plssvm::detail Namespace Reference

Namespace containing implementation details. Should not directly be used by users. More...

Namespaces
	cmd
	Namespace containing implementation details for the command line interface functionality. Should not directly be used by users.
	io
	Namespace containing implementation details for the IO related functions. Should not directly be used by users.

Classes
struct	csvm_backend_exists< cuda::csvm >
	Sets the value to true since C-SVMs using the CUDA backend are available. More...
class	gpu_csvm
	A C-SVM implementation for all GPU backends to reduce code duplication. More...
class	gpu_device_ptr
	Small wrapper class around a GPU device pointer together with commonly used device functions for all GPU backends to reduce code duplication. More...
struct	csvm_backend_exists< hip::csvm >
	Sets the value to true since C-SVMs using the HIP backend are available. More...
struct	csvm_backend_exists< opencl::csvm >
	Sets the value to true since C-SVMs using the OpenCL backend are available. More...
struct	csvm_backend_exists< openmp::csvm >
	Sets the value to true since C-SVMs using the OpenMP backend are available. More...
struct	csvm_backend_exists< dpcpp::csvm >
	Sets the value to true since C-SVMs using the SYCL backend with DPC++ as SYCL implementation are available. More...
struct	csvm_backend_exists< hipsycl::csvm >
	Sets the value to true since C-SVMs using the SYCL backend with hipSYCL as SYCL implementation are available. More...
class	execution_range
	Class specifying a backend independent execution range. More...
struct	tracking_entry
	A single tracking entry containing a specific category, a unique name, and the actual value to be tracked. More...
struct	is_tracking_entry
	Check whether T is a tracking entry. Ignores all top-level const, volatile, and reference qualifiers. More...
class	performance_tracker
	Store the tracked information during calls to plssvm-train, plssvm-predict, and plssvm-scale. More...
class	sha256
	A hashing struct used for sha256 hashing. More...
struct	real_type_label_type_combination
	Encapsulates a combination of a used real_type (float or double) and label_type (an arithmetic type or std::string). More...
struct	concat_tuple_types
struct	concat_tuple_types< std::tuple< FirstTupleTypes... >, std::tuple< SecondTupleTypes... > >
	Concatenate the types of the two tuples to a new tuple type. More...
struct	cartesian_type_product
struct	cartesian_type_product< std::tuple< FirstTupleType, FirstTupleRemainingTypes... >, std::tuple< SecondTupleTypes... > >
	Calculate the cartesian product of the types in two tuples and return a new tuple with the corresponding real_type_label_type_combination types. More...
struct	cartesian_type_product< std::tuple<>, std::tuple< SecondTupleTypes... > >
struct	type_list_contains
	Checks whether the type T is present in the Tuple. More...
struct	type_list_contains< T, std::tuple< Types... > >
	Checks whether the type T is present in the tuple Types. More...
struct	is_array
	Type trait to check whether T is a std::array. More...
struct	is_array< std::array< T, I > >
	Type trait to check whether T is a std::array. More...
struct	is_vector
	Type trait to check whether T is a std::vector. More...
struct	is_vector< std::vector< T > >
	Type trait to check whether T is a std::vector. More...
struct	is_deque
	Type trait to check whether T is a std::deque. More...
struct	is_deque< std::deque< T > >
	Type trait to check whether T is a std::deque. More...
struct	is_forward_list
	Type trait to check whether T is a std::forward_list. More...
struct	is_forward_list< std::forward_list< T > >
	Type trait to check whether T is a std::forward_list. More...
struct	is_list
	Type trait to check whether T is a std::list. More...
struct	is_list< std::list< T > >
	Type trait to check whether T is a std::list. More...
struct	is_set
	Type trait to check whether T is a std::set. More...
struct	is_set< std::set< T > >
	Type trait to check whether T is a std::set. More...
struct	is_map
	Type trait to check whether T is a std::map. More...
struct	is_map< std::map< Key, T > >
	Type trait to check whether T is a std::map. More...
struct	is_multiset
	Type trait to check whether T is a std::multiset. More...
struct	is_multiset< std::multiset< T > >
	Type trait to check whether T is a std::multiset. More...
struct	is_multimap
	Type trait to check whether T is a std::multimap. More...
struct	is_multimap< std::multimap< Key, T > >
	Type trait to check whether T is a std::multimap. More...
struct	is_unordered_set
	Type trait to check whether T is a std::unordered_set. More...
struct	is_unordered_set< std::unordered_set< T > >
	Type trait to check whether T is a std::unordered_set. More...
struct	is_unordered_map
	Type trait to check whether T is a std::unordered_map. More...
struct	is_unordered_map< std::unordered_map< Key, T > >
	Type trait to check whether T is a std::unordered_map. More...
struct	is_unordered_multiset
	Type trait to check whether T is a std::unordered_multiset. More...
struct	is_unordered_multiset< std::unordered_multiset< T > >
	Type trait to check whether T is a std::unordered_multiset. More...
struct	is_unordered_multimap
	Type trait to check whether T is a std::unordered_multimap. More...
struct	is_unordered_multimap< std::unordered_multimap< Key, T > >
	Type trait to check whether T is a std::unordered_multimap. More...
struct	parameter
	Class for encapsulating all important C-SVM parameters. More...

