FunctionTraits.h

Go to the documentation of this file.

// # FunctionTraits.h: Function data types for parameters and arguments
// # Copyright (C) 2001,2002
// # Associated Universities, Inc. Washington DC, USA.
// #
// # This library is free software; you can redistribute it and/or modify it
// # under the terms of the GNU Library General Public License as published by
// # the Free Software Foundation; either version 2 of the License, or (at your
// # option) any later version.
// #
// # This library is distributed in the hope that it will be useful, but WITHOUT
// # ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
// # FITNESS FOR A PARTICULAR PURPOSE.  See the GNU Library General Public
// # License for more details.
// #
// # You should have received a copy of the GNU Library General Public License
// # along with this library; if not, write to the Free Software Foundation,
// # Inc., 675 Massachusetts Ave, Cambridge, MA 02139, USA.
// #
// # Correspondence concerning AIPS++ should be addressed as follows:
// #        Internet email: casa-feedback@nrao.edu.
// #        Postal address: AIPS++ Project Office
// #                        National Radio Astronomy Observatory
// #                        520 Edgemont Road
// #                        Charlottesville, VA 22903-2475 USA
 
#ifndef SCIMATH_FUNCTIONTRAITS_H
#define SCIMATH_FUNCTIONTRAITS_H
 
// # Includes
#include <casacore/casa/aips.h>
#include <casacore/scimath/Mathematics/AutoDiff.h>
#include <casacore/scimath/Mathematics/AutoDiffA.h>
#include <casacore/scimath/Mathematics/AutoDiffX.h>
 
namespace casacore {  // # NAMESPACE CASACORE - BEGIN
 
//
// <summary> Function data types for parameters and arguments
// </summary>
// <use visibility=local>
//
// <reviewed reviewer="UNKNOWN" date="before2004/08/25" tests="t">
// </reviewed>
 
// <prerequisite>
//   <li> <linkto class="Function">Function</linkto>
//   <li> <linkto class="AutoDiff">AutoDiff</linkto>
// </prerequisite>
 
// <etymology>
// A trait is a characteristic feature. FunctionTraits defines relationships
// between the data types of result, parameters and arguments of function
// objects.
// </etymology>
//
// <synopsis>
// This templated class contains a number of typedefs that
// describe the relationship between different numeric data types used for
// the calculation of <src>Function</src> and the function parameter and
// arguments.
//
// Its main use is to optimize the speed of the calculation of function
// values and derivatives in the case of <src>AutoDiff</src> use and
// manual calculation of derivatives, by allowing the data type of the
// the arguments and/or parameters to be plain numeric in cases where the
// derivatives wrt these are not needed.
// To enable this, the following definitions are used for the use of the
// Function template. Bear in mind that the Function operator is defined
// as <src>result = f<T>(x; parameters)</src>.
// <ol>
//  <li> Simple numeric type (Double, Complex, etc): result, parameters and
//  arguments: all the one defined templated type.
//  <li> <src>AutoDiff<T></src> indicates the calculation (either automatic or
//  with specialized implementations) of the result with a <src>T</src>
//  function value for <src>T</src> arg, and <src>AutoDiff<T></src>
//  parameters.
//   <li> <src>AutoDiffA</src> calculate form <src>AutoDiff<T></src>
//  arguments and parameters (note that either could be simple
//  values with zero derivatives)
//   <li> <src>AutoDiffX<T></src> : calculate only with respect to
//  the arguments the derivatives, by using <src>T</src>
//  parameters
// </ol>
// The following types are defined:
// <dl>
// <dt> <src>Type</src>
// <dd> The template argument
// <dt> <src>BaseType</src>
// <dd> One down in the template hierarchy if possible (e.g. <src>Double</src>
//  for <src>AutoDiff<Double></src>)
// <dt> <src>NumericType</src>
// <dd> Ultimate numeric type (e.g. <src>Double</src> for
//  <src>AutoDiff<AutoDiff<Double> ></src>
// <dt> <src>ParamType</src>
// <dd> Type used for parameters
// <dt> <src>ArgType</src>
// <dd> Type used for arguments
// <dt> <src>DiffType</src>
// <dd> The default differentiation type (e.g. <src>AutoDiff<Double></src>
//      for <src>AutoDiff<Double></src>)
// <dt> <src>getValue()</src>
// <dd>     get the value of a simple numeric or of an <src>AutoDiff</src>
// <dt> <src>setValue()</src>
// <dd>     set the value of a simple numeric or of an <src>AutoDiff</src>
// </dl>
//
// The specializations are done in such a way that higher order
// derivatives (e.g. <src>AutoDiff<AutoDiff<Double> ></src>) are catered for.
//
// Note that the class names in the following definitions are extended with
// some individual id (like <src>_PA</src>): do not use them in programming,
// they are only necessary for the <src>cxx2html</src> interpreter)
//
// This class is implemented as a number of specializations for the
// following data types.
// <ul>
// <li> <src>T</src>
// <li> <src>AutoDiff<T></src>
// <li> <src>AutoDiffA<T></src>
// <li> <src>AutoDiffX<T></src>
// </ul>
// </synopsis>
//
// <example>
// See the <linkto class=Function>Function</linkto> class code.
// </example>
//
// <motivation>
// To keep the Function class single templated
// </motivation>
//
// <todo asof="2002/06/19">
//  <li> Additional <src>AutoDiff*</src> classes if and when needed
// </todo>
//
 
