hillclimber 0.1.6__cp314-cp314t-manylinux_2_27_aarch64.manylinux_2_28_aarch64.whl

plumed/_lib/include/plumed/core/Value.h

@@ -0,0 +1,547 @@
+/* +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
+   Copyright (c) 2011-2023 The plumed team
+   (see the PEOPLE file at the root of the distribution for a list of names)
+
+   See http://www.plumed.org for more information.
+
+   This file is part of plumed, version 2.
+
+   plumed is free software: you can redistribute it and/or modify
+   it under the terms of the GNU Lesser General Public License as published by
+   the Free Software Foundation, either version 3 of the License, or
+   (at your option) any later version.
+
+   plumed 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 Lesser General Public License for more details.
+
+   You should have received a copy of the GNU Lesser General Public License
+   along with plumed.  If not, see <http://www.gnu.org/licenses/>.
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ */
+#ifndef __PLUMED_core_Value_h
+#define __PLUMED_core_Value_h
+
+#include <vector>
+#include <string>
+#include <map>
+#include "../tools/Exception.h"
+#include "../tools/Tools.h"
+#include "../tools/AtomNumber.h"
+#include "../tools/Vector.h"
+#include "../tools/View.h"
+
+namespace PLMD {
+
+class OFile;
+class Communicator;
+class ActionWithValue;
+class ActionAtomistic;
+
+/// \ingroup TOOLBOX
+/// A class for holding the value of a function together with its derivatives.
+/// Typically, an  object of type PLMD::ActionWithValue will contain one
+/// object of type PLUMD::Value that will be named after the label.  If the
+/// PLMD::ActionWithValue is part of a class that calculates multiple components
+/// then the class will contain multiple that will be called label.component-name
+/// This class is used to pass information between different PLMD::Action
+/// objects.  However, if you find a use for a tempory PLMD::Value in some method
+/// you are implementing please feel free to use it.
+class Value {
+  friend struct ArgumentsBookkeeping;
+  friend class ActionWithValue;
+  friend class ActionWithVector;
+  friend class ActionWithMatrix;
+  friend class ActionAtomistic;
+  friend class ActionWithArguments;
+  friend class ActionWithVirtualAtom;
+  friend class DomainDecomposition;
+  template<typename T>
+  friend class DataPassingObjectTyped;
+private:
+/// The action in which this quantity is calculated
+  ActionWithValue* action=nullptr;
+/// Had the value been set
+  bool value_set=false;
+/// The value of the quantity
+  std::vector<double> data;
+/// The force acting on this quantity
+  std::vector<double> inputForce;
+/// A flag telling us we have a force acting on this quantity
+  bool hasForce=false;
+/// The way this value is used in the code
+/// normal = regular value that is determined during calculate
+/// constant = constnt value that is determined during startup and that doesn't change during simulation
+/// average = value that is averaged/collected over multiple steps of trajectory
+/// calcFromAverage = value that is calculated from an average value
+  enum {normal,constant,average,calcFromAverage} valtype=normal;
+/// This is used by ActionWithValue to set the valtype
+  void setValType( const std::string& vtype );
+/// This is used by ActionWithValue to determine if we need to calculate on update
+  bool calculateOnUpdate() const ;
+/// The derivatives of the quantity stored in value
+  std::map<AtomNumber,Vector> gradients;
+/// The name of this quantiy
+  std::string name;
+/// What is the shape of the value (0 dimensional=scalar, n dimensional with derivatives=grid, 1 dimensional no derivatives=vector, 2 dimensional no derivatives=matrix)
+  std::vector<std::size_t> shape;
+/// Does this quanity have derivatives
+  bool hasDeriv=true;
+/// Variables for storing data
+  unsigned bufstart=0;
+  unsigned ngrid_der=0;
+  std::size_t ncols=0;
+/// If we are storing a matrix is it symmetric?