Typedefs
template<typename... T>
using	concat_tuple_types_t = typename concat_tuple_types< T... >::type
	Shorthand for the typename concat_tuple_types<...>::type type.
template<typename... T>
using	cartesian_type_product_t = typename cartesian_type_product< T... >::type
	Shorthand for the typename cartesian_type_product<...>::type type.
using	real_type_list = std::tuple< float, double >
	A type list of all supported real types as std::tuple.
using	label_type_list = std::tuple< bool, char, signed char, unsigned char, short, unsigned short, int, unsigned int, long, unsigned long, long long, unsigned long long, float, double, long double, std::string >
	A type list of all supported label types (currently arithmetic types and std::string) as std::tuple.
using	real_type_label_type_combination_list = detail::cartesian_type_product_t< real_type_list, label_type_list >
	The cartesian product of all real types and label types as std::tuple.
template<typename T >
using	remove_cvref_t = std::remove_cv_t< std::remove_reference_t< T > >
	Remove the topmost reference- and cv-qualifiers. More...

Enumerations
enum class	layout_type { aos , soa }
	Enum class for all available layout types. More...

Functions
template<typename csvm_type , typename... Args>
std::unique_ptr< csvm >	make_csvm_default_impl ([[maybe_unused]] Args &&...args)
	Construct a C-SVM using the parameters args. More...
template<typename... Args>
std::unique_ptr< csvm >	make_csvm_sycl_impl ([[maybe_unused]] Args &&...args)
	Construct a SYCL C-SVM using the parameters args. More...
template<typename... Args>
std::unique_ptr< csvm >	make_csvm_impl (const backend_type backend, Args &&...args)
	Create a new C-SVM using the backend type and the additional parameter args. More...
template<typename T >
constexpr std::string_view	arithmetic_type_name ()=delete
	Tries to convert the given type to its name as string representation including possible const and/or volatile qualifiers. More...
template<typename... Args>
void	check_assertion (const bool cond, const std::string_view cond_str, const source_location &loc, const std::string_view msg, Args &&...args)
	Function called by the PLSSVM_ASSERT macro. Checks the assertion condition. If the condition evaluates to false, prints the assertion condition together with additional information (e.g., plssvm::source_location information) and aborts the program. More...
std::ostream &	operator<< (std::ostream &out, const execution_range &range)
	Output the execution range to the given output-stream out. More...
std::ostream &	operator<< (std::ostream &out, layout_type layout)
	Output the layout to the given output-stream out. More...
std::istream &	operator>> (std::istream &in, layout_type &layout)
	Use the input-stream in to initialize the layout type. More...
template<typename real_type >
std::vector< real_type >	transform_to_aos_layout (const std::vector< std::vector< real_type >> &matrix, const std::size_t boundary_size, const std::size_t num_points, const std::size_t num_features)
	Convert a 2D matrix into a 1D array in Array-of-Structs layout adding boundary_size values per data point for padding. More...
template<typename real_type >
std::vector< real_type >	transform_to_soa_layout (const std::vector< std::vector< real_type >> &matrix, const std::size_t boundary_size, const std::size_t num_points, const std::size_t num_features)
	Convert a 2D matrix into a 1D array in Struct-of-Arrays layout adding boundary_size values per feature for padding. More...
template<typename real_type >
std::vector< real_type >	transform_to_layout (const layout_type layout, const std::vector< std::vector< real_type >> &matrix, const std::size_t boundary_size, const std::size_t num_points)
	Convert a 2D matrix into a 1D array in the layout adding boundary_size values per data point or feature respectively for padding. More...
template<typename... Args>
void	log (const verbosity_level verb, const std::string_view msg, Args &&...args)
template<typename T >
std::ostream &	operator<< (std::ostream &out, const tracking_entry< T > &entry)
	Output the tracking entry to the given output-stream out. Only the tracked value is output excluding the category and name. More...
template<typename T , typename Exception = std::runtime_error, PLSSVM_REQUIRES((std::is_arithmetic_v< T >||std::is_same_v< remove_cvref_t< T >, std::string >)) >
T	convert_to (const std::string_view str)
	Converts the string str to a value of type T. More...
template<typename T >
T	extract_first_integer_from_string (std::string_view str)
