diff --git a/src/ActionWithDistribution.cpp b/src/ActionWithDistribution.cpp
index 855539e404d85ee01ec78cef715c8910d4249bde..889fa8ba202d0188cfd60b7e3f7e998feb631719 100644
--- a/src/ActionWithDistribution.cpp
+++ b/src/ActionWithDistribution.cpp
@@ -4,15 +4,6 @@ using namespace std;
using namespace PLMD;
void ActionWithDistribution::registerKeywords(Keywords& keys){
- keys.add("nohtml","MIN","take the minimum value from these variables. The continuous version of the minimum described above is calculated and beta must be specified in input");
-// keys.add("optional","MAX", "take the maximum value from these variables");
- keys.add("nohtml","SUM", "take the sum of these variables");
- keys.add("nohtml","AVERAGE", "take the average value of these variables");
- keys.add("nohtml","LESS_THAN", "take the number of variables less than the specified target. This quantity is made differentiable using a switching function. You can control the parameters of this switching function by specifying three numbers to the keyword (r_0, nn and mm). If you are happy with the default values of nn=6 and mm=12 then you need only specify the target r_0. The number of values less than the target is stored in a value called lt<target>.");
- keys.add("nohtml","MORE_THAN", "take the number of variables more than the specified target. This quantity is made differentiable using a switching function. You can control the parameters of this switching function by specifying three numbers to the keyword (r_0, nn and mm). If you are happy with the default values of nn=6 and mm=12 then you need only specify the target r_0. The number of values less than the target is stored in a value called gt<target>.");
- keys.add("nohtml","HISTOGRAM", "create a discretized histogram of the distribution. This keyword's input should consist of one or two numbers");
- keys.add("nohtml","RANGE", "the range in which to calculate the histogram");
- keys.add("nohtml", "WITHIN", "The number of values within a certain range. This keyword's input should consist of one or two numbers.");
keys.reserve("optional","NL_STRIDE","the frequency with which the neighbor list should be updated");
}
@@ -57,72 +48,11 @@ void ActionWithDistribution::addDistributionFunction( std::string name, Distribu
log.printf(" value %s contains %s\n",( (a->copyOutput( fno ))->getName() ).c_str(),( functions[fno]->message() ).c_str() );
}
-void ActionWithDistribution::readDistributionKeywords(){
+void ActionWithDistribution::requestDistribution(){
read=true; bool dothis; std::vector<std::string> params;
ActionWithValue*a=dynamic_cast<ActionWithValue*>(this);
plumed_massert(a,"can only do distribution on ActionsWithValue");
- // Read SUM keyword
- parseFlag("SUM",dothis);
- if( dothis && keywords.exists("NL_CHEAT") && updateFreq>0 ) error("cannot use SUM and neighbor list for this keyword");
- if( dothis ){
- addDistributionFunction( "sum", new sum(params) );
- }
- // Read AVERAGE keyword
- parseFlag("AVERAGE",dothis);
- if( dothis && keywords.exists("NL_CHEAT") && updateFreq>0 ) error("cannot use AVERAGE and neighbor list for this keyword");
- if( dothis ){
- params.resize(1); Tools::convert(getNumberOfFunctionsInDistribution(),params[0]);
- addDistributionFunction( "average", new mean(params) );
- }
- // Read MIN keyword
- parseVector("MIN",params);
- if( params.size()!=0 && keywords.exists("NL_CHEAT") && updateFreq>0 ) error("cannot use MIN and neighbor list for this keyword");
- if( params.size()!=0 ){
- addDistributionFunction( "min", new min(params) );
- }
- // Read Less_THAN keyword
- parseVector("LESS_THAN",params);
- if( params.size()!=0 ){
- addDistributionFunction( "lt" + params[0], new less_than(params) );
- }
- // Read MORE_THAN keyword
- parseVector("MORE_THAN",params);
- if( params.size()!=0 ){
- addDistributionFunction( "gt" + params[0], new more_than(params) );
- }
- // Read HISTOGRAM keyword
- parseVector("HISTOGRAM",params);
- if( params.size()!=0 ){
- std::vector<double> range(2); parseVector("RANGE",range);
- if(range[1]<range[0]) error("range is meaningless");
