Reference documentation for deal.II version 9.6.0
|
#include <deal.II/base/patterns.h>
Public Types | |
enum | OutputStyle { Machine , Text , LaTeX } |
Public Member Functions | |
Double (const double lower_bound=min_double_value, const double upper_bound=max_double_value) | |
virtual bool | match (const std::string &test_string) const override |
virtual std::string | description (const OutputStyle style=Machine) const override |
virtual std::unique_ptr< PatternBase > | clone () const override |
virtual std::size_t | memory_consumption () const |
Static Public Member Functions | |
static std::unique_ptr< Double > | create (const std::string &description) |
Static Public Attributes | |
static const double | min_double_value = std::numeric_limits<double>::lowest() |
static const double | max_double_value = std::numeric_limits<double>::max() |
Private Attributes | |
const double | lower_bound |
const double | upper_bound |
Static Private Attributes | |
static const char * | description_init = "[Double" |
Test for the string being a double
. If bounds are given to the constructor, then the integer given also needs to be within the interval specified by these bounds. Note that unlike common convention in the C++ standard library, both bounds of this interval are inclusive; the reason is that in practice in most cases, one needs closed intervals, but these can only be realized with inclusive bounds for non-integer values. We thus stay consistent by always using closed intervals.
If the upper bound given to the constructor is smaller than the lower bound, then every double precision number is allowed.
Giving bounds may be useful if for example a value can only be positive and less than a reasonable upper bound (for example damping parameters are frequently only reasonable if between zero and one), or in many other cases.
Definition at line 290 of file patterns.h.
|
inherited |
List of possible description output formats.
Capitalization chosen for similarity to ParameterHandler::OutputStyle.
Definition at line 96 of file patterns.h.
Patterns::Double::Double | ( | const double | lower_bound = min_double_value, |
const double | upper_bound = max_double_value ) |
Constructor. Bounds can be specified within which a valid parameter has to be. If the upper bound is smaller than the lower bound, then the entire set of double precision numbers is implied. The default values are chosen such that no bounds are enforced on parameters.
Definition at line 349 of file patterns.cc.
|
overridevirtual |
Return true
if the string is a number and its value is within the specified range.
Implements Patterns::PatternBase.
Definition at line 357 of file patterns.cc.
|
overridevirtual |
Return a description of the pattern that valid strings are expected to match. If bounds were specified to the constructor, then include them into this description.
Implements Patterns::PatternBase.
Definition at line 380 of file patterns.cc.
|
overridevirtual |
Return a copy of the present object, which is newly allocated on the heap. Ownership of that object is transferred to the caller of this function.
Implements Patterns::PatternBase.
Definition at line 480 of file patterns.cc.
|
static |
Creates a new object on the heap using new
if the given description
is a valid format (for example created by calling description() on an existing object), or nullptr
otherwise. Ownership of the returned object is transferred to the caller of this function, which should be freed using delete
.
Definition at line 488 of file patterns.cc.
|
virtualinherited |
Determine an estimate for the memory consumption (in bytes) of this object. To avoid unnecessary overhead, we do not force derived classes to provide this function as a virtual overloaded one, but rather try to cast the present object to one of the known derived classes and if that fails then take the size of this base class instead and add 32 byte (this value is arbitrary, it should account for virtual function tables, and some possible data elements). Since there are usually not many thousands of objects of this type around, and since the memory_consumption mechanism is used to find out where memory in the range of many megabytes is, this seems like a reasonable approximation.
On the other hand, if you know that your class deviates from this assumption significantly, you can still overload this function.
Reimplemented in Patterns::List, Patterns::Map, Patterns::MultipleSelection, Patterns::Selection, and Patterns::Tuple.
Definition at line 189 of file patterns.cc.
|
static |
Minimal double value used as default value, taken from std::numeric_limits
.
Definition at line 297 of file patterns.h.
|
static |
Maximal double value used as default value, taken from std::numeric_limits
.
Definition at line 303 of file patterns.h.
|
private |
Value of the lower bound. A number that satisfies the match operation of this class must be equal to this value or larger, if the bounds of the interval form a valid range.
Definition at line 355 of file patterns.h.
|
private |
Value of the upper bound. A number that satisfies the match operation of this class must be equal to this value or less, if the bounds of the interval form a valid range.
Definition at line 363 of file patterns.h.
|
staticprivate |
Initial part of description
Definition at line 368 of file patterns.h.