	Extract the first integer from the given string str and converts it to T ignoring a potential sign. More...
template<typename T >
std::vector< T >	split_as (const std::string_view str, const char delim=' ')
	Split the string str at the positions with delimiter delim and return the sub-strings converted to the type T. More...
bool	starts_with (std::string_view str, std::string_view sv) noexcept
	Checks if the string str starts with the prefix sv. More...
bool	starts_with (std::string_view str, char c) noexcept
	Checks if the string str starts with the character c. More...
bool	ends_with (std::string_view str, std::string_view sv) noexcept
	Checks if the string str ends with the suffix sv. More...
bool	ends_with (std::string_view str, char c) noexcept
	Checks if the string str ends with the character c. More...
bool	contains (std::string_view str, std::string_view sv) noexcept
	Checks if the string str contains the string sv. More...
bool	contains (std::string_view str, char c) noexcept
	Checks if the string str contains the character c. More...
std::string_view	trim_left (std::string_view str) noexcept
	Returns a new std::string_view equal to str where all leading whitespaces are removed. More...
std::string_view	trim_right (std::string_view str) noexcept
	Returns a new std::string_view equal to str where all trailing whitespaces are removed. More...
std::string_view	trim (std::string_view str) noexcept
	Returns a new std::string_view equal to str where all leading and trailing whitespaces are removed. More...
std::string &	replace_all (std::string &str, std::string_view what, std::string_view with)
	Replaces all occurrences of what with with in the string str. More...
std::string &	to_lower_case (std::string &str)
	Convert the string str to its all lower case representation. More...
std::string	as_lower_case (std::string_view str)
	Return a new string with the same content as str but all lower case. More...
std::string &	to_upper_case (std::string &str)
	Convert the string str to its all upper case representation. More...
std::string	as_upper_case (std::string_view str)
	Return a new string with the same content as str but all upper case. More...
std::vector< std::string_view >	split (std::string_view str, char delim=' ')
	Split the string str at the positions with delimiter delim and return the sub-strings. More...
void	unreachable ()
	Invokes undefined behavior. Used to mark code paths that may never be reachable. More...
template<std::size_t I, typename... Types>
constexpr decltype(auto)	get (Types &&...args) noexcept
	Get the I-th element of the parameter pack args at compile-time. More...
template<typename Enum >
constexpr std::underlying_type_t< Enum >	to_underlying (const Enum e) noexcept
	Converts an enumeration to its underlying type. More...
template<typename Enum >
constexpr std::underlying_type_t< Enum >	to_underlying (const default_value< Enum > &e) noexcept
	Converts an enumeration wrapped in a plssvm::default_value to its underlying type. More...
template<typename Container , typename Pred , PLSSVM_REQUIRES(is_container_v< Container >) >
Container::size_type	erase_if (Container &c, Pred pred)
	Implements an erase_if function for different containers according to https://en.cppreference.com/w/cpp/container/map/erase_if. More...
template<typename Container , typename T , PLSSVM_REQUIRES(is_container_v< Container >) >
bool	contains (const Container &c, const T &val)
	Check whether the Container c contains the value val. More...
std::string	current_date_time ()
	Return the current date time in the format "YYYY-MM-DD hh:mm:ss". More...
template<typename ExpectedType , typename IgorParser , typename NamedArgType >
ExpectedType	get_value_from_named_parameter (const IgorParser &parser, const NamedArgType &named_arg)
	Parse the value hold be named_arg and return it converted to the ExpectedType. More...
template<typename T >
constexpr bool	operator== (const parameter< T > &lhs, const parameter< T > &rhs) noexcept
	Compares the two parameter sets lhs and rhs for equality. More...
template<typename T >
constexpr bool	operator!= (const parameter< T > &lhs, const parameter< T > &rhs) noexcept
	Compares the two parameter sets lhs and rhs for inequality. More...
template<typename T >
constexpr bool	equivalent (const parameter< T > &lhs, const parameter< T > &rhs) noexcept
	Checks whether the two parameter sets lhs and rhs are equivalent. More...
template<typename T >
std::ostream &	operator<< (std::ostream &out, const parameter< T > &params)
	Output all parameters encapsulated by params to the given output-stream out. More...