- int nbins; Tools::convert(params[0],nbins);
- std::vector<std::string> hparams(2);
- if(params.size()==2){
- hparams.resize(3);
- hparams[2]=params[1];
- } else if(params.size()!=1){
- error("Histogram keyword should either specify just the number"
- " of bins or the number of bins and the ammount of smearing");
- }
- double lb,ub,delr=(range[1]-range[0])/static_cast<double>(nbins);
- for(int i=0;i<nbins;++i){
- lb=range[0]+i*delr; Tools::convert( lb, hparams[0] );
- ub=range[0]+(i+1)*delr; Tools::convert( ub, hparams[1] );
- addDistributionFunction( "between" + hparams[0] + "&" +hparams[1], new within(hparams) );
- }
- }
- // Read within keywords
- parseVector("WITHIN",params);
- if( params.size()!=0 ){
- addDistributionFunction( "between" + params[0] + "&" +params[1], new within(params) );
- } else {
- for(unsigned i=1;;++i){
- if( !parseNumberedVector("WITHIN",i,params) ) break;
- addDistributionFunction( "between" + params[0] + "&" +params[1], new within(params) );
- }
- }
-
if(all_values){
if( a->checkNumericalDerivatives() && getNumberOfFunctionsInDistribution()>1 ){
error("Can only use numerical derivatives for distribution functions or single colvars");
@@ -157,7 +87,7 @@ void ActionWithDistribution::prepare(){
}
void ActionWithDistribution::calculate(){
- plumed_massert( read, "you must have a call to readDistributionKeywords somewhere" );
+ plumed_massert( read, "you must have a call to requestDistribution somewhere" );
unsigned stride=comm.Get_size();
unsigned rank=comm.Get_rank();
diff --git a/src/ActionWithDistribution.h b/src/ActionWithDistribution.h
index 3ca4d93bd050135e00b5abf92fe07c3a0a79d405..4644cce1103674a812f7d1af4bfdc37d9ebadfde 100644
--- a/src/ActionWithDistribution.h
+++ b/src/ActionWithDistribution.h
@@ -32,10 +32,10 @@ private:
/// The list of quantities that should be calculated
DynamicList members;
protected:
-/// Add a distribution function to the list
+/// Add a distribution function to the list (this routine must be called after construction of ActionWithValue)
void addDistributionFunction( std::string name, DistributionFunction* fun );
-/// Read the keywords for the distribution (this routine must be called after construction of ActionWithValue)
- void readDistributionKeywords();
+/// Complete the setup of this object (this routine must be called after construction of ActionWithValue)
+ void requestDistribution();
/// Find out if we are running the calculation without mpi
bool getSerial() const ;
/// This resets members so that we calculate all functions - this is used for neighbour list update
diff --git a/src/DistributionFunctions.cpp b/src/DistributionFunctions.cpp
index eb12cf62ae488368e9654089610930c83a55f5df..36086f2e708f115596519528503719ec79dc60db 100644
--- a/src/DistributionFunctions.cpp
+++ b/src/DistributionFunctions.cpp
@@ -2,6 +2,20 @@
namespace PLMD {
+DistributionFunctionDocs::DistributionFunctionDocs(){
+ std::string docs;
+ min::writeDocs( docs ); alldocs.insert( std::pair<std::string,std::string>("MIN",docs) );
+ sum::writeDocs( docs ); alldocs.insert( std::pair<std::string,std::string>("SUM",docs) );
+ mean::writeDocs( docs ); alldocs.insert( std::pair<std::string,std::string>("AVERAGE",docs) );
+ less_than::writeDocs( docs ); alldocs.insert( std::pair<std::string,std::string>("LESS_THAN",docs) );
+ more_than::writeDocs( docs ); alldocs.insert( std::pair<std::string,std::string>("MORE_THAN",docs) );
+ within::writeDocs( docs ); alldocs.insert( std::pair<std::string,std::string>("WITHIN",docs) );
+}
+
+void DistributionFunctionDocs::printDocs(const std::string& name){
+ if( alldocs.count(name)>0 ) std::cout<<alldocs[name];
+}
+
DistributionFunction::DistributionFunction( const std::vector<std::string>& parameters ):
fine(true)
{
diff --git a/src/DistributionFunctions.h b/src/DistributionFunctions.h
index 2733f0fef2a55ce40a61c612ed43a13866cce357..c2ae1909d8586353a83cacbd67ac9c2d133cbddc 100644
--- a/src/DistributionFunctions.h
+++ b/src/DistributionFunctions.h
@@ -22,6 +22,15 @@ on. This class is used in PLMD::ActionWithDistribution.
