Base class for Attributes. More...

#include <Attribute.h>

Inheritance diagram for ioda::detail::Attribute_Base< Attribute_Implementation >:

[legend]

Public Member Functions
Writing Data
Note Writing metadata is an all-or-nothing-process, unlike writing segments of data to a variable. Dimensions are fixed. Attribute are not resizable.
virtual Attribute_Implementation	write (gsl::span< char > data, const Type &type)
	The fundamental write function. Backends overload this function to implement all write operations. More...

template<class DataType , class Marshaller = ioda::Object_Accessor<DataType>, class TypeWrapper = Types::GetType_Wrapper<DataType>>
Attribute_Implementation	write (gsl::span< const DataType > data)
	Write data. More...

template<class DataType , class Marshaller = ioda::Object_Accessor<DataType>, class TypeWrapper = Types::GetType_Wrapper<DataType>>
Attribute_Implementation	write (const std::vector< DataType > &data)
	Write data. More...

template<class DataType >
Attribute_Implementation	write (std::initializer_list< DataType > data)
	Write data. More...

template<class DataType >
Attribute_Implementation	write (DataType data)
	Write a datum. More...

template<class EigenClass >
Attribute_Implementation	writeWithEigenRegular (const EigenClass &d)
	Write an Eigen object (a Matrix, an Array, a Block, a Map). More...

template<class EigenClass >
Attribute_Implementation	writeWithEigenTensor (const EigenClass &d)
	Write an Eigen Tensor-like object. More...

Reading Data
virtual Attribute_Implementation	read (gsl::span< char > data, const Type &in_memory_dataType) const
	The fundamental read function. Backends overload this function to implement all read operations. More...

template<class DataType , class Marshaller = ioda::Object_Accessor<DataType>, class TypeWrapper = Types::GetType_Wrapper<DataType>>
Attribute_Implementation	read (gsl::span< DataType > data) const
	Read data. More...

template<class DataType >
Attribute_Implementation	read (std::vector< DataType > &data) const
	Vector read convenience function. More...

template<class DataType >
Attribute_Implementation	read (std::valarray< DataType > &data) const
	Valarray read convenience function. More...

template<class DataType >
Attribute_Implementation	read (DataType &data) const
	Read into a single value (convenience function). More...

template<class DataType >
DataType	read () const
	Read a single value (convenience function). More...

template<class DataType >
std::vector< DataType >	readAsVector () const
	Read into a new vector. Python convenience function. More...

template<class EigenClass , bool Resize = detail::EigenCompat::CanResize<EigenClass>::value>
Attribute_Implementation	readWithEigenRegular (EigenClass &res) const
	Read data into an Eigen::Array, Eigen::Matrix, Eigen::Map, etc. More...

template<class EigenClass >
Attribute_Implementation	readWithEigenTensor (EigenClass &res) const
	Read data into an Eigen::Array, Eigen::Matrix, Eigen::Map, etc. More...

template<class EigenClass >
EigenClass	_readWithEigenRegular_python () const

Type-querying Functions
virtual Type	getType () const
	Get Attribute type. More...

Type	type () const
	Get Attribute type. More...

virtual detail::Type_Provider *	getTypeProvider () const
	Query the backend and get the type provider. More...

template<class DataType >
bool	isA () const
	Convenience function to check an Attribute's storage type. More...

virtual bool	isA (Type lhs) const
	Hand-off to the backend to check equivalence. More...

bool	isA (BasicTypes dataType)
	Python compatability function. More...

bool	_py_isA2 (BasicTypes dataType)

Data Space-Querying Functions
virtual Dimensions	getDimensions () const
	Get Attribute's dimensions. More...

Protected Attributes
std::shared_ptr< Attribute_Backend >	backend_
	Using an opaque object to implement the backend. More...

Friends
class	Variable_Backend

General Functions
	Attribute_Base (std::shared_ptr< Attribute_Backend >)

virtual	~Attribute_Base ()

Detailed Description

template<class Attribute_Implementation = Attribute>
class ioda::detail::Attribute_Base< Attribute_Implementation >

Base class for Attributes.

You might wonder why we have this class as a template. This is because we are using a bit of compile-time template polymorphism to return Attribute objects from base classes before Attribute is fully declared. This is a variation of the (Curiously Recurring Template Pattern)[https://en.wikipedia.org/wiki/Curiously_recurring_template_pattern].

Definition at line 50 of file Attribute.h.