Variables
template<typename T >
constexpr bool	is_tracking_entry_v = is_tracking_entry<T>::value
std::shared_ptr< performance_tracker >	global_tracker
	The global performance tracker instance used for the default tracking.
template<typename T , typename Tuple >
constexpr bool	type_list_contains_v = type_list_contains<T, Tuple>::value
	Checks whether the type T is present in the Tuple.
template<typename >
constexpr bool	always_false_v = false
	Type-dependent expression that always evaluates to false.
template<typename T >
constexpr bool	is_array_v = is_array<T>::value
	Type trait to check whether T is a std::array.
template<typename T >
constexpr bool	is_vector_v = is_vector<T>::value
	Type trait to check whether T is a std::vector.
template<typename T >
constexpr bool	is_deque_v = is_deque<T>::value
	Type trait to check whether T is a std::deque.
template<typename T >
constexpr bool	is_forward_list_v = is_forward_list<T>::value
	Type trait to check whether T is a std::forward_list.
template<typename T >
constexpr bool	is_list_v = is_list<T>::value
	Type trait to check whether T is a std::list.
template<typename T >
constexpr bool	is_sequence_container_v = is_array_v<T> || is_vector_v<T> || is_deque_v<T> || is_forward_list_v<T> || is_list_v<T>
	Type trait to check whether T is a sequence container.
template<typename T >
constexpr bool	is_set_v = is_set<T>::value
	Type trait to check whether T is a std::set.
template<typename T >
constexpr bool	is_map_v = is_map<T>::value
	Type trait to check whether T is a std::map.
template<typename T >
constexpr bool	is_multiset_v = is_multiset<T>::value
	Type trait to check whether T is a std::multiset.
template<typename T >
constexpr bool	is_multimap_v = is_multimap<T>::value
	Type trait to check whether T is a std::multimap.
template<typename T >
constexpr bool	is_associative_container_v = is_set_v<T> || is_map_v<T> || is_multimap_v<T> || is_multiset_v<T>
	Type trait to check whether T is a associative container.
template<typename T >
constexpr bool	is_unordered_set_v = is_unordered_set<T>::value
	Type trait to check whether T is a std::unordered_set.
template<typename T >
constexpr bool	is_unordered_map_v = is_unordered_map<T>::value
	Type trait to check whether T is a std::unordered_map.
template<typename T >
constexpr bool	is_unordered_multiset_v = is_unordered_multiset<T>::value
	Type trait to check whether T is a std::unordered_multiset.
template<typename T >
constexpr bool	is_unordered_multimap_v = is_unordered_multimap<T>::value
	Type trait to check whether T is a std::unordered_multimap.
template<typename T >
constexpr bool	is_unordered_associative_container_v = is_unordered_set_v<T> || is_unordered_map_v<T> || is_unordered_multimap_v<T> || is_unordered_multiset_v<T>
	Type trait to check whether T is a unordered associative container.
template<typename T >
constexpr bool	is_container_v = is_sequence_container_v<T> || is_associative_container_v<T> || is_unordered_associative_container_v<T>
	Type trait to check whether T is a container.
template<typename... Args>
constexpr bool	has_only_named_args_v = !igor::has_unnamed_arguments<Args...>()
	Trait to check whether Args only contains named-parameter.
template<typename... Args>
constexpr bool	has_only_parameter_named_args_v = !igor::has_other_than<Args...>(plssvm::kernel_type, plssvm::gamma, plssvm::degree, plssvm::coef0, plssvm::cost)
	Trait to check whether Args only contains named-parameter that can be used to initialize a plssvm::parameter struct.
template<typename... Args>
constexpr bool	has_only_sycl_parameter_named_args_v = !igor::has_other_than<Args...>(plssvm::kernel_type, plssvm::gamma, plssvm::degree, plssvm::coef0, plssvm::cost, plssvm::sycl_implementation_type, plssvm::sycl_kernel_invocation_type)
	Trait to check whether Args only contains named-parameter that can be used to initialize a plssvm::parameter struct including SYCL specific named-parameters.

Detailed Description

Namespace containing implementation details. Should not directly be used by users.

Typedef Documentation

remove_cvref_t

template<typename T >

using plssvm::detail::remove_cvref_t = typedef std::remove_cv_t<std::remove_reference_t<T> >

Remove the topmost reference- and cv-qualifiers.

For more information see std::remove_cvref_t.

Enumeration Type Documentation

layout_type

enum plssvm::detail::layout_type

strong

Enum class for all available layout types.

Enumerator
aos	Array-of-Structs (AoS)
soa	Structs-of-Arrays (SoA)

Function Documentation

make_csvm_default_impl()

template<typename csvm_type , typename... Args>

std::unique_ptr<csvm> plssvm::detail::make_csvm_default_impl ( [[maybe_unused] ] Args &&...  args )

inline

Construct a C-SVM using the parameters args.

The default case, no special parameters for the C-SVMs are necessary.

