diff --git a/regtest/crystallization/rt-pamm-aperiodic/clusters.dat b/regtest/crystallization/rt-pamm-aperiodic/clusters.dat
index 21c770a1ff454ff10d3329095afc00069a4c0cb3..9402d5ff064e47554baa9bf619bc788078480107 100755
--- a/regtest/crystallization/rt-pamm-aperiodic/clusters.dat
+++ b/regtest/crystallization/rt-pamm-aperiodic/clusters.dat
@@ -1,3 +1,5 @@
-#! FIELDS weight center-fcc width-fcc-fcc
+#! FIELDS height fcc sigma_fcc_fcc
+#! SET multivariate von-misses
+#! SET kerneltype gaussian
4.13163724E-0001 1.44833887E-0002 2.52246630E-0002
5.86836276E-0001 8.72359830E-0001 1.71959514E-0002
diff --git a/regtest/crystallization/rt-pamm-aperiodic/plumed.dat b/regtest/crystallization/rt-pamm-aperiodic/plumed.dat
index a241bed3d80e618570c8613813d2ea2cdc24adb1..7b7615778293dc052edb70eea0b3ce21c8dbc75c 100755
--- a/regtest/crystallization/rt-pamm-aperiodic/plumed.dat
+++ b/regtest/crystallization/rt-pamm-aperiodic/plumed.dat
@@ -1,6 +1,6 @@
FCCUBIC SPECIESA=1-2 SPECIESB=1-1200 SWITCH={CUBIC D_0=1.2 D_MAX=1.5} MEAN LABEL=fcc
-PAMM DATA=fcc CLUSTERS=clusters.dat MEAN LABEL=pamm
+PAMM DATA=fcc CLUSTERS=clusters.dat MEAN REGULARISE=0.0 LABEL=pamm
# PAMM DATA=fcc CLUSTERS=clusters.dat MEAN NUMERICAL_DERIVATIVES LABEL=pammn
PRINT ARG=pamm.* FILE=colvar FMT=%8.4f
diff --git a/regtest/multicolvar/rt-pamm-periodic/2D-testc-0.75.pammp b/regtest/multicolvar/rt-pamm-periodic/2D-testc-0.75.pammp
index 651fe29725a9721663ced7d8107729ea65d21057..9778b62d483ada4fdc813538ff32edd416619fc7 100755
--- a/regtest/multicolvar/rt-pamm-periodic/2D-testc-0.75.pammp
+++ b/regtest/multicolvar/rt-pamm-periodic/2D-testc-0.75.pammp
@@ -1,4 +1,6 @@
-#! FIELDS weight center-phi center-psi width-phi-phi width-phi-psi width-psi-phi width-psi-psi
+#! FIELDS height phi psi sigma_phi_phi sigma_phi_psi sigma_psi_phi sigma_psi_psi
+#! SET multivariate von-misses
+#! SET kerneltype gaussian
2.97197455E-0001 -1.91983118E+0000 2.25029540E+0000 2.45960237E-0001 -1.30615381E-0001 -1.30615381E-0001 2.40239117E-0001
2.29131448E-0002 1.39809354E+0000 9.54585380E-0002 9.61755708E-0002 -3.55657919E-0002 -3.55657919E-0002 1.06147253E-0001
5.06676398E-0001 -1.09648066E+0000 -7.17867907E-0001 1.40523052E-0001 -1.05385552E-0001 -1.05385552E-0001 1.63290557E-0001
diff --git a/regtest/multicolvar/rt-pamm-periodic/plumed.dat b/regtest/multicolvar/rt-pamm-periodic/plumed.dat
index 356d6a0796ac58e4b85783669f234d0c5ea8b9d3..d92e01bca8c8a137edd8ae7679ce3f7ad2ba4c5a 100755
--- a/regtest/multicolvar/rt-pamm-periodic/plumed.dat
+++ b/regtest/multicolvar/rt-pamm-periodic/plumed.dat
@@ -17,8 +17,8 @@ TORSIONS ...
