalised to appear first during analysis.

Domain inference is performed on a bounded reservoir sample of valid input rows.

Remarks

Rows whose normalised width differs from the header are ignored during domain inference.

build() [2/2]

template<DataframeMatrix R>

void ultra::src::columns_info::build	(	const R &	exs,
		std::optional< std::size_t >	output_index )

Infer column metadata from a dataframe-like range.

Template Parameters

R	a DataframeMatrix type

Parameters

[in]	exs	a non-empty range of rows representing the dataset. The first row must contain column headers
[in]	output_index	optional index of the output column. If not provided, a surrogate output column may be inserted

Precondition

The range must not be empty

Rows must have non-zero size.

The first row of exs is interpreted as the header row. An optional output column may be designated via output_index; when present, the corresponding column is treated as the output and is normalised to appear first during analysis.

Domain inference is performed on a bounded reservoir sample of valid input rows.

Remarks

Rows whose normalised width differs from the header are ignored during domain inference.

data_typing()

void ultra::src::columns_info::data_typing ( typing t )

noexcept

Set the data typing system used for category identification.

Parameters

[in]

t

type of data typing (see typing for details)

domain_of_category()

domain_t ultra::src::columns_info::domain_of_category ( symbol::category_t target ) const

nodiscard

Return the domain associated with a symbol category.

Parameters

[in]

target

category of interest

Returns

domain corresponding to the given category

front()

const columns_info::column_info & ultra::src::columns_info::front ( ) const

nodiscard

Returns a reference to the first element in the container.

Warning

Calling front on an empty container causes undefined behavior.

is_valid()

bool ultra::src::columns_info::is_valid ( ) const

nodiscard

Returns

true if all columns have valid domains and consistent categories

operator=()

columns_info & ultra::src::columns_info::operator=

(

const columns_info &

other

)

Copy assignment operator.

Parameters

[in]

other

another container to use as the data source

Returns

*this

Replaces the contents with a copy of the contents of other.

operator[]() [1/3]

const columns_info::column_info & ultra::src::columns_info::operator[] ( const std::string & name ) const

nodiscard

Returns a reference to column_info with specified column header.

Parameters

[in]

name

the name of the column_info to find

Returns

a reference to the column_info with the given column header

Exceptions

`std::out_of_range`

if the container does not contain a column with the specified name

operator[]() [2/3]

columns_info::column_info & ultra::src::columns_info::operator[] ( size_type i )

nodiscard

Access column metadata by index.

Parameters

[in]

i

column index

Returns

reference to the corresponding column_info

operator[]() [3/3]

const columns_info::column_info & ultra::src::columns_info::operator[] ( size_type i ) const

nodiscard

Access column metadata by index.

Parameters

[in]

i

column index

Returns

constant reference to the corresponding column_info

pop_back()

void ultra::src::columns_info::pop_back

(

)

Removes the last column from the container.

Warning

Ensure the container is not empty before calling (possible undefined behavior).

push_back()

void ultra::src::columns_info::push_back

(

const column_info &

v

)

Adds a new column at the end of the column list.

Parameters

[in]

v

information about the new column

push_front()

void ultra::src::columns_info::push_front

(

const column_info &

v

)

Inserts a new column at the beginning of the list, moving existing columns forward.

Parameters

[in]

v

information about the new column

size()

columns_info::size_type ultra::src::columns_info::size ( ) const

nodiscardnoexcept

Returns

the number of columns in the container

task()

task_t ultra::src::columns_info::task ( ) const

nodiscardnoexcept

Return the learning task inferred for the columns.

Returns

current task type

used_categories()

std::set< symbol::category_t > ultra::src::columns_info::used_categories

(

)

const

Returns

a set of all categories used across columns

---

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

• columns_info.h
• columns_info.cc
• columns_info.tcc