*/
//+ENDDEVELDOC
+class DistributionFunctionDocs {
+/// The documentation for this particular thing
+ std::map<std::string,std::string> alldocs;
+public:
+ DistributionFunctionDocs();
+/// Print the documentation for the object named
+ void printDocs(const std::string& name);
+};
+
class DistributionFunction {
private:
bool fine;
@@ -55,6 +64,7 @@ bool DistributionFunction::check() const {
class sum : public DistributionFunction {
public:
+ static void writeDocs( std::string& docs );
sum( const std::vector<std::string>& parameters );
double calculate( Value* value_in, std::vector<Value>& aux, Value* value_out );
void finish( const double& total, Value* value_out );
@@ -65,6 +75,7 @@ class mean : public DistributionFunction {
private:
double nvalues;
public:
+ static void writeDocs( std::string& docs );
mean( const std::vector<std::string>& parameters );
double calculate( Value* value_in, std::vector<Value>& aux, Value* value_out );
void finish( const double& total, Value* value_out );
@@ -77,6 +88,7 @@ private:
unsigned nn,mm;
SwitchingFunction sf;
public:
+ static void writeDocs( std::string& docs );
less_than( const std::vector<std::string>& parameters );
double calculate( Value* value_in, std::vector<Value>& aux, Value* value_out );
void finish( const double& total, Value* value_out );
@@ -89,6 +101,7 @@ private:
unsigned nn,mm;
SwitchingFunction sf;
public:
+ static void writeDocs( std::string& docs );
more_than( const std::vector<std::string>& parameters );
double calculate( Value* value_in, std::vector<Value>& aux, Value* value_out );
void finish( const double& total, Value* value_out );
@@ -100,6 +113,7 @@ private:
double a,b,sigma;
HistogramBead hist;
public:
+ static void writeDocs( std::string& docs );;
within( const std::vector<std::string>& parameters );
double calculate( Value* value_in, std::vector<Value>& aux, Value* value_out );
void finish( const double& total, Value* value_out );
@@ -110,6 +124,7 @@ class min : public DistributionFunction {
private:
double beta;
public:
+ static void writeDocs( std::string& docs );
min( const std::vector<std::string>& parameters );
double calculate( Value* value_in, std::vector<Value>& aux, Value* value_out );
void finish( const double& total, Value* value_out );
diff --git a/src/DistributionLessThan.cpp b/src/DistributionLessThan.cpp
index 247813234fd1293ecf060ff31be925cb06ff5542..cef345dee6c5f2b56c6fb65f3c681f722cfe0def 100644
--- a/src/DistributionLessThan.cpp
+++ b/src/DistributionLessThan.cpp
@@ -2,20 +2,18 @@
namespace PLMD {
-//+PLUMEDOC MODIFIER LESS_THAN
-/**
-
-Calculate the number of functions that are less than a target value. To make this quantity continuous it is calculated using:
-\f$
-\sigma = \sum_i \frac{ 1 - \left(\frac{ s_i }{ r_0 }\right)^{n} }{ 1 - \left(\frac{ s_i }{ r_0 }\right)^{m} }
-\f$
-, where \f$r_0\f$ is specied after the LESS_THAN keyword. By default \f$n\f$ and \f$m\f$ are set equal to 6 and 12 respectively.
-These values can be adjusted by using using three arguments for less than (\f$r_0\f$, \f$n\f$, \f$m\f$). Once calculated the final value is referenced
-using label.lt\f$r_0\f$.