+  bool symmetric=false;
+/// This is a bookeeping array that holds the non-zero elements of the "sparse" matrix
+  std::vector<unsigned> matrix_bookeeping;
+/// Is this quantity periodic
+  enum {unset,periodic,notperiodic} periodicity=unset;
+/// Various quantities that describe the domain of this value
+  std::string str_min;
+  std::string str_max;
+  double min=0.0;
+  double max=0.0;
+  double max_minus_min=0.0;
+  double inv_max_minus_min=0.0;
+/// Is the derivative of this quantity zero when the value is zero
+  bool derivativeIsZeroWhenValueIsZero=false;
+/// Complete the setup of the periodicity
+  void setupPeriodicity();
+// bring value within PBCs
+  void applyPeriodicity( const unsigned& ival );
+public:
+/// A constructor that can be used to make Vectors of values
+  Value();
+/// A constructor that can be used to make Vectors of named values
+  explicit Value(const std::string& valname);
+/// A constructor that is used throughout the code to setup the value poiters
+  Value(ActionWithValue* av, const std::string& valname, const bool withderiv,const std::vector<std::size_t>&ss=std::vector<std::size_t>());
+/// Set the shape of the Value
+  void setShape( const std::vector<std::size_t>&ss );
+/// Set the value of the function
+  void set(double);
+/// Set the value of the stored data
+  void set(const std::size_t& n, const double& v );
+/// Add something to the value of the function
+  void add(double);
+/// Add something to the ith element of the data array
+  void add(const std::size_t& n, const double& v );
+/// Get the location of this element of in the store
+  std::size_t getIndexInStore( const std::size_t& ival ) const ;
+/// Get the value of the function
+  double get( const std::size_t ival=0, const bool trueind=true ) const;
+/// A variant of get() for checking that at least one of the values on the row is !=0
+  bool checkValueIsActiveForMMul(std::size_t task) const;
+/// A variant of get() for assigning data to an external view (assuems trueind=false), returns the number of arguments assigned
+  std::size_t assignValues(View<double> target);
+/// Find out if the value has been set
+  bool valueHasBeenSet() const;
+/// Check if the value is periodic
+  bool isPeriodic() const;
+/// Set the function not periodic
+  void setNotPeriodic();
+/// Set the domain of the function
+  void setDomain(const std::string&, const std::string&);
+/// Get the domain of the quantity
+  void getDomain(std::string&,std::string&) const;
+/// Get the domain of the quantity
+  void getDomain(double&,double&) const;
+/// Get the name of the quantity
+  const std::string& getName() const;
+/// Check whether or not this particular quantity has derivatives
+  bool hasDerivatives()const;
+/// Get the number of derivatives that this particular value has
+  unsigned getNumberOfDerivatives() const;
+/// Set the number of derivatives
+  void resizeDerivatives(int n);
+/// Set all the derivatives to zero
+  void clearDerivatives( const bool force=false );
+/// Add some derivative to the ith component of the derivatives array
+  void addDerivative(unsigned i,double d);
+/// Set the value of the ith component of the derivatives array
+  void setDerivative(unsigned i, double d);
+/// Get the derivative with respect to component n
+  double getDerivative(const unsigned n) const;
+/// Clear the input force on the variable
+  void clearInputForce();
+/// Special method for clearing forces on variables used by DataPassingObject
+  void clearInputForce( const std::vector<AtomNumber>& index );
+/// Set hasForce equal to true
+  void addForce();
+/// Add some force on this value
+  void addForce(double f);
+/// Add some force on the ival th component of this value
+  void addForce( const std::size_t& ival, double f, const bool trueind=true );
+  ///Add forces from a vector, imples trueInd=false and retunrs the number of forces assigned
+  std::size_t addForces(View<const double> f);
+/// Get the value of the force on this colvar
+  double getForce( const std::size_t& ival=0 ) const ;