Constructor & Destructor Documentation

◆ Attribute_Base()

ioda::detail::Attribute_Base::Attribute_Base ( std::shared_ptr< Attribute_Backend > hnd_attr )

protected

Definition at line 17 of file Attribute.cpp.

◆ ~Attribute_Base()

ioda::detail::Attribute_Base::~Attribute_Base ( )

virtualdefault

Member Function Documentation

◆ _py_isA2()

template<class Attribute_Implementation = Attribute>

bool ioda::detail::Attribute_Base< Attribute_Implementation >::_py_isA2 ( BasicTypes dataType )

inline

Definition at line 458 of file Attribute.h.

Here is the call graph for this function:

Here is the caller graph for this function:

◆ _readWithEigenRegular_python()

template<class Attribute_Implementation = Attribute>

template<class EigenClass >

EigenClass ioda::detail::Attribute_Base< Attribute_Implementation >::_readWithEigenRegular_python ( ) const

inline

Definition at line 423 of file Attribute.h.

Here is the call graph for this function:

◆ getDimensions()

Dimensions ioda::detail::Attribute_Base::getDimensions

virtual

Get Attribute's dimensions.

Reimplemented in ioda::Engines::ObsStore::ObsStore_Attribute_Backend, and ioda::detail::Engines::HH::HH_Attribute.

Definition at line 35 of file Attribute.cpp.

Here is the caller graph for this function:

◆ getType()

Type ioda::detail::Attribute_Base::getType

virtual

Get Attribute type.

Reimplemented in ioda::Engines::ObsStore::ObsStore_Attribute_Backend, and ioda::detail::Engines::HH::HH_Attribute.

Definition at line 25 of file Attribute.cpp.

Here is the caller graph for this function:

◆ getTypeProvider()

detail::Type_Provider * ioda::detail::Attribute_Base::getTypeProvider

virtual

Query the backend and get the type provider.

Reimplemented in ioda::Engines::ObsStore::ObsStore_Attribute_Backend, and ioda::detail::Engines::HH::HH_Attribute.

Definition at line 55 of file Attribute.cpp.

Here is the caller graph for this function:

◆ isA() [1/3]

template<class Attribute_Implementation = Attribute>

template<class DataType >

bool ioda::detail::Attribute_Base< Attribute_Implementation >::isA ( ) const

inline

Convenience function to check an Attribute's storage type.

Template Parameters

DataType is the type of the data. I.e. float, int, int32_t, uint16_t, std::string, etc.

Returns: True if the type matches; False (0) if the type does not match

Exceptions

jedi::xError if an error occurred.

Definition at line 447 of file Attribute.h.

Here is the call graph for this function:

Here is the caller graph for this function:

◆ isA() [2/3]

template<class Attribute_Implementation = Attribute>

bool ioda::detail::Attribute_Base< Attribute_Implementation >::isA ( BasicTypes dataType )

inline

Python compatability function.

Definition at line 456 of file Attribute.h.

Here is the call graph for this function:

Here is the caller graph for this function:

◆ isA() [3/3]

bool ioda::detail::Attribute_Base::isA ( Type lhs ) const

virtual

Hand-off to the backend to check equivalence.

Reimplemented in ioda::Engines::ObsStore::ObsStore_Attribute_Backend, and ioda::detail::Engines::HH::HH_Attribute.

Definition at line 45 of file Attribute.cpp.

◆ read() [1/6]

template<class Attribute_Implementation = Attribute>

template<class DataType >

DataType ioda::detail::Attribute_Base< Attribute_Implementation >::read ( ) const

inline

Read a single value (convenience function).

Template Parameters

DataType is the type of the data. I.e. float, int, int32_t, uint16_t, std::string, etc.

Returns: A datum of type DataType.

Exceptions

jedi::xError if the underlying data have size greater than 1.

Note: The Python function is read_datum_*

Definition at line 319 of file Attribute.h.

Here is the caller graph for this function:

◆ read() [2/6]

template<class Attribute_Implementation = Attribute>

template<class DataType >

Attribute_Implementation ioda::detail::Attribute_Base< Attribute_Implementation >::read ( DataType & data ) const

inline

Read into a single value (convenience function).

Template Parameters

DataType is the type of the data. I.e. float, int, int32_t, uint16_t, std::string, etc.

Parameters

data	is where the datum is read to.

Returns: Another instance of this Attribute. Used for operation chaining.

Exceptions

jedi::xError if the underlying data have multiple elements.

Definition at line 303 of file Attribute.h.