-
-*/
-//+ENDPLUMEDOC
-
+void less_than::writeDocs(std::string& docs){
+ std::ostringstream ostr;
+ ostr<<"\\par LESS_THAN\n\n";
+ ostr<<"Calculate the number of functions that are less than a target value. To make this quantity continuous it is calculated using:\n";
+ ostr<<"\\f$\n";
+ ostr<<"\\sigma = \\sum_i \\frac{ 1 - \\left(\\frac{ s_i }{ r_0 }\\right)^{n} }{ 1 - \\left(\\frac{ s_i }{ r_0 }\\right)^{m} }\n";
+ ostr<<"\\f$\n";
+ ostr<<", where \\f$r_0\\f$ is specied after the LESS_THAN keyword. By default \\f$n\\f$ and \\f$m\\f$ are set equal to 6 and 12 respectively.\n";
+ ostr<<"These values can be adjusted by using using three arguments for less than (\\f$r_0\\f$, \\f$n\\f$, \\f$m\\f$). Once calculated the final value is referenced\n";
+ ostr<<"using label.lt\\f$r_0\\f$.\n";
+ docs=ostr.str();
+}
less_than::less_than( const std::vector<std::string>& parameters ) :
DistributionFunction(parameters)
diff --git a/src/DistributionMean.cpp b/src/DistributionMean.cpp
index 9587fa1bffae2fd75f97c8bf0e9a6cc015164b98..f9456e9509774baa59b158b1c0a660723b903a14 100644
--- a/src/DistributionMean.cpp
+++ b/src/DistributionMean.cpp
@@ -2,15 +2,13 @@
namespace PLMD {
-//+PLUMEDOC MODIFIER AVERAGE
-/**
-
-Calculate the average of all the colvars in the distribution. Once calculated the final value is referenced
-using lable.average.
-
-*/
-//+ENDPLUMEDOC
-
+void mean::writeDocs(std::string& docs){
+ std::ostringstream ostr;
+ ostr<<"\\par AVERAGE \n\n";
+ ostr<<"Calculate the average of all the colvars in the distribution. Once calculated the final value is referenced\n";
+ ostr<<"using lable.average.\n";
+ docs=ostr.str();
+}
mean::mean( const std::vector<std::string>& parameters ) :
DistributionFunction(parameters)
diff --git a/src/DistributionMin.cpp b/src/DistributionMin.cpp
index 22523ae0c39995db7e5e0619402f42b93d5e380c..5c5a77e9f0f0a1f29d37ecccf4823b8b8f06fc8a 100644
--- a/src/DistributionMin.cpp
+++ b/src/DistributionMin.cpp
@@ -2,20 +2,20 @@
namespace PLMD {
-//+PLUMEDOC MODIFIER MIN
-/**
-
-Calculate the minimum from the set of defined colvars. To make this quantity continuous the minimum is calculated using:
-
-\f[
-\textrm{min} = \frac{\beta}{ \log \sum_i \exp\left( \frac{\beta}{s_i} \right) }
-\f]
-
-The keyword MIN=\f$\beta\f$ tells plumed to calculate this quantity and sets \f$\beta\f$ to an appropriate value. Once
-calcualted the final value is referneced using label.min
-
-*/
-//+ENDPLUMEDOC
+void min::writeDocs(std::string& docs){
+ std::ostringstream ostr;
+ ostr<<"\\par MIN"<<std::endl;
+ ostr<<std::endl;
+ ostr<<"Calculate the minimum value. To make this quantity continuous the minimum is calculated using:"<<std::endl;
+ ostr<<std::endl;
+ ostr<<"\\f["<<std::endl;
+ ostr<<"\\textrm{min} = \\frac{\\beta}{ \\log \\sum_i \\exp\\left( \\frac{\\beta}{s_i} \\right) }"<<std::endl;
+ ostr<<"\\f]"<<std::endl;
+ ostr<<std::endl;
+ ostr<<"The keyword MIN=\\f$\\beta\\f$ tells plumed to calculate this quantity and sets \\f$\\beta\\f$ to an appropriate value. Once"<<std::endl;
+ ostr<<"calcualted the final value is referneced using label.min"<<std::endl;
+ docs=ostr.str();
+}
min::min( const std::vector<std::string>& parameters ) :
DistributionFunction(parameters)
diff --git a/src/DistributionMoreThan.cpp b/src/DistributionMoreThan.cpp
index 3db5f50fda2a9dc74f815daff6bcd0e6ad0eae9a..13a5d422f26b06a7a9e3cb669e18e931705852bf 100644
--- a/src/DistributionMoreThan.cpp
+++ b/src/DistributionMoreThan.cpp
@@ -2,20 +2,18 @@
namespace PLMD {
-//+PLUMEDOC MODIFIER LESS_THAN
-/**
-
-Calculate the number of functions that are more than a target value. To make this quantity continuous it is calculated using:
-\f$
-\sigma = 1.0 - \sum_i \frac{ 1 - \left(\frac{ s_i }{ r_0 }\right)^{n} }{ 1 - \left(\frac{ s_i }{ r_0 }\right)^{m} }
-\f$
-, where \f$r_0\f$ is specied after the LESS_THAN keyword. By default \f$n\f$ and \f$m\f$ are set equal to 6 and 12 respectively.