+/// Apply the forces to the derivatives using the chain rule (if there are no forces this routine returns false)
+  bool applyForce( std::vector<double>& forces ) const ;
+/// Calculate the difference between the instantaneous value of the function and some other point: other_point-inst_val
+  double difference(double)const;
+/// Calculate the difference between two values of this function: d2 -d1
+  double difference(double d1,double d2)const;
+/// This returns the pointer to the action where this value is calculated
+  ActionWithValue* getPntrToAction();
+/// Bring back one value into the correct pbc if needed, else give back the value
+  double bringBackInPbc(double d1)const;
+/// Get the difference between max and minimum of domain
+  double getMaxMinusMin()const;
+/// This sets up the gradients
+  void setGradients( ActionAtomistic* aa, unsigned& start );
+/// This passes gradients from one action to another
+  void passGradients( const double& der, std::map<AtomNumber,Vector>& g ) const ;
+  static double projection(const Value&,const Value&);
+/// Get the rank of the object that is contained in this value
+  unsigned getRank() const ;
+/// Get the shape of the object that is contained in this value
+  const std::vector<std::size_t>& getShape() const ;
+/// Reshape the storage for sparse matrices
+  void reshapeMatrixStore( const unsigned& n );
+/// Copy the matrix bookeeping stuff
+  void copyBookeepingArrayFromArgument( Value* myarg );
+/// Set the symmetric flag equal true for this matrix
+  void setSymmetric( const bool& sym );
+/// Get the total number of scalars that are stored here
+  std::size_t getNumberOfValues() const ;
+/// Get the number of values that are actually stored here once sparse matrices are taken into account
+  std::size_t getNumberOfStoredValues() const ;
+/// Get the number of threads to use when assigning this value
+  unsigned getGoodNumThreads( const unsigned& j, const unsigned& k ) const ;
+/// These are used for passing around the data in this value when we are doing replica exchange
+  void writeBinary(std::ostream&o) const ;
+  void readBinary(std::istream&i);
+/// These are used for making constant values
+  bool isConstant() const ;
+  void setConstant();
+  void reshapeConstantValue( const std::vector<std::size_t>& sh );
+/// Check if forces have been added on this value
+  bool forcesWereAdded() const ;
+/// Set a bool that tells us if the derivative is zero when the value is zero true
+  void setDerivativeIsZeroWhenValueIsZero();
+/// Return a bool that tells us if the derivative is zero when the value is zero
+  bool isDerivativeZeroWhenValueIsZero() const ;
+/// Convert the input index to its corresponding indices
+  void convertIndexToindices(const std::size_t& index, std::vector<unsigned>& indices ) const ;
+/// Print out all the values in this Value
+  void print( OFile& ofile ) const ;
+/// Print out all the forces in this Value
+  void printForce( OFile& ofile ) const ;
+/// Are we to ignore the stored value
+  bool ignoreStoredValue(const std::string& n) const ;
+/// Set a matrix element to be non zero
+  void setMatrixBookeepingElement( const unsigned& i, const unsigned& n );
+///
+  unsigned getRowLength( const std::size_t& irow ) const ;
+///
+  unsigned getRowIndex( std::size_t irow, std::size_t jind ) const ;
+///
+  void setRowIndices( const std::size_t& irow, const std::vector<std::size_t>& ind );
+///
+  std::size_t getNumberOfColumns() const ;
+///
+  bool isSymmetric() const ;
+/// Retrieve the non-zero edges in a matrix
+  void retrieveEdgeList( unsigned& nedge, std::vector<std::pair<unsigned,unsigned> >& active, std::vector<double>& elems );
+/// Get the number of derivatives that the grid has
+  unsigned getNumberOfGridDerivatives() const ;
+/// get the derivative of a grid at a point n with resepct to argument j
+  double getGridDerivative(const unsigned& n, const unsigned& j ) const ;
+/// Add the derivatives of the grid to the corner