Here is the call graph for this function:

◆ read() [3/6]

Attribute ioda::detail::Attribute_Base::read	(	gsl::span< char >	data,
		const Type &	in_memory_dataType
	)		const

virtual

The fundamental read function. Backends overload this function to implement all read operations.

This function reads in a span of characters from the backend attribute storage. No type conversions take place here (see the templated conversion function, below).

Parameters

data	is a span of data that has length of getStorageSize().
in_memory_datatype	is an opaque (backend-level) object that describes the placement of the data in memory. Usually ignorable - needed for complex data structures.

Exceptions

jedi::xError if data has the wrong size.

Returns: The attribute (for chaining).

Reimplemented in ioda::Engines::ObsStore::ObsStore_Attribute_Backend, and ioda::detail::Engines::HH::HH_Attribute.

Definition at line 75 of file Attribute.cpp.

Here is the caller graph for this function:

◆ read() [4/6]

template<class Attribute_Implementation = Attribute>

template<class DataType , class Marshaller = ioda::Object_Accessor<DataType>, class TypeWrapper = Types::GetType_Wrapper<DataType>>

Attribute_Implementation ioda::detail::Attribute_Base< Attribute_Implementation >::read ( gsl::span< DataType > data ) const

inline

Read data.

This is a fundamental function that reads a span of characters from backend storage, and then performs the appropriate type conversion / deserialization into objects in data.

Template Parameters

DataType	is the type if the data to be read. I.e. float, int, int32_t, uint16_t, std::string, etc.
Marshaller	is the class that performs the deserialization operation.
TypeWrapper	is a helper class that passes Type information to the backend.

Parameters

data	is a pointer-size pair to the data buffer that is filled with the metadata's contents. It should be pre-sized to accomodate all of the matadata. See getDimensions().numElements. data will be filled in row-major order.
in_memory_datatype	is an opaque (backend-level) object that describes the placement of the data in memory. Usually this does not need to be set to anything other than its default value. Kept as a parameter for debugging purposes.

Returns: Another instance of this Attribute. Used for operation chaining.

Exceptions

jedi::xError if data.size() != getDimensions().numElements.

See also: getDimensions for buffer size information.

Definition at line 251 of file Attribute.h.

Here is the call graph for this function:

◆ read() [5/6]

template<class Attribute_Implementation = Attribute>

template<class DataType >

Attribute_Implementation ioda::detail::Attribute_Base< Attribute_Implementation >::read ( std::valarray< DataType > & data ) const

inline

Valarray read convenience function.

Template Parameters

DataType is the type of the data. I.e. float, int, int32_t, uint16_t, std::string, etc.

Parameters

data	is a valarray acting as a data buffer that is filled with the metadata's contents. It gets resized as needed.

Returns: Another instance of this Attribute. Used for operation chaining.

Note: data will be stored in row-major order.

Definition at line 292 of file Attribute.h.

Here is the call graph for this function:

◆ read() [6/6]

template<class Attribute_Implementation = Attribute>

template<class DataType >

Attribute_Implementation ioda::detail::Attribute_Base< Attribute_Implementation >::read ( std::vector< DataType > & data ) const

inline

Vector read convenience function.

Template Parameters

DataType is the type of the data. I.e. float, int, int32_t, uint16_t, std::string, etc.

Parameters

data	is a vector acting as a data buffer that is filled with the metadata's contents. It gets resized as needed.

Returns: Another instance of this Attribute. Used for operation chaining.

Note: data will be stored in row-major order.

Definition at line 280 of file Attribute.h.

Here is the call graph for this function:

◆ readAsVector()

template<class Attribute_Implementation = Attribute>

template<class DataType >

std::vector<DataType> ioda::detail::Attribute_Base< Attribute_Implementation >::readAsVector ( ) const

inline

Read into a new vector. Python convenience function.

Template Parameters

DataType is the type of the data.

Note: The Python function is read_list_*

Definition at line 329 of file Attribute.h.

Here is the call graph for this function:

◆ readWithEigenRegular()

template<class Attribute_Implementation = Attribute>

template<class EigenClass , bool Resize = detail::EigenCompat::CanResize<EigenClass>::value>

Attribute_Implementation ioda::detail::Attribute_Base< Attribute_Implementation >::readWithEigenRegular ( EigenClass & res ) const

inline

Read data into an Eigen::Array, Eigen::Matrix, Eigen::Map, etc.