DUMPMULTICOLVAR DATA=psi FILE=psi.xyz
DUMPMULTICOLVAR DATA=phi FILE=phi.xyz
-PAMM DATA=phi,psi CLUSTERS=2D-testc-0.75.pammp MEAN1={COMPONENT=1} HISTOGRAM={GAUSSIAN NBINS=2 LOWER=0.0 UPPER=0.5 COMPONENT=2 SMEAR=0.1} HIGHEST={COMPONENT=4} MOMENTS1={MOMENTS=2-3 COMPONENT=1} MOMENTS2={MOMENTS=2-3 COMPONENT=5} LABEL=p
-# PAMM DATA=phi,psi CLUSTERS=2D-testc-0.75.pammp MEAN1={COMPONENT=1} HISTOGRAM={GAUSSIAN NBINS=2 LOWER=0.0 UPPER=0.5 COMPONENT=2} HIGHEST={COMPONENT=4} MOMENTS1={MOMENTS=2-3 COMPONENT=1} MOMENTS2={MOMENTS=2-3 COMPONENT=5} NUMERICAL_DERIVATIVES LABEL=pn
+PAMM DATA=phi,psi CLUSTERS=2D-testc-0.75.pammp MEAN1={COMPONENT=1} HISTOGRAM={GAUSSIAN NBINS=2 LOWER=0.0 UPPER=0.5 COMPONENT=2 SMEAR=0.1} HIGHEST={COMPONENT=4} MOMENTS1={MOMENTS=2-3 COMPONENT=1} MOMENTS2={MOMENTS=2-3 COMPONENT=5} REGULARISE=0.0 LABEL=p
+# PAMM DATA=phi,psi CLUSTERS=2D-testc-0.75.pammp MEAN1={COMPONENT=1} HISTOGRAM={GAUSSIAN NBINS=2 LOWER=0.0 UPPER=0.5 COMPONENT=2} HIGHEST={COMPONENT=4} MOMENTS1={MOMENTS=2-3 COMPONENT=1} MOMENTS2={MOMENTS=2-3 COMPONENT=5} NUMERICAL_DERIVATIVES REGULARISE=0.0 LABEL=pn
PRINT ARG=p.* FILE=colvar FMT=%8.4f
DUMPDERIVATIVES ARG=p.* FILE=deriv FMT=%8.4f
diff --git a/src/multicolvar/PAMM.cpp b/src/multicolvar/PAMM.cpp
index 700057fe5167c4426384449185d0b542d28214db..ca43717f3e6138e6524ca5dbb30fde8c2c33e81c 100644
--- a/src/multicolvar/PAMM.cpp
+++ b/src/multicolvar/PAMM.cpp
@@ -68,13 +68,15 @@ PRINT ARG=p.mean-1,mean-2 FILE=colvar
The best place to start our explanation is to look at the contents of the clusters.dat file
\verbatim
-#! FIELDS weight center-phi center-psi width-phi-phi width-phi-psi width-psi-phi width-psi-psi
+#! FIELDS height phi psi sigma_phi_phi sigma_phi_psi sigma_psi_phi sigma_psi_psi
+#! SET multivariate von-misses
+#! SET kerneltype gaussian
0.4 -1.0 -1.0 0.2 -0.1 -0.1 0.2
0.6 1.0 +1.0 0.1 -0.03 -0.03 0.1
\endverbatim
This files contains the parameters of two two-dimensional Gaussian functions. Each of these Gaussians has a weight, \f$w_k\f$,
-a vector that specifies the position of its centre, \f$\mathbf{c}_\f$, and a covariance matrix, \f$\Sigma_k\f$. The \f$\phi_k\f$ functions that
+a vector that specifies the position of its centre, \f$\mathbf{c}_k\f$, and a covariance matrix, \f$\Sigma_k\f$. The \f$\phi_k\f$ functions that
we use to calculate our PAMM components are thus:
\f[
@@ -107,7 +109,7 @@ namespace multicolvar {
class PAMM : public MultiColvarFunction {
private:
double regulariser;
- std::vector<KernelFunctions> kernels;
+ std::vector<KernelFunctions*> kernels;
std::vector<Value*> pos;
public:
static void registerKeywords( Keywords& keys );
@@ -171,49 +173,33 @@ MultiColvarFunction(ao)
pos[i]->setDomain( min, max );