+  void addGridDerivatives( const unsigned& n, const unsigned& j, const double& val );
+///
+  void setGridDerivatives( const unsigned& n, const unsigned& j, const double& val );
+/// Add another value to the end of the data vector held by this value.  This is used in COLLECT
+  void push_back( const double& val );
+/// Get the type of value that is stored here
+  std::string getValueType() const ;
+/// Check if all the elements in the value have the same value
+  bool allElementsEqual() const ;
+};
+
+inline
+void Value::applyPeriodicity(const unsigned& ival) {
+  if(periodicity==periodic) {
+    data[ival]=min+difference(min,data[ival]);
+    if(data[ival]<min) {
+      data[ival]+=max_minus_min;
+    }
+  }
+}
+
+inline
+void Value::set(double v) {
+  value_set=true;
+  data[0]=v;
+  applyPeriodicity(0);
+}
+
+inline
+void Value::add(double v) {
+  value_set=true;
+  data[0]+=v;
+  applyPeriodicity(0);
+}
+
+inline
+void Value::add(const std::size_t& n, const double& v ) {
+  value_set=true;
+  data[n]+=v;
+  applyPeriodicity(n);
+}
+
+inline
+bool Value::valueHasBeenSet() const {
+  return value_set;
+}
+
+inline
+const std::string& Value::getName()const {
+  return name;
+}
+
+inline
+unsigned Value::getNumberOfDerivatives() const {
+  plumed_massert(hasDeriv,"the derivatives array for this value has zero size");
+  if( shape.size()>0 ) {
+    return shape.size();
+  }
+  return data.size() - 1;
+}
+
+inline
+double Value::getDerivative(const unsigned n) const {
+  plumed_dbg_massert(n<getNumberOfDerivatives(),"you are asking for a derivative that is out of bounds");
+  return data[1+n];
+}
+
+inline
+bool Value::hasDerivatives() const {
+  return hasDeriv;
+}
+
+inline
+void Value::resizeDerivatives(int n) {
+  if( shape.size()>0 ) {
+    return;
+  }
+  if(hasDeriv) {
+    data.resize(1+n);
+  }
+}
+
+inline
+void Value::addDerivative(unsigned i,double d) {
+  plumed_dbg_massert(i<getNumberOfDerivatives(),"derivative is out of bounds");
+  data[1+i]+=d;
+}
+
+inline
+void Value::setDerivative(unsigned i, double d) {
+  plumed_dbg_massert(i<getNumberOfDerivatives(),"derivative is out of bounds");
+  data[1+i]=d;
+}
+
+inline
+void Value::clearInputForce() {
+  if( !hasForce ) {
+    return;
+  }
+  hasForce=false;
+  std::fill(inputForce.begin(),inputForce.end(),0);
+}
+
+inline
+void Value::clearInputForce( const std::vector<AtomNumber>& index ) {
+  if( !hasForce ) {
+    return;
+  }
+  hasForce=false;
+  for(const auto & p : index) {
+    inputForce[p.index()]=0;
+  }
+}
+
+inline
+void Value::clearDerivatives( const bool force ) {
+  if( !force && (valtype==constant || valtype==average) ) {
+    return;
+  }
+
+  value_set=false;
+  if( shape.size()>0 ) {
+    std::fill(data.begin(), data.end(), 0);
+  } else if( data.size()>1 ) {
+    std::fill(data.begin()+1, data.end(), 0);
+  }
+}
+
+inline
+void Value::addForce() {
+  hasForce=true;
+}
+
+inline
+void Value::addForce(double f) {
+  hasForce=true;
+  inputForce[0]+=f;
+}
+
+inline
+bool Value::forcesWereAdded() const {
+  return hasForce;
+}
+
+inline
+double Value::getForce( const std::size_t& ival ) const {
+  plumed_dbg_assert( ival<inputForce.size() );
+  return inputForce[ival];
+}
+
+/// d2-d1
+inline
+double Value::difference(double d1,double d2)const {
+  if(periodicity==notperiodic) {
+    return d2-d1;
+  } else if(periodicity==periodic) {
+    double s=(d2-d1)*inv_max_minus_min;
+    // remember: pbc brings the difference in a range of -0.5:0.5
+    s=Tools::pbc(s);
+    return s*max_minus_min;
+  } else {
+    plumed_merror("periodicity should be set to compute differences");
+  }
+}
+
+inline
+double Value::bringBackInPbc(double d1)const {
+  return min+max_minus_min/2.+difference(min+max_minus_min/2., d1);
+}
+
+inline
+double Value::difference(double d)const {
+  return difference(get(),d);
+}
+
+inline
+double Value::getMaxMinusMin()const {
+  plumed_dbg_assert( periodicity==periodic );
+  return max_minus_min;
+}
+
+inline
+unsigned Value::getRank() const {
+  if( valtype==constant && shape.size()==1 && shape[0]==1 ) {