Template Parameters

EigenClass	is a template pointing to the Eigen object. This template must provide the EigenClass::Scalar typedef.
Resize	indicates whether the Eigen object should be resized if there is a dimension mismatch. Not all Eigen objects can be resized.

Parameters

res	is the Eigen object.

Returns: Another instance of this Attribute. Used for operation chaining.

Exceptions

jedi::xError	if the attribute's dimensionality is too high.
jedi::xError	if resize = false and there is a dimension mismatch.

Note: When reading in a 1-D object, the data are read as a column vector.

Definition at line 347 of file Attribute.h.

Here is the call graph for this function:

Here is the caller graph for this function:

◆ readWithEigenTensor()

template<class Attribute_Implementation = Attribute>

template<class EigenClass >

Attribute_Implementation ioda::detail::Attribute_Base< Attribute_Implementation >::readWithEigenTensor ( EigenClass & res ) const

inline

Read data into an Eigen::Array, Eigen::Matrix, Eigen::Map, etc.

Template Parameters

EigenClass is a template pointing to the Eigen object. This template must provide the EigenClass::Scalar typedef.

Parameters

res	is the Eigen object.

Returns: Another instance of this Attribute. Used for operation chaining.

Exceptions

jedi::xError if there is a size mismatch.

Note: When reading in a 1-D object, the data are read as a column vector.

Definition at line 401 of file Attribute.h.

Here is the call graph for this function:

Here is the caller graph for this function:

◆ type()

template<class Attribute_Implementation = Attribute>

Type ioda::detail::Attribute_Base< Attribute_Implementation >::type ( ) const

inline

Get Attribute type.

Definition at line 436 of file Attribute.h.

Here is the call graph for this function:

◆ write() [1/5]

template<class Attribute_Implementation = Attribute>

template<class DataType , class Marshaller = ioda::Object_Accessor<DataType>, class TypeWrapper = Types::GetType_Wrapper<DataType>>

Attribute_Implementation ioda::detail::Attribute_Base< Attribute_Implementation >::write ( const std::vector< DataType > & data )

inline

Write data.

Note: Normally the gsl::span write is fine. This one exists for easy Python binding.

Template Parameters

DataType	is the type of the data. I.e. float, int, int32_t, uint16_t, std::string, etc.
Marshaller	is the class that serializes the data type into something that the backend library can use.
TypeWrapper	is a helper class that creates Type objects for the backend.

Parameters

data	is a std::vector that contains the data to be written.
in_memory_dataType	is the memory layout needed to parse data's type.

Returns: Another instance of this Attribute. Used for operation chaining.

Exceptions

jedi::xError if data.size() does not match getDimensions().numElements.

See also: gsl::span for details of how to make a span.; gsl::make_span for details on how to make a span.

Definition at line 129 of file Attribute.h.

Here is the caller graph for this function:

◆ write() [2/5]

template<class Attribute_Implementation = Attribute>

template<class DataType >

Attribute_Implementation ioda::detail::Attribute_Base< Attribute_Implementation >::write ( DataType data )

inline

Write a datum.

Template Parameters

DataType is the type of the data. I.e. float, int, int32_t, uint16_t, std::string, etc.

Parameters

data	is the data to be written.

Returns: Another instance of this Attribute. Used for operation chaining.

Exceptions

jedi::xError if the Attribute dimensions are larger than a single point.

Definition at line 153 of file Attribute.h.

Here is the call graph for this function:

Here is the caller graph for this function:

◆ write() [3/5]

Attribute ioda::detail::Attribute_Base::write	(	gsl::span< char >	data,
		const Type &	type
	)

virtual

The fundamental write function. Backends overload this function to implement all write operations.

This function writes a span of bytes (characters) to the backend attribute storage. No type conversions take place here (see the templated conversion function, below).

Parameters

data	is a span of data.
in_memory_datatype	is an opaque (backend-level) object that describes the placement of the data in memory. Usually ignorable - needed for complex data structures.

Exceptions

jedi::xError if data has the wrong size.

Returns: The attribute (for chaining).

Reimplemented in ioda::Engines::ObsStore::ObsStore_Attribute_Backend, and ioda::detail::Engines::HH::HH_Attribute.

Definition at line 65 of file Attribute.cpp.

Here is the caller graph for this function:

◆ write() [4/5]

template<class Attribute_Implementation = Attribute>

template<class DataType , class Marshaller = ioda::Object_Accessor<DataType>, class TypeWrapper = Types::GetType_Wrapper<DataType>>

Attribute_Implementation ioda::detail::Attribute_Base< Attribute_Implementation >::write ( gsl::span< const DataType > data )

inline

Write data.

