Flow123d  release_2.2.0-914-gf1a3a4f
Public Types | Public Member Functions | Static Public Member Functions | Static Public Attributes | Protected Member Functions | Protected Attributes | List of all members
FieldAlgorithmBase< spacedim, Value > Class Template Referenceabstract

#include <field_algo_base.hh>

Inheritance diagram for FieldAlgorithmBase< spacedim, Value >:
Inheritance graph
[legend]
Collaboration diagram for FieldAlgorithmBase< spacedim, Value >:
Collaboration graph
[legend]

Public Types

typedef Space< spacedim >::Point Point
 

Public Member Functions

 FieldAlgorithmBase (unsigned int n_comp=0)
 
virtual void init_from_input (const Input::Record &rec, const struct FieldAlgoBaseInitData &init_data)
 
virtual bool set_time (const TimeStep &time)
 
virtual void set_mesh (const Mesh *mesh, bool boundary_domain)
 
void set_component_idx (unsigned int idx)
 
unsigned int n_comp () const
 
FieldResult field_result () const
 
virtual double next_change_time ()
 
virtual Value::return_type const & value (const Point &p, const ElementAccessor< spacedim > &elm)=0
 
virtual void value_list (const std::vector< Point > &point_list, const ElementAccessor< spacedim > &elm, std::vector< typename Value::return_type > &value_list)=0
 
virtual ~FieldAlgorithmBase ()
 

Static Public Member Functions

static std::string template_name ()
 
static Input::Type::Abstractget_input_type ()
 
static const Input::Type::Instanceget_input_type_instance (Input::Type::Selection value_selection=Input::Type::Selection())
 
static const Input::Type::Recordget_field_algo_common_keys ()
 
static std::shared_ptr< FieldAlgorithmBase< spacedim, Value > > function_factory (const Input::AbstractRecord &rec, const struct FieldAlgoBaseInitData &init_data)
 

Static Public Attributes

static const unsigned int spacedim_ =spacedim
 
static constexpr bool is_enum_valued = boost::is_same<typename Value::element_type, FieldEnum>::value
 

Protected Member Functions

void init_unit_conversion_coefficient (const Input::Record &rec, const struct FieldAlgoBaseInitData &init_data)
 Init value of unit_conversion_coefficient_ from input. More...
 

Protected Attributes

TimeStep time_
 Actual time level; initial value is -infinity. More...
 
Value value_
 Last value, prevents passing large values (vectors) by value. More...
 
Value::return_type r_value_
 
FieldResult field_result_
 Indicator of particular values (zero, one) constant over space. More...
 
unsigned int component_idx_
 Specify if the field is part of a MultiField and which component it is. More...
 
double unit_conversion_coefficient_
 Coeficient of conversion of user-defined unit. More...
 

Detailed Description

template<int spacedim, class Value>
class FieldAlgorithmBase< spacedim, Value >

Base class for space-time function classes.

Definition at line 86 of file field_algo_base.hh.

Member Typedef Documentation

template<int spacedim, class Value>
typedef Space<spacedim>::Point FieldAlgorithmBase< spacedim, Value >::Point

Definition at line 89 of file field_algo_base.hh.

Constructor & Destructor Documentation

template<int spacedim, class Value >
FieldAlgorithmBase< spacedim, Value >::FieldAlgorithmBase ( unsigned int  n_comp = 0)

Kind of default constructor , with possible setting of the initial time. Fields that returns variable size vectors accepts number of components n_comp.

Definition at line 52 of file field_algo_base.impl.hh.

template<int spacedim, class Value>
virtual FieldAlgorithmBase< spacedim, Value >::~FieldAlgorithmBase ( )
inlinevirtual

Virtual destructor.

Definition at line 216 of file field_algo_base.hh.

Member Function Documentation

template<int spacedim, class Value>
FieldResult FieldAlgorithmBase< spacedim, Value >::field_result ( ) const
inline

Special field values spatially constant. Could allow optimization of tensor multiplication and tensor or vector addition. field_result_ should be set in constructor and in set_time method of particular Field implementation.

Definition at line 172 of file field_algo_base.hh.

template<int spacedim, class Value >
shared_ptr< FieldAlgorithmBase< spacedim, Value > > FieldAlgorithmBase< spacedim, Value >::function_factory ( const Input::AbstractRecord rec,
const struct FieldAlgoBaseInitData init_data 
)
static

This static method gets accessor to abstract record with function input, dispatch to correct constructor and initialize appropriate function object from the input. Returns shared pointer to FunctionBase<>.