-These values can be adjusted by using using three arguments for less than (\f$r_0\f$, \f$n\f$, \f$m\f$). Once calculated the final value is referenced
-using label.lt\f$r_0\f$.
-
-*/
-//+ENDPLUMEDOC
-
+void more_than::writeDocs(std::string& docs){
+ std::ostringstream ostr;
+ ostr<<"\\par MORE_THAN\n\n";
+ ostr<<"Calculate the number of functions that are more than a target value. To make this quantity continuous it is calculated using:\n";
+ ostr<<"\\f$\n";
+ ostr<<"\\sigma = 1.0 - \\sum_i \\frac{ 1 - \\left(\\frac{ s_i }{ r_0 }\\right)^{n} }{ 1 - \\left(\\frac{ s_i }{ r_0 }\\right)^{m} }\n";
+ ostr<<"\\f$\n";
+ ostr<<", where \\f$r_0\\f$ is specied after the LESS_THAN keyword. By default \\f$n\\f$ and \\f$m\\f$ are set equal to 6 and 12 respectively.\n";
+ ostr<<"These values can be adjusted by using using three arguments for less than (\\f$r_0\\f$, \\f$n\\f$, \\f$m\\f$). Once calculated the final value is referenced\n";
+ ostr<<"using label.lt\\f$r_0\\f$.\n";
+ docs=ostr.str();
+}
more_than::more_than( const std::vector<std::string>& parameters ) :
DistributionFunction(parameters)
diff --git a/src/DistributionSum.cpp b/src/DistributionSum.cpp
index f112c2b746b6d32173a9a9d34dab9d6445f521ec..541f83da4cba8ad1bd33bbe53df982de55806358 100644
--- a/src/DistributionSum.cpp
+++ b/src/DistributionSum.cpp
@@ -2,15 +2,13 @@
namespace PLMD {
-//+PLUMEDOC MODIFIER SUM
-/**
-
-Calculate the sum of all the colvars in the distribution. Once calculated the final value is referenced
-using lable.sum.
-
-*/
-//+ENDPLUMEDOC
-
+void sum::writeDocs(std::string& docs){
+ std::ostringstream ostr;
+ ostr<<"\\par SUM \n\n";
+ ostr<<"Calculate the sum of all the colvars in the distribution. Once calculated the final value is referenced\n";
+ ostr<<"using lable.sum.\n";
+ docs=ostr.str();
+}
sum::sum( const std::vector<std::string>& parameters ) :
DistributionFunction(parameters)
diff --git a/src/DistributionWithin.cpp b/src/DistributionWithin.cpp
index fa752b6c39dc7ba8703a40361c05d878b26e14ed..5634318cf382e4b58a0f548617ec05a53a077e08 100644
--- a/src/DistributionWithin.cpp
+++ b/src/DistributionWithin.cpp
@@ -2,25 +2,22 @@
namespace PLMD {
-//+PLUMEDOC MODIFIER WITHIN
-/**
-
-Calculates the number of colvars that are within a certain range. To make this quantity continuous it is calculated using:
-
-\f[
-S = \sum_i \int_a^b G( s_i, \sigma*(b-a) )
-\f]
-
-where \f$G( s_i, \sigma )\f$ is a normalized Gaussian function of width \f$\sigma\f$ centered at the value of the colvar \f$s_i\f$. \f$a\f$ and \f$b\f$ are
-the lower and upper bounds of the range of interest respectively. The values of \f$a\f$ and \f$b\f$ must be specified using the WITHIN keyword. If this keyword
-has three input numbers then the third is assumed to be the value of \f$\sigma\f$. You can specify that you want to investigate multiple rangles by using multiple instances
-of the WITHIN keyword (WITHIN1,WITHIN2 etc). Alternatively, if you want to calculate a discretized distribution function you can use the HISTOGRAM keyword in
-tandem with the RANGE keyword. HISTOGRAM takes as input the number of bins in your distribution and (optionally) the value of the smearing parameter \f$\sigma\f$.