Template Parameters

DataType	is the type of the data. I.e. float, int, int32_t, uint16_t, std::string, etc.
Marshaller	is the class that serializes the data type into something that the backend library can use.
TypeWrapper	is a helper class that creates Type objects for the backend.

Parameters

data	is a gsl::span (a pointer-length pair) that contains the data to be written.
in_memory_dataType	is the memory layout needed to parse data's type.

Returns: Another instance of this Attribute. Used for operation chaining.

Exceptions

jedi::xError if data.size() does not match getDimensions().numElements.

See also: gsl::span for details of how to make a span.; gsl::make_span

Definition at line 101 of file Attribute.h.

Here is the call graph for this function:

Here is the caller graph for this function:

◆ write() [5/5]

template<class Attribute_Implementation = Attribute>

template<class DataType >

Attribute_Implementation ioda::detail::Attribute_Base< Attribute_Implementation >::write ( std::initializer_list< DataType > data )

inline

Write data.

Template Parameters

DataType is the type of the data. I.e. float, int, int32_t, uint16_t, std::string, etc.

Parameters

data	is an initializer list that contains the data to be written.
in_memory_dataType	is the memory layout needed to parse data's type.

Returns: Another instance of this Attribute. Used for operation chaining.

Exceptions

jedi::xError if data.size() does not match getDimensions().numElements.

See also: gsl::span for details of how to make a span.

Definition at line 142 of file Attribute.h.

Here is the caller graph for this function:

◆ writeWithEigenRegular()

template<class Attribute_Implementation = Attribute>

template<class EigenClass >

Attribute_Implementation ioda::detail::Attribute_Base< Attribute_Implementation >::writeWithEigenRegular ( const EigenClass & d )

inline

Write an Eigen object (a Matrix, an Array, a Block, a Map).

Template Parameters

EigenClass is the Eigen object to write.

Parameters

d	is the data to be written.

Exceptions

jedi::xError on a dimension mismatch.

Returns: the attribute

Definition at line 169 of file Attribute.h.

◆ writeWithEigenTensor()

template<class Attribute_Implementation = Attribute>

template<class EigenClass >

Attribute_Implementation ioda::detail::Attribute_Base< Attribute_Implementation >::writeWithEigenTensor ( const EigenClass & d )

inline

Write an Eigen Tensor-like object.

Template Parameters

EigenClass is the Eigen tensor to write.

Parameters

d	is the data to be written.

Exceptions

jedi::xError on a dimension mismatch.

Returns: the attribute

Definition at line 195 of file Attribute.h.

Here is the call graph for this function:

Here is the caller graph for this function:

Friends And Related Function Documentation

◆ Variable_Backend

template<class Attribute_Implementation = Attribute>

friend class Variable_Backend

friend

Definition at line 56 of file Attribute.h.

Member Data Documentation

◆ backend_

template<class Attribute_Implementation = Attribute>

std::shared_ptr<Attribute_Backend> ioda::detail::Attribute_Base< Attribute_Implementation >::backend_

protected

Using an opaque object to implement the backend.

Definition at line 53 of file Attribute.h.

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

Public Member Functions

Protected Attributes

Friends

General Functions

Detailed Description

template<class Attribute_Implementation = Attribute> class ioda::detail::Attribute_Base< Attribute_Implementation >

Constructor & Destructor Documentation

◆ Attribute_Base()

◆ ~Attribute_Base()

Member Function Documentation

◆ _py_isA2()

◆ _readWithEigenRegular_python()

◆ getDimensions()

◆ getType()

◆ getTypeProvider()

◆ isA() [1/3]

◆ isA() [2/3]

◆ isA() [3/3]

◆ read() [1/6]

◆ read() [2/6]

◆ read() [3/6]

◆ read() [4/6]

◆ read() [5/6]

◆ read() [6/6]

◆ readAsVector()

◆ readWithEigenRegular()

◆ readWithEigenTensor()

◆ type()

◆ write() [1/5]

◆ write() [2/5]

◆ write() [3/5]

◆ write() [4/5]

◆ write() [5/5]

◆ writeWithEigenRegular()

◆ writeWithEigenTensor()

Friends And Related Function Documentation

◆ Variable_Backend

Member Data Documentation

◆ backend_

template<class Attribute_Implementation = Attribute>
class ioda::detail::Attribute_Base< Attribute_Implementation >