Template Parameters

csvm_type	the type of the C-SVM
Args	the types of the parameters to initialize the C-SVM

Parameters

[in]

args

the parameters used to initialize the C-SVM

Exceptions

plssvm::unsupported_backend_exception

if the backend is not recognized

Returns

the C-SVM ([[nodiscard]])

make_csvm_sycl_impl()

template<typename... Args>

std::unique_ptr<csvm> plssvm::detail::make_csvm_sycl_impl ( [[maybe_unused] ] Args &&...  args )

inline

Construct a SYCL C-SVM using the parameters args.

The special case for the SYCL backend to handle the SYCL specific parameters.

Template Parameters

Args	the types of the parameters to initialize the SYCL C-SVM

Parameters

[in]

args

the parameters used to initialize the SYCL C-SVM

Exceptions

plssvm::unsupported_backend_exception

if the backend is not recognized

Returns

the SYCL C-SVM ([[nodiscard]])

make_csvm_impl()

template<typename... Args>

std::unique_ptr<csvm> plssvm::detail::make_csvm_impl ( const backend_type  backend, Args &&...  args  )

inline

Create a new C-SVM using the backend type and the additional parameter args.

Template Parameters

Args	the types of the parameters to initialize the C-SVM

Parameters

[in]	backend	the backend to use
[in]	args	the parameters used to initialize the respective C-SVM

Exceptions

plssvm::unsupported_backend_exception

if the backend is not recognized

Returns

the C-SVM ([[nodiscard]])

arithmetic_type_name()

template<typename T >

constexpr std::string_view plssvm::detail::arithmetic_type_name ( )

inlineconstexprdelete

Tries to convert the given type to its name as string representation including possible const and/or volatile qualifiers.

The definition is marked as deleted if T isn't an arithmetic type.

Template Parameters

T	the type to convert to a string

Returns

the name of T ([[nodiscard]])

check_assertion()

template<typename... Args>

void plssvm::detail::check_assertion ( const bool  cond, const std::string_view  cond_str, const source_location &  loc, const std::string_view  msg, Args &&...  args  )

inline

Function called by the PLSSVM_ASSERT macro. Checks the assertion condition. If the condition evaluates to false, prints the assertion condition together with additional information (e.g., plssvm::source_location information) and aborts the program.

Template Parameters

Args	the placeholder types

Parameters

[in]	cond	the assertion condition, aborts the program if evaluated to false
[in]	cond_str	the assertion condition as string
[in]	loc	the source location where the assertion appeared
[in]	msg	the custom assertion message
[in]	args	the placeholder values

operator<<() [1/4]

std::ostream& plssvm::detail::operator<<	(	std::ostream &	out,
		const execution_range &	range
	)

Output the execution range to the given output-stream out.

Parameters

[in,out]	out	the output-stream to write the execution range to
[in]	range	the execution range

Returns

the output-stream

operator<<() [2/4]

std::ostream& plssvm::detail::operator<<	(	std::ostream &	out,
		layout_type	layout
	)

Output the layout to the given output-stream out.

Parameters

[in,out]	out	the output-stream to write the layout type to
[in]	layout	the layout type

Returns

the output-stream

operator>>()

std::istream& plssvm::detail::operator>>	(	std::istream &	in,
		layout_type &	layout
	)

Use the input-stream in to initialize the layout type.

Parameters

[in,out]	in	input-stream to extract the layout type from
[in]	layout	the layout type

Returns

the input-stream

transform_to_aos_layout()

template<typename real_type >

std::vector<real_type> plssvm::detail::transform_to_aos_layout ( const std::vector< std::vector< real_type >> &  matrix, const std::size_t  boundary_size, const std::size_t  num_points, const std::size_t  num_features  )

inline

Convert a 2D matrix into a 1D array in Array-of-Structs layout adding boundary_size values per data point for padding.

The 1D array is assembled data point wise. For more information see: https://www.wikiwand.com/en/AoS_and_SoA.

Template Parameters

real_type

the type of the values in the 2D matrix and the transformed 1D array

Parameters

[in]	matrix	the 2D matrix to transform
[in]	boundary_size	the number of boundary elements to insert as padding
[in]	num_points	the number of data points to transform
[in]	num_features	the number of features per data point in the 2D matrix

Returns

the transformed 1D array in Array-of-Structs layout ([[nodiscard]])

transform_to_soa_layout()

template<typename real_type >

std::vector<real_type> plssvm::detail::transform_to_soa_layout ( const std::vector< std::vector< real_type >> &  matrix, const std::size_t  boundary_size, const std::size_t  num_points, const std::size_t  num_features  )

inline

Convert a 2D matrix into a 1D array in Struct-of-Arrays layout adding boundary_size values per feature for padding.

The 1D array is assembled feature wise. For more information see: https://www.wikiwand.com/en/AoS_and_SoA.