-RANGE takes the upper and lower bound of the histogram. The interval between the upper and lower bound specified using RANGE is then divided into
-equal-width bins.
-
-*/
-//+ENDPLUMEDOC
+void within::writeDocs(std::string& docs){
+ std::ostringstream ostr;
+ ostr<<"\\par WITHIN/HISTOGRAM \n\n";
+ ostr<<"Calculates the number of colvars that are within a certain range. To make this quantity continuous it is calculated using: \n";
+ ostr<<"\\f[\n";
+ ostr<<"S = \\sum_i \\int_a^b G( s_i, \\sigma*(b-a) )";
+ ostr<<"\\f]\n";
+ ostr<<"where \\f$G( s_i, \\sigma )\\f$ is a normalized Gaussian function of width \\f$\\sigma\\f$ centered at the value of the colvar \\f$s_i\\f$. \\f$a\\f$ and \\f$b\\f$ are\n";
+ ostr<<"the lower and upper bounds of the range of interest respectively. The values of \\f$a\\f$ and \\f$b\\f$ must be specified using the WITHIN keyword. If this keyword\n";
+ ostr<<"has three input numbers then the third is assumed to be the value of \\f$\\sigma\\f$. You can specify that you want to investigate multiple rangles by using multiple instances\n";
+ ostr<<"of the WITHIN keyword (WITHIN1,WITHIN2 etc). Alternatively, if you want to calculate a discretized distribution function you can use the HISTOGRAM keyword in \n";
+ ostr<<"tandem with the RANGE keyword. HISTOGRAM takes as input the number of bins in your distribution and (optionally) the value of the smearing parameter \\f$\\sigma\\f$.\n";
+ ostr<<"RANGE takes the upper and lower bound of the histogram. The interval between the upper and lower bound specified using RANGE is then divided into \n";
+ ostr<<"equal-width bins.\n";
+ docs=ostr.str();
+}
within::within( const std::vector<std::string>& parameters ) :
DistributionFunction(parameters)
diff --git a/src/Keywords.cpp b/src/Keywords.cpp
index 978fbacab478fa5353049a19f6f2a65d183bc2f4..3b9cff517da81b69e024d531b4f6f37582aa0277 100644
--- a/src/Keywords.cpp
+++ b/src/Keywords.cpp
@@ -1,4 +1,5 @@
#include "Keywords.h"
+#include "DistributionFunctions.h"
using namespace PLMD;
@@ -213,6 +214,22 @@ void Keywords::print_html() const {
std::cout<<"\n";
}
std::cout<<"</table>\n\n";
+
+ nkeys=0;
+ for(unsigned i=0;i<keys.size();++i){
+ if( types[i].isNoHTML() ) nkeys++;
+ }
+ if( nkeys>0 ){
+ std::cout<<"\\section mod Working with distributions \n\n";
+ std::cout<<"This action can be used to calculate a number of quantities from a distribution of quantities of this type.\n";
+ std::cout<<"Plumed allows you to then calculate functions of the resulting distribution and store\n";
+ std::cout<<"the value of these distributions and the derivatives in the values of the colvar.\n";
+ std::cout<<"The options for doing this for this particular Action are as follows.\n\n";
+ DistributionFunctionDocs ddDoc;
+ for(unsigned i=0;i<keys.size();++i){
+ if( types[i].isNoHTML() ) ddDoc.printDocs(keys[i]);
+ }
+ }
}
void Keywords::print( Log& log ) const {
diff --git a/src/MultiColvar.cpp b/src/MultiColvar.cpp
index 000fc9d143b3d4d166b8fa4c2ed90fca6da3bb34..ac9fdeda9c4eabf839f06252e5ab11da0b8444c9 100644
--- a/src/MultiColvar.cpp
+++ b/src/MultiColvar.cpp
@@ -33,6 +33,15 @@ void MultiColvar::registerKeywords( Keywords& keys ){
"of the atoms specifies using SPECIESB is within the specified cutoff");
keys.reserve("atoms","SPECIESB","this keyword is used for colvars such as the coordination number. It must appear with SPECIESA. For a full explanation see "