Definition at line 129 of file field_algo_base.impl.hh.

Here is the caller graph for this function:

template<int spacedim, class Value >
const Input::Type::Record & FieldAlgorithmBase< spacedim, Value >::get_field_algo_common_keys ( )
static

Returns auxiliary record with keys common to all field algorithms.

Definition at line 96 of file field_algo_base.impl.hh.

template<int spacedim, class Value >
Input::Type::Abstract & FieldAlgorithmBase< spacedim, Value >::get_input_type ( )
static

Returns whole tree of input types for FieldBase with all descendants based on element input type (namely for FieldConstant) given by element_input_type pointer.

Definition at line 71 of file field_algo_base.impl.hh.

Here is the caller graph for this function:

template<int spacedim, class Value >
const Input::Type::Instance & FieldAlgorithmBase< spacedim, Value >::get_input_type_instance ( Input::Type::Selection  value_selection = Input::Type::Selection())
static

Returns parameterized whole tree of input types for FieldBase with all descendants based on element input type (namely for FieldConstant) given by element_input_type pointer.

Definition at line 83 of file field_algo_base.impl.hh.

Here is the caller graph for this function:

template<int spacedim, class Value >
void FieldAlgorithmBase< spacedim, Value >::init_from_input ( const Input::Record rec,
const struct FieldAlgoBaseInitData init_data 
)
virtual

Function can provide way to initialize itself from the input data.

TODO: make protected, should be called through function factory

Reimplemented in FieldFE< spacedim, Value >, FieldElementwise< spacedim, Value >, FieldConstant< spacedim, Value >, FieldFormula< spacedim, Value >, FieldPython< spacedim, Value >, FieldTimeFunction< spacedim, Value >, and FieldInterpolatedP0< spacedim, Value >.

Definition at line 140 of file field_algo_base.impl.hh.

Here is the caller graph for this function:

template<int spacedim, class Value >
void FieldAlgorithmBase< spacedim, Value >::init_unit_conversion_coefficient ( const Input::Record rec,
const struct FieldAlgoBaseInitData init_data 
)
protected

Init value of unit_conversion_coefficient_ from input.

Definition at line 184 of file field_algo_base.impl.hh.

Here is the caller graph for this function:

template<int spacedim, class Value >
unsigned int FieldAlgorithmBase< spacedim, Value >::n_comp ( ) const

Returns number of rows, i.e. number of components for variable size vectors. For values of fixed size returns zero.

Definition at line 162 of file field_algo_base.impl.hh.

Here is the caller graph for this function:

template<int spacedim, class Value>
virtual double FieldAlgorithmBase< spacedim, Value >::next_change_time ( )
inlinevirtual

Method for getting some information about next time where the function change its character. Used to add appropriate TimeMarks. TODO: think what kind of information we may need, is the next time value enough?

Definition at line 180 of file field_algo_base.hh.

template<int spacedim, class Value>
void FieldAlgorithmBase< spacedim, Value >::set_component_idx ( unsigned int  idx)
inline

Sets component_idx_

Definition at line 160 of file field_algo_base.hh.

template<int spacedim, class Value >
void FieldAlgorithmBase< spacedim, Value >::set_mesh ( const Mesh mesh,
bool  boundary_domain 
)
virtual

Is used only by some Field implementations, but can be used to check validity of incoming ElementAccessor in value methods.

Optional parameter boundary_domain can be used to specify, that the field will be evaluated only on the boundary part of the mesh. TODO: make separate mesh for the boundary, then we can drop this parameter.

Reimplemented in FieldFE< spacedim, Value >, FieldAddPotential< spacedim, Value >, and FieldElementwise< spacedim, Value >.

Definition at line 156 of file field_algo_base.impl.hh.

template<int spacedim, class Value >
bool FieldAlgorithmBase< spacedim, Value >::set_time ( const TimeStep time)
virtual

Set new time value. Some Fields may and some may not implement time dependent values and possibly various types of interpolation. There can not be unified approach to interpolation (at least not on this abstraction level) since some fields (FieldFormula, FieldPython) provides naturally time dependent functions other fields like (FieldConstant, ...), however, can be equipped by various time interpolation schemes. In future, we obviously need time interpolation of higher order so that we can use ODE integrators of higher order.

The method returns true if the value of the field has changed in the new time step.

Reimplemented in FieldFE< spacedim, Value >, FieldAddPotential< spacedim, Value >, FieldElementwise< spacedim, Value >, FieldFormula< spacedim, Value >, FieldTimeFunction< spacedim, Value >, and FieldInterpolatedP0< spacedim, Value >.