Template Parameters

real_type

the type of the values in the 2D matrix and the transformed 1D array

Parameters

[in]	matrix	the 2D matrix to transform
[in]	boundary_size	the number of boundary elements to insert as padding
[in]	num_points	the number of data points to transform
[in]	num_features	the number of features per data point in the 2D matrix

Returns

the transformed 1D array in Struct-of-Arrays layout ([[nodiscard]])

transform_to_layout()

template<typename real_type >

std::vector<real_type> plssvm::detail::transform_to_layout ( const layout_type  layout, const std::vector< std::vector< real_type >> &  matrix, const std::size_t  boundary_size, const std::size_t  num_points  )

inline

Convert a 2D matrix into a 1D array in the layout adding boundary_size values per data point or feature respectively for padding.

For more information regarding AoS and SoA see: https://www.wikiwand.com/en/AoS_and_SoA.

Template Parameters

real_type

the type of the values in the 2D matrix and the transformed 1D array

Parameters

[in]	layout	the layout type to transform the 2D matrix to, either Array-of-Structs (AoS) or Struct-of-Arrays (SoA)
[in]	matrix	the 2D matrix to transform
[in]	boundary_size	the number of boundary elements to insert as padding
[in]	num_points	the number of data points to transform

Returns

the transformed 1D array in the specified layout ([[nodiscard]])

log()

template<typename... Args>

void plssvm::detail::log	(	const verbosity_level	verb,
		const std::string_view	msg,
		Args &&...	args
	)

@breif Output the message msg filling the {fmt} like placeholders with args to the standard output stream.

If a value in Args is of type plssvm::detail::tracking_entry and performance tracking is enabled, this is also added to the plssvm::detail::performance_tracker. Only logs the message if the verbosity level matches the plssvm::verbosity level.

Template Parameters

Args	the types of the placeholder values

Parameters

[in]	verb	the verbosity level of the message to log; must match the plssvm::verbosity level to log the message
[in]	msg	the message to print on the standard output stream if requested (i.e., plssvm::verbose is true)
[in]	args	the values to fill the {fmt}-like placeholders in msg

operator<<() [3/4]

template<typename T >

std::ostream& plssvm::detail::operator<<	(	std::ostream &	out,
		const tracking_entry< T > &	entry
	)

Output the tracking entry to the given output-stream out. Only the tracked value is output excluding the category and name.

Template Parameters

T	the type of the value tracked in the tracking entry

Parameters

[in,out]	out	the output-stream to write the tracking entry to
[in]	entry	the tracking entry

Returns

the output-stream

convert_to()

template<typename T , typename Exception = std::runtime_error, PLSSVM_REQUIRES((std::is_arithmetic_v< T >||std::is_same_v< remove_cvref_t< T >, std::string >)) >

T plssvm::detail::convert_to ( const std::string_view  str )

inline

Converts the string str to a value of type T.

If T is an integral type std::from_chars is used, if T is a floating point type float_fast::from_chars is used, if T is a std::string the normal std::string constructor is used, and if T is neither of the three, an exception is thrown.

Template Parameters

T	the type to convert the value of str to, must be an arithmetic type or a std::string
Exception	the exception type to throw in case that str can't be converted to a value of T (default: std::runtime_error).

Parameters

[in]

str

the string to convert

Exceptions

Exception

if str can't be converted to a value of type T

Returns

the value of type T denoted by str ([[nodiscard]])

extract_first_integer_from_string()

template<typename T >

T plssvm::detail::extract_first_integer_from_string ( std::string_view  str )

inline

Extract the first integer from the given string str and converts it to T ignoring a potential sign.

Template Parameters

T	the type to convert the first integer to

Parameters

[in]

str

the string to check

Exceptions

std::runtime_error

if str doesn't contain an integer

Returns

the converted first integer of type T ([[nodiscard]])

split_as()

template<typename T >

std::vector<T> plssvm::detail::split_as ( const std::string_view  str, const char  delim = ' '  )

inline

Split the string str at the positions with delimiter delim and return the sub-strings converted to the type T.

Template Parameters

T	the type to convert the value of str to, must be an arithmetic type or a std::string

Parameters

[in]	str	the string to split
[in]	delim	the split delimiter

Returns

the split sub-strings converted to the type T ([[nodiscard]])

starts_with() [1/2]

bool plssvm::detail::starts_with ( std::string_view  str, std::string_view  sv  )

noexcept

Checks if the string str starts with the prefix sv.

Parameters

[in]	str	the string to check
[in]	sv	the string to match against the start of str

Returns

true if str starts with the string sv, otherwise false ([[nodiscard]])

starts_with() [2/2]

bool plssvm::detail::starts_with ( std::string_view  str, char  c  )

noexcept

Checks if the string str starts with the character c.