}
}
+
+ // Get labels for base multicolvars
+ std::vector<std::string> valnames( getNumberOfBaseMultiColvars() );
+ for(unsigned i=0;i<valnames.size();++i) valnames[i]=getBaseMultiColvar(i)->getLabel();
- std::string filename; parse("CLUSTERS",filename);
- unsigned ncolv = getNumberOfBaseMultiColvars();
- IFile ifile; Matrix<double> covar( ncolv, ncolv ), icovar( ncolv, ncolv );
+ std::string filename; parse("CLUSTERS",filename); IFile ifile;
if( !ifile.FileExist(filename) ) error("could not find file named " + filename);
- ifile.open(filename); ifile.allowIgnoredFields(); double weight, wij;
- std::vector<double> center( ncolv );
- std::vector<double> sig( 0.5*ncolv*(ncolv-1) + ncolv );
- while( ifile.scanField("weight",weight) ) {
- for(unsigned i=0;i<center.size();++i){
- ifile.scanField("center-"+getBaseMultiColvar(i)->getLabel(),center[i]);
- }
- for(unsigned i=0;i<center.size();++i){
- for(unsigned j=0;j<center.size();++j){
- ifile.scanField("width-"+getBaseMultiColvar(i)->getLabel()+"-" + getBaseMultiColvar(j)->getLabel(),covar(i,j) );
- }
- }
- Invert( covar, icovar );
- unsigned k=0;
- for(unsigned i=0;i<ncolv;i++){
- for(unsigned j=i;j<ncolv;j++){ sig[k]=icovar(i,j); k++; }
- }
+ ifile.open(filename); ifile.allowIgnoredFields();
+
+ for(unsigned i=0;;++i){
+ KernelFunctions* kk = KernelFunctions::read( &ifile, false, valnames );
+ if( !kk ) break ;
+ kk->normalize( pos );
+ kernels.push_back( kk );
ifile.scanField();
- kernels.push_back( KernelFunctions( center, sig, "GAUSSIAN", "VON-MISES", weight ) );
- kernels[kernels.size()-1].normalize( pos );
}
ifile.close();
// This builds the lists
buildSets();
-
- // I think we can put this back for anything with not usespecies
- // Now need to pass cutoff information to underlying distance multicolvars
- // for(unsigned i=0;i<getNumberOfBaseMultiColvars();++i){
- // Distances* mydist = dynamic_cast<Distances*>( getBaseMultiColvar(i) );
- // for(unsigned k=0;k<kernels.size();++k){
- // mydist->setLinkCellCutoff( kernels[k].getCenter()[i]+kernels[k].getContinuousSupport()[i] );
- // }
- // }
+ // Read the regularisation parameter
+ parse("REGULARISE",regulariser);
}
PAMM::~PAMM(){
for(unsigned i=0;i<pos.size();++i) delete pos[i];
+ for(unsigned i=0;i<kernels.size();++i) delete kernels[i];
}
unsigned PAMM::getNumberOfQuantities(){
@@ -258,7 +244,7 @@ double PAMM::compute( const unsigned& tindex, AtomValuePack& myatoms ) const {
// Evaluate the set of kernels
double denom=regulariser;
for(unsigned i=0;i<kernels.size();++i){
- vals[i]=kernels[i].evaluate( pos, tderiv[i] );
+ vals[i]=kernels[i]->evaluate( pos, tderiv[i] );
denom+=vals[i];
for(unsigned j=0;j<nvars;++j) dderiv[j] += tderiv[i][j];
}