"the documentation for that keyword");
+ keys.add("nohtml","MIN","take the minimum value from these variables. The continuous version of the minimum described above is calculated and beta must be specified in input");
+// keys.add("optional","MAX", "take the maximum value from these variables");
+ keys.add("nohtml","SUM", "take the sum of these variables");
+ keys.add("nohtml","AVERAGE", "take the average value of these variables");
+ keys.add("nohtml","LESS_THAN", "take the number of variables less than the specified target. This quantity is made differentiable using a switching function. You can control the parameters of this switching function by specifying three numbers to the keyword (r_0, nn and mm). If you are happy with the default values of nn=6 and mm=12 then you need only specify the target r_0. The number of values less than the target is stored in a value called lt<target>.");
+ keys.add("nohtml","MORE_THAN", "take the number of variables more than the specified target. This quantity is made differentiable using a switching function. You can control the parameters of this switching function by specifying three numbers to the keyword (r_0, nn and mm). If you are happy with the default values of nn=6 and mm=12 then you need only specify the target r_0. The number of values less than the target is stored in a value called gt<target>.");
+ keys.add("nohtml","HISTOGRAM", "create a discretized histogram of the distribution. This keyword's input should consist of one or two numbers");
+ keys.add("nohtml","RANGE", "the range in which to calculate the histogram");
+ keys.add("nohtml", "WITHIN", "The number of values within a certain range. This keyword's input should consist of one or two numbers.");
ActionWithDistribution::registerKeywords( keys );
}
@@ -82,8 +91,72 @@ void MultiColvar::readAtoms( int& natoms ){
}
all_atoms.updateActiveMembers();
- readDistributionKeywords(); // And read the ActionWithDistributionKeywords
- requestAtoms(); // Request the atoms in ActionAtomistic and set up the value sizes
+ // -- Now read in distribution keywords -- //
+ bool dothis; std::vector<std::string> params;
+
+ // Read SUM keyword
+ parseFlag("SUM",dothis);
+ if( dothis && keywords.exists("NL_CHEAT") && isTimeForNeighborListUpdate() ) error("cannot use SUM and neighbor list for this keyword");
+ if( dothis ){
+ addDistributionFunction( "sum", new sum(params) );
+ }
+ // Read AVERAGE keyword
+ parseFlag("AVERAGE",dothis);
+ if( dothis && keywords.exists("NL_CHEAT") && isTimeForNeighborListUpdate() ) error("cannot use AVERAGE and neighbor list for this keyword");
+ if( dothis ){
+ params.resize(1); Tools::convert(getNumberOfFunctionsInDistribution(),params[0]);
+ addDistributionFunction( "average", new mean(params) );
+ }
+ // Read MIN keyword
+ parseVector("MIN",params);
+ if( params.size()!=0 && keywords.exists("NL_CHEAT") && isTimeForNeighborListUpdate() ) error("cannot use MIN and neighbor list for this keyword");
+ if( params.size()!=0 ){
+ addDistributionFunction( "min", new min(params) );
+ }
+ // Read Less_THAN keyword
+ parseVector("LESS_THAN",params);
+ if( params.size()!=0 ){
+ addDistributionFunction( "lt" + params[0], new less_than(params) );
+ }
+ // Read MORE_THAN keyword
+ parseVector("MORE_THAN",params);
+ if( params.size()!=0 ){