Parameters

[in]	str	the string to check
[in]	c	the character to match against the first character of str

Returns

true if str starts with the character c, otherwise false ([[nodiscard]])

ends_with() [1/2]

bool plssvm::detail::ends_with ( std::string_view  str, std::string_view  sv  )

noexcept

Checks if the string str ends with the suffix sv.

Parameters

[in]	str	the string to check
[in]	sv	the string to match against the end of str

Returns

true if str ends with the string sv, otherwise false ([[nodiscard]])

ends_with() [2/2]

bool plssvm::detail::ends_with ( std::string_view  str, char  c  )

noexcept

Checks if the string str ends with the character c.

Parameters

[in]	str	the string to check
[in]	c	the character to match against the last character of str

Returns

true if str ends with the character c, otherwise false ([[nodiscard]])

contains() [1/3]

bool plssvm::detail::contains ( std::string_view  str, std::string_view  sv  )

noexcept

Checks if the string str contains the string sv.

Parameters

[in]	str	the string to check
[in]	sv	the string to find

Returns

true if str contains the string sv, otherwise false ([[nodiscard]])

contains() [2/3]

bool plssvm::detail::contains ( std::string_view  str, char  c  )

noexcept

Checks if the string str contains the character c.

Parameters

[in]	str	the string to check
[in]	c	the character to find

Returns

true if str contains the character c, otherwise false ([[nodiscard]])

trim_left()

std::string_view plssvm::detail::trim_left ( std::string_view  str )

noexcept

Returns a new std::string_view equal to str where all leading whitespaces are removed.

Parameters

[in]

str

the string to remove the leading whitespaces

Returns

the string str without leading whitespace ([[nodiscard]])

trim_right()

std::string_view plssvm::detail::trim_right ( std::string_view  str )

noexcept

Returns a new std::string_view equal to str where all trailing whitespaces are removed.

Parameters

[in]

str

the string to remove the trailing whitespaces

Returns

the string str without trailing whitespace ([[nodiscard]])

trim()

std::string_view plssvm::detail::trim ( std::string_view  str )

noexcept

Returns a new std::string_view equal to str where all leading and trailing whitespaces are removed.

Parameters

[in]

str

the string to remove the leading and trailing whitespaces

Returns

the string str without leading and trailing whitespace ([[nodiscard]])

replace_all()

std::string& plssvm::detail::replace_all	(	std::string &	str,
		std::string_view	what,
		std::string_view	with
	)

Replaces all occurrences of what with with in the string str.

Parameters

[in,out]	str	the string to replace the values
[in]	what	the string to replace
[in]	with	the string to replace with

Returns

the replaced string

to_lower_case()

std::string& plssvm::detail::to_lower_case

(

std::string &

str

)

Convert the string str to its all lower case representation.

Parameters

[in,out]

str

the string to transform

Returns

the transformed string

as_lower_case()

std::string plssvm::detail::as_lower_case

(

std::string_view

str

)

Return a new string with the same content as str but all lower case.

In contrast to to_lower_case(std::string&) this function does not change the input string str.

Parameters

[in]

str

the string to use in the transformation

Returns

the transformed string ([[nodiscard]])

to_upper_case()

std::string& plssvm::detail::to_upper_case

(

std::string &

str

)

Convert the string str to its all upper case representation.

Parameters

[in,out]

str

the string to transform

Returns

the transformed string

as_upper_case()

std::string plssvm::detail::as_upper_case

(

std::string_view

str

)

Return a new string with the same content as str but all upper case.

In contrast to to_upper_case(std::string&) this function does not change the input string str.

Parameters

[in]

str

the string to use in the transformation

Returns

the transformed string ([[nodiscard]])

split()

std::vector<std::string_view> plssvm::detail::split	(	std::string_view	str,
		char	delim = ' '
	)

Split the string str at the positions with delimiter delim and return the sub-strings.

Parameters

[in]	str	the string to split
[in]	delim	the split delimiter

Returns

the split sub-strings ([[nodiscard]])

unreachable()

void plssvm::detail::unreachable ( )

inline

Invokes undefined behavior. Used to mark code paths that may never be reachable.

See: C++23 std::unreachable

get()

template<std::size_t I, typename... Types>

constexpr decltype(auto) plssvm::detail::get ( Types &&...  args )

constexprnoexcept

Get the I-th element of the parameter pack args at compile-time.

Template Parameters

I	the index of the element to get
Types	the types inside the parameter pack

Parameters

[in]

args

the values of the parameter pack

Returns

the I-th element of args ([[nodiscard]])

to_underlying() [1/2]

template<typename Enum >

constexpr std::underlying_type_t<Enum> plssvm::detail::to_underlying ( const Enum  e )

constexprnoexcept

Converts an enumeration to its underlying type.