+    return 0;
+  }
+  return shape.size();
+}
+
+inline
+const std::vector<std::size_t>& Value::getShape() const {
+  return shape;
+}
+
+inline
+std::size_t Value::getNumberOfValues() const {
+  std::size_t size=1;
+  for(unsigned i=0; i<shape.size(); ++i) {
+    size *= shape[i];
+  }
+  return size;
+}
+
+inline
+std::size_t Value::getNumberOfStoredValues() const {
+  if( getRank()==2 && !hasDeriv ) {
+    return shape[0]*ncols;
+  }
+  return getNumberOfValues();
+}
+
+inline
+bool Value::isConstant() const {
+  return valtype==constant;
+}
+
+inline
+void Value::setDerivativeIsZeroWhenValueIsZero() {
+  derivativeIsZeroWhenValueIsZero=true;
+}
+
+inline
+bool Value::isDerivativeZeroWhenValueIsZero() const {
+  return derivativeIsZeroWhenValueIsZero;
+}
+
+inline
+void Value::setMatrixBookeepingElement( const unsigned& i, const unsigned& n ) {
+  plumed_dbg_assert( i<matrix_bookeeping.size() );
+  matrix_bookeeping[i]=n;
+}
+
+inline
+unsigned Value::getRowLength( const std::size_t& irow ) const {
+  if( matrix_bookeeping.size()==0 ) {
+    return 0;
+  }
+  plumed_dbg_assert( (1+ncols)*irow<matrix_bookeeping.size() );
+  return matrix_bookeeping[(1+ncols)*irow];
+}
+
+inline
+unsigned Value::getRowIndex(const std::size_t irow, const std::size_t jind ) const {
+  plumed_dbg_massert( (1+ncols)*irow+1+jind<matrix_bookeeping.size() && jind<matrix_bookeeping[(1+ncols)*irow], "failing in value " + name );
+  return matrix_bookeeping[(1+ncols)*irow+1+jind];
+}
+
+inline
+void Value::setRowIndices( const std::size_t& irow, const std::vector<std::size_t>& ind ) {
+  plumed_dbg_massert( (1+ncols)*irow+1+ind.size()<=matrix_bookeeping.size(), "problem in " + name );
+  std::size_t istart = (1+ncols)*irow;
+  matrix_bookeeping[istart] = ind.size();
+  ++istart;
+  for(unsigned i=0; i<ind.size(); ++i) {
+    matrix_bookeeping[istart] = ind[i];
+    ++istart;
+  }
+}
+
+inline
+std::size_t Value::getNumberOfColumns() const {
+  return ncols;
+}
+
+inline
+bool Value::isSymmetric() const {
+  return symmetric;
+}
+
+inline
+unsigned Value::getNumberOfGridDerivatives() const {
+  return ngrid_der;
+}
+
+inline
+double Value::getGridDerivative(const unsigned& n, const unsigned& j ) const {
+  plumed_dbg_assert( hasDeriv && n*(1+ngrid_der) + 1 + j < data.size() );
+  return data[n*(1+ngrid_der) + 1 + j];
+}
+
+inline
+void Value::addGridDerivatives( const unsigned& n, const unsigned& j, const double& val ) {
+  plumed_dbg_assert( hasDeriv && n*(1+ngrid_der) + 1 + j < data.size() );
+  data[n*(1+ngrid_der) + 1 + j] += val;
+}
+
+inline
+void Value::setGridDerivatives( const unsigned& n, const unsigned& j, const double& val ) {
+  plumed_dbg_assert( hasDeriv && n*(1+ngrid_der) + 1 + j < data.size() );
+  data[n*(1+ngrid_der) + 1 + j] = val;
+}
+
+}
+#endif
+

plumed/_lib/include/plumed/core/WithCmd.h

@@ -0,0 +1,93 @@
+/* +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
+   Copyright (c) 2011-2023 The plumed team
+   (see the PEOPLE file at the root of the distribution for a list of names)
+
+   See http://www.plumed.org for more information.
+
+   This file is part of plumed, version 2.
+
+   plumed is free software: you can redistribute it and/or modify
+   it under the terms of the GNU Lesser General Public License as published by
+   the Free Software Foundation, either version 3 of the License, or
+   (at your option) any later version.
+
+   plumed 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 Lesser General Public License for more details.
+
+   You should have received a copy of the GNU Lesser General Public License
+   along with plumed.  If not, see <http://www.gnu.org/licenses/>.
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ */
+#ifndef __PLUMED_core_WithCmd_h
+#define __PLUMED_core_WithCmd_h
+
+#include "../tools/TypesafePtr.h"
+#include <string>
+#include <string_view>
+
+namespace PLMD {
+
+/// Base for classes with cmd() method.
+/// This is an abstract base class for classes with
+/// cmd() method.