template <class T>
class FunctionTraits {
 public:
  // Actual template type
  typedef T Type;
  // Template base type
  typedef T BaseType;
  // Numeric type of template
  typedef T NumericType;
  // Type for parameters
  typedef T ParamType;
  // Type for arguments
  typedef T ArgType;
  // Default type for differentiation
  typedef AutoDiff<T> DiffType;
  // Get the value
  static const T &getValue(const T &in) { return in; }
  // Set a value (and possible derivative)
  static void setValue(T &out, const T &val, const uInt, const uInt) { out = val; }
};
 
// # Following are specializations. Naming only for documentation
// # purposes (a problem with cxx2html)
 
#define FunctionTraits_P FunctionTraits
 
// <summary> FunctionTraits specialization for AutoDiff
// </summary>
 
template <class T>
class FunctionTraits_P<AutoDiff<T>> {
 public:
  // Actual template type
  typedef AutoDiff<T> Type;
  // Template base type
  typedef T BaseType;
  // Template numeric type
  typedef typename FunctionTraits_P<T>::NumericType NumericType;
  // Type for parameters
  typedef AutoDiff<T> ParamType;
  // Type for arguments
  typedef T ArgType;
  // Default type for differentiation
  typedef AutoDiff<T> DiffType;
  // Get the value
  static const T &getValue(const Type &in) { return FunctionTraits<T>::getValue(in.value()); }
  // Set a value (and possible derivative)
  static void setValue(Type &out, const T &val, const uInt nder, const uInt i) {
    out = Type(val, nder, i);
  }
};
 
#undef FunctionTraits_P
 
#define FunctionTraits_PA FunctionTraits
 
// <summary> FunctionTraits specialization for AutoDiffA
// </summary>
 
template <class T>
class FunctionTraits_PA<AutoDiffA<T>> {
 public:
  // Actual template type
  typedef AutoDiffA<T> Type;
  // Template base type
  typedef T BaseType;
  // Template numeric type
  typedef typename FunctionTraits_PA<T>::NumericType NumericType;
  // Type for parameters
  typedef AutoDiffA<T> ParamType;
  // Type for arguments
  typedef AutoDiffA<T> ArgType;
  // Default type for differentiation
  typedef AutoDiffA<T> DiffType;
  // Get the value
  static const T &getValue(const Type &in) { return FunctionTraits<T>::getValue(in.value()); }
  // Set a value (and possible derivative)
  static void setValue(Type &out, const T &val, const uInt nder, const uInt i) {
    out = Type(val, nder, i);
  }
};
 
#undef FunctionTraits_PA
 
#define FunctionTraits_PX FunctionTraits
 
// <summary> FunctionTraits specialization for AutoDiffX
// </summary>
 
template <class T>
class FunctionTraits_PX<AutoDiffX<T>> {
 public:
  // Actual template type
  typedef AutoDiffX<T> Type;
  // Template base type
  typedef T BaseType;
  // Template numeric type
  typedef typename FunctionTraits_PX<T>::NumericType NumericType;
  // Type for parameters
  typedef T ParamType;
  // Type for arguments
  typedef AutoDiffX<T> ArgType;
  // Default type for differentiation
  typedef AutoDiffX<T> DiffType;
  // Get the value
  static const T &getValue(const Type &in) { return FunctionTraits<T>::getValue(in.value()); }
  // Set a value (and possible derivative)
  static void setValue(Type &out, const T &val, const uInt nder, const uInt i) {
    out = Type(val, nder, i);
  }
};
 
#undef FunctionTraits_PX
 
}  // namespace casacore
 
#endif