+ addDistributionFunction( "gt" + params[0], new more_than(params) );
+ }
+ // Read HISTOGRAM keyword
+ parseVector("HISTOGRAM",params);
+ if( params.size()!=0 ){
+ std::vector<double> range(2); parseVector("RANGE",range);
+ if(range[1]<range[0]) error("range is meaningless");
+ int nbins; Tools::convert(params[0],nbins);
+ std::vector<std::string> hparams(2);
+ if(params.size()==2){
+ hparams.resize(3);
+ hparams[2]=params[1];
+ } else if(params.size()!=1){
+ error("Histogram keyword should either specify just the number"
+ " of bins or the number of bins and the ammount of smearing");
+ }
+ double lb,ub,delr=(range[1]-range[0])/static_cast<double>(nbins);
+ for(int i=0;i<nbins;++i){
+ lb=range[0]+i*delr; Tools::convert( lb, hparams[0] );
+ ub=range[0]+(i+1)*delr; Tools::convert( ub, hparams[1] );
+ addDistributionFunction( "between" + hparams[0] + "&" +hparams[1], new within(hparams) );
+ }
+ }
+ // Read within keywords
+ parseVector("WITHIN",params);
+ if( params.size()!=0 ){
+ addDistributionFunction( "between" + params[0] + "&" +params[1], new within(params) );
+ } else {
+ for(unsigned i=1;;++i){
+ if( !parseNumberedVector("WITHIN",i,params) ) break;
+ addDistributionFunction( "between" + params[0] + "&" +params[1], new within(params) );
+ }
+ }
+
+ requestDistribution(); // And setup the ActionWithDistribution
+ requestAtoms(); // Request the atoms in ActionAtomistic and set up the value sizes
}
void MultiColvar::createNeighborList( std::vector<std::pair<unsigned,unsigned> >& pairs ){
diff --git a/user-doc/extract b/user-doc/extract
index cddbe0bda8ec0e8a57f296909e5fac08231b639d..284a54e351e71a52275ed5fd7fa7b0f78625442b 100755
--- a/user-doc/extract
+++ b/user-doc/extract
@@ -30,28 +30,6 @@ done
# fi
#done
-# Generate all the documentation for colvar modifiers
-rm -f modifier_file
-echo "\\section mod Working with distributions" > modifier_file
-echo "As described above you can use this colvar to calculate one or multiple CVs. Obviously, if you have calculated multiple CVs then you have a distribution of CVs. Plumed allows you to calculate a number of functions of this distribution. These quantities and their derivatives are then stored as the values of the colvar. N.B. You cannot calcualte the values of the colvars and these functions of the distribution from a single line in input. You can however calculate multiple functions of the distribution." >> modifier_file
-echo "<table align=center frame=void width=95%% cellpadding=5%%>" >> modifier_file
-cat ../*/*cpp |
-awk '{
- if($1=="//+ENDPLUMEDOC" && inside==1){
- print "</td> </tr>" >> "modifier_file"
- inside=0
- }
- if(inside && $1!="/**" && $1!="*/") print $0 >> "modifier_file"
- if($1=="//+PLUMEDOC"){
- if($2=="MODIFIER"){
- print "<tr> <td width=5%> <b> ", $3, " </b></td>" >> "modifier_file"
- print "<td>" >> "modifier_file"
- inside=1;
- }
- }
-}'
-echo "</table>" >> modifier_file
-
# Generate all the documentation pages from comments
cat ../*/*cpp |
awk '{
@@ -104,14 +82,6 @@ do
if [ "$LINE" = "\par Examples" ]
then
cat automatic/$flag.man
- iscol=`grep $flag automatic/COLVAR.list | wc -l`
- if [ $iscol -gt 0 ] ; then
- fname=`grep PLUMED_REGISTER_ACTION ../src/*.cpp | grep $flag | sed -e 's/:/ /g' | awk '{print $1}'`
- hasmod=`grep MultiColvar $fname | wc -l | awk '{print $1}'`
- if [ $hasmod -gt 0 ] ; then
- cat modifier_file
- fi
- fi
echo -E "$LINE"
else
echo -E "$LINE"