+class WithCmd {
+  /// Small structure used to pass elements of a shape initializer_list
+  struct SizeLike {
+    std::size_t size;
+    SizeLike(short unsigned dim): size(dim) {}
+    SizeLike(unsigned dim): size(dim) {}
+    SizeLike(long unsigned dim): size(dim) {}
+    SizeLike(long long unsigned dim): size(dim) {}
+    SizeLike(short dim): size(std::size_t(dim)) {}
+    SizeLike(int dim): size(std::size_t(dim)) {}
+    SizeLike(long int dim): size(std::size_t(dim)) {}
+    SizeLike(long long int dim): size(std::size_t(dim)) {}
+  };
+public:
+  /// This is the preferred method as it avoid allocations of temporaries.
+  /// If this is not overridded, it will call the legacy method.
+  virtual void cmd(std::string_view key,const TypesafePtr & val=nullptr) {
+    cmd(std::string(key),val);
+  }
+  /// This is the legacy method we used in older plumed versions, so it is still possible.
+  /// If this is not overridden, it will call the preferred method
+  virtual void cmd(const std::string& key,const TypesafePtr & val=nullptr) {
+    cmd(std::string_view(key),val);
+  }
+  void cmd(std::string_view key,const TypesafePtr & val,const std::size_t* shape) {
+    cmd(key,TypesafePtr::setNelemAndShape(val,0,shape));
+  }
+  void cmd(std::string_view key,const TypesafePtr & val,std::initializer_list<SizeLike> shape) {
+    if(shape.size()>4) {
+      plumed_error() << "Maximum shape size is 4";
+    }
+    std::array<std::size_t,5> shape_;
+    unsigned j=0;
+    for(auto i : shape) {
+      shape_[j]=i.size;
+      j++;
+    }
+    shape_[j]=0;
+    cmd(key,val,shape_.data());
+  }
+  template<typename I, typename std::enable_if<std::is_integral<I>::value, int>::type = 0>
+  void cmd(std::string_view key,const TypesafePtr & val,I nelem, const std::size_t* shape=nullptr) {
+    cmd(key,TypesafePtr::setNelemAndShape(val,nelem,shape));
+  }
+  /// This is needed to avoid ambiguities
+  void cmd(const char* key,const TypesafePtr & val=nullptr) {
+    cmd(std::string_view(key),val);
+  }
+  virtual ~WithCmd();
+};
+
+inline
+WithCmd::~WithCmd() {
+// do nothing
+// here just to allow inheriting from this class properly
+}
+
+}
+
+#endif

plumed/_lib/include/plumed/dimred/SMACOF.h

@@ -0,0 +1,55 @@
+/* +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
+   Copyright (c) 2015-2020 The plumed team
+   (see the PEOPLE file at the root of the distribution for a list of names)
+
+   See http://www.plumed.org for more information.
+
+   This file is part of plumed, version 2.
+
+   plumed is free software: you can redistribute it and/or modify
+   it under the terms of the GNU Lesser General Public License as published by
+   the Free Software Foundation, either version 3 of the License, or
+   (at your option) any later version.
+
+   plumed 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 Lesser General Public License for more details.
+
+   You should have received a copy of the GNU Lesser General Public License
+   along with plumed.  If not, see <http://www.gnu.org/licenses/>.
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ */
+#ifndef __PLUMED_dimred_SMACOF_h
+#define __PLUMED_dimred_SMACOF_h
+
+#include <vector>
+#include "../core/Value.h"
+#include "../tools/Matrix.h"
+
+namespace PLMD {
+namespace dimred {
+
+class SMACOF {
+private:
+  Matrix<double> Distances, Weights;
+  double calculateSigma( const Matrix<double>& InitialZ, Matrix<double>& dists );
+public:
+  explicit SMACOF( const Value* mysquaredists );
+  void optimize( const double& tol, const unsigned& maxloops, std::vector<double>& proj);
+  double getDistance( const unsigned& i, const unsigned& j ) const ;
+  void setWeight( const unsigned& i, const unsigned& j, const double& ww );
+};
+
+inline
+double SMACOF::getDistance( const unsigned& i, const unsigned& j ) const {
+  return Distances( i, j );
+}
+
+inline
+void SMACOF::setWeight( const unsigned& i, const unsigned& j, const double& ww ) {
+  Weights(i,j) = Weights(j,i ) = ww;
+}
+
+}
+}
+#endif