Template Parameters

Enum	the enumeration type

Parameters

[in]

e

enumeration value to convert

Returns

the integer value of the underlying type of Enum, converted from e ([[nodiscard]])

to_underlying() [2/2]

template<typename Enum >

constexpr std::underlying_type_t<Enum> plssvm::detail::to_underlying ( const default_value< Enum > &  e )

constexprnoexcept

Converts an enumeration wrapped in a plssvm::default_value to its underlying type.

Template Parameters

Enum	the enumeration type

Parameters

[in]

e

enumeration value to convert wrapped in a plssvm::default_value

Returns

the integer value of the underlying type of Enum, converted from e ([[nodiscard]])

erase_if()

template<typename Container , typename Pred , PLSSVM_REQUIRES(is_container_v< Container >) >

Container::size_type plssvm::detail::erase_if ( Container &  c, Pred  pred  )

inline

Implements an erase_if function for different containers according to https://en.cppreference.com/w/cpp/container/map/erase_if.

Template Parameters

Container	the container type
Pred	the type of the unary-predicate

Parameters

[in,out]	c	the container to erase the elements that match the predicate pred from
[in]	pred	the predicate the to be erased elements must match

Returns

the number of erased elements

contains() [3/3]

template<typename Container , typename T , PLSSVM_REQUIRES(is_container_v< Container >) >

bool plssvm::detail::contains ( const Container &  c, const T &  val  )

inline

Check whether the Container c contains the value val.

Template Parameters

Container	the container type
T	the type of the value to check

Parameters

[in]	c	the container to check if it contains the value val
[in]	val	the value to check

Returns

true if the val exists in the container c, otherwise false ([[nodiscard]])

current_date_time()

std::string plssvm::detail::current_date_time

(

)

Return the current date time in the format "YYYY-MM-DD hh:mm:ss".

Returns

the current date time ([[nodiscard]])

get_value_from_named_parameter()

template<typename ExpectedType , typename IgorParser , typename NamedArgType >

ExpectedType plssvm::detail::get_value_from_named_parameter	(	const IgorParser &	parser,
		const NamedArgType &	named_arg
	)

Parse the value hold be named_arg and return it converted to the ExpectedType.

Template Parameters

ExpectedType	the type the value of the named argument should be converted to
IgorParser	the type of the named-parameter parser
NamedArgType	the type of the named-parameter (necessary since they are struct tags)

Parameters

[in]	parser	the named-parameter parser
[in]	named_arg	the named-parameter argument

Returns

the value of named_arg converted to ExpectedType ([[nodiscard]])

operator==()

template<typename T >

constexpr bool plssvm::detail::operator== ( const parameter< T > &  lhs, const parameter< T > &  rhs  )

constexprnoexcept

Compares the two parameter sets lhs and rhs for equality.

Two parameter sets are equal if and only if all SVM parameters are equal.

Template Parameters

T	the real_type

Parameters

[in]	lhs	the first parameter set
[in]	rhs	the second parameter set

Returns

true if both parameter sets are equal, false otherwise ([[nodiscard]])

operator!=()

template<typename T >

constexpr bool plssvm::detail::operator!= ( const parameter< T > &  lhs, const parameter< T > &  rhs  )

constexprnoexcept

Compares the two parameter sets lhs and rhs for inequality.

Two parameter sets are unequal if any of the SVM parameters are unequal.

Template Parameters

T	the real_type

Parameters

[in]	lhs	the first parameter set
[in]	rhs	the second parameter set

Returns

true if both parameter sets are unequal, false otherwise ([[nodiscard]])

equivalent()

template<typename T >

constexpr bool plssvm::detail::equivalent ( const parameter< T > &  lhs, const parameter< T > &  rhs  )

constexprnoexcept

Checks whether the two parameter sets lhs and rhs are equivalent.

Compares the member variables based on the kernel, i.e., for example for the rbf kernel both parameter sets must only have the same gamma and cost values but may differ in the values of degree or coef0. If all members should be compared regardless of the kernel type, one can use the operator== overload.

Parameters

[in]	lhs	the first parameter set
[in]	rhs	the second parameter set

Returns

true if both parameter sets are equivalent, false otherwise ([[nodiscard]])

operator<<() [4/4]

template<typename T >

std::ostream& plssvm::detail::operator<<	(	std::ostream &	out,
		const parameter< T > &	params
	)

Output all parameters encapsulated by params to the given output-stream out.

Template Parameters

T	the type of the data

Parameters

[in,out]	out	the output-stream to write the parameters to
[in]	params	the parameter set

Returns

the output-stream

Variable Documentation

is_tracking_entry_v

template<typename T >

constexpr bool plssvm::detail::is_tracking_entry_v = is_tracking_entry<T>::value

constexpr

A shorthand for plssvm::is_tracking_entry::value.