Definition at line 148 of file field_algo_base.impl.hh.

template<int spacedim, class Value >
string FieldAlgorithmBase< spacedim, Value >::template_name ( )
static

Returns template parameters as string in order to distinguish name of Abstracts for initialization of different instances of the FieldBase template.

Definition at line 64 of file field_algo_base.impl.hh.

Here is the caller graph for this function:

template<int spacedim, class Value>
virtual Value::return_type const& FieldAlgorithmBase< spacedim, Value >::value ( const Point p,
const ElementAccessor< spacedim > &  elm 
)
pure virtual

Returns one value in one given point on an element given by ElementAccessor elm. It returns reference to he actual value in order to avoid temporaries for vector and tensor values.

This method just call the later one value(Point, ElementAccessor, Value) and drops the FieldResult.

Usual implementation of this method fills member r_value_ through unified envelope value_ as general tensor and then returns member r_value_. However, some particular Fields may have result stored elsewhere, in such a case the reference to the result can be returned directly without using the member value_. Keeping such wide space for optimization has drawback in slow generic implementation of the value_list method that fills whole vector of values for vector of points. Its generic implementation has to copy all values instead of directly store them into the vector of result values.

So the best practice when implementing value and methods in particular FieldBase descendant is implement some thing like value(point, elm, Value::return_type &value) and using s having in part

Implemented in FieldElementwise< spacedim, Value >, FieldAddPotential< spacedim, Value >, FieldFE< spacedim, Value >, FieldPython< spacedim, Value >, FieldConstant< spacedim, Value >, FieldFormula< spacedim, Value >, and FieldInterpolatedP0< spacedim, Value >.

Here is the caller graph for this function:

template<int spacedim, class Value>
void FieldAlgorithmBase< spacedim, Value >::value_list ( const std::vector< Point > &  point_list,
const ElementAccessor< spacedim > &  elm,
std::vector< typename Value::return_type > &  value_list 
)
pure virtual

Returns std::vector of scalar values in several points at once. The base class implements trivial implementation using the value(,,) method. This is not optimal as it involves lot of virtual calls, but this overhead can be negligible for more complex fields as Python of Formula.

FieldAlgorithmBase provides a slow implementation using the value() method. Derived Field can implement its value_list method as call of FieldAlgoritmBase<...>::value_list().

Implemented in FieldElementwise< spacedim, Value >, FieldAddPotential< spacedim, Value >, FieldFE< spacedim, Value >, FieldPython< spacedim, Value >, FieldConstant< spacedim, Value >, FieldFormula< spacedim, Value >, and FieldInterpolatedP0< spacedim, Value >.

Definition at line 169 of file field_algo_base.impl.hh.

Here is the caller graph for this function:

Member Data Documentation

template<int spacedim, class Value>
unsigned int FieldAlgorithmBase< spacedim, Value >::component_idx_
protected

Specify if the field is part of a MultiField and which component it is.

Definition at line 230 of file field_algo_base.hh.

template<int spacedim, class Value>
FieldResult FieldAlgorithmBase< spacedim, Value >::field_result_
protected

Indicator of particular values (zero, one) constant over space.

Definition at line 228 of file field_algo_base.hh.

template<int spacedim, class Value>
constexpr bool FieldAlgorithmBase< spacedim, Value >::is_enum_valued = boost::is_same<typename Value::element_type, FieldEnum>::value
static

Definition at line 91 of file field_algo_base.hh.

template<int spacedim, class Value>
Value::return_type FieldAlgorithmBase< spacedim, Value >::r_value_
protected

Definition at line 226 of file field_algo_base.hh.

template<int spacedim, class Value>
const unsigned int FieldAlgorithmBase< spacedim, Value >::spacedim_ =spacedim
static

Definition at line 90 of file field_algo_base.hh.

template<int spacedim, class Value>
TimeStep FieldAlgorithmBase< spacedim, Value >::time_
protected

Actual time level; initial value is -infinity.

Definition at line 223 of file field_algo_base.hh.

template<int spacedim, class Value>
double FieldAlgorithmBase< spacedim, Value >::unit_conversion_coefficient_
protected

Coeficient of conversion of user-defined unit.

Definition at line 232 of file field_algo_base.hh.

template<int spacedim, class Value>
Value FieldAlgorithmBase< spacedim, Value >::value_
protected

Last value, prevents passing large values (vectors) by value.

Definition at line 225 of file field_algo_base.hh.


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