root/applications/bdmtoolbox/mex/estimator.cpp @ 1344

/*!
\file
\brief Application Estimator

The general task of estimation is defined on the following scheme:
\dot
digraph estimation{
        node [shape=box];
        {rank="same"; "Data Source"; "Bayesian Model"}
        "Data Source" -> "Bayesian Model" [label="data"];
        "Bayesian Model" -> "Result Logger" [label="estimated\n statistics"];
        "Data Source" -> "Result Logger" [label="Simulated\n data"];
}
\enddot

Here,
\li Data Source is an object (class DS) providing sequential data, \f$ [d_1, d_2, \ldots d_t] \f$.
\li Bayesian Model is an object (class BM) performing Bayesian filtering,
\li Result Logger is an object (class logger) dedicated to storing important data from the experiment.

\section  cmd Command-line usage
Execute command:
\code
$> estimator config_file.cfg
\endcode

Full description of the experiment is in the file config_file.cfg which is expected to have the following structure:
\code
system = {type = "DS_offspring", ...};      // definition of a data source
estimator = {type = "BM_offspring", ...};   // definition of an estimator
logger = {type = "logger_type",...};        // definition of a logger
experiment = {ndat = 11000; };              // definition of number of data records
\endcode

The above description must be specialized to specific classes. See, \subpage arx_ui how to do it for estimation of an ARX model.

\section ex Matlab usage
Execute command:
\code
>> estimator
\endcode
and follow the help there.

 */

#include <estim/arx_ext.h>
#include <stat/emix.h>
#include <base/datasources.h>
#include <base/loggers.h>
#include <estim/particles.h>
#include <estim/kalman.h>

//#include "mex_datasource.h"

using namespace bdm;

#ifdef MEX
#include <itpp/itmex.h>
#include <mex/mex_BM.h>
#include <mex/mex_logger.h>
#include <mex/mex_datasource.h>
#include <mex/mex_function.h>

void mexFunction ( int n_output, mxArray *output[], int n_input, const mxArray *input[] ) {
        // Check the number of inputs and output arguments
        if ( n_input<2 ) mexErrMsgTxt ( "Usage:\n"
                                                "[Res,estimators,Res2]=estimator(system, estimators, experiment, logger)\n"
                                                "  system     = struct('class','datasource',...);  % Estimated system\n"
                                                "  estimators = {struct('class','estimator',...),  % Estimators\n"
                                                "                struct('class','estimator',...),...} \n"
                                                "  === optional ==="
                                                "  experiment = struct('ndat',10);                 % number of data in experiment, full length of finite datasources, 10 otherwise \n"
                                                "  logger     = struct('class','mexlogger');       % How to store results, default=mexlog, i.e. matlab structure\n\n"
                                                "Output:\n"
                                                "  Res          Matlab structure with logged results, \n"  
                                                "  estimators   Array of estimators updated with data \n"
                                                "  Res2         When logfull log_level is on, this structure is filled with structures of posterior densities\n\n"
                                                "see documentation of classes datasource, BM, and mexlogger and their offsprings in BDM." );

        RV::clear_all();
        //CONFIG
        UImxArray Cfg;
        try {
                Cfg.addGroup ( input[0],"system" );
                Cfg.addList ( input[1],"estimators" );
                if ( n_input>2 ) {
                        Cfg.addGroup ( input[2],"experiment" );
                }
                if ( n_input>3 ) {
                        Cfg.addGroup ( input[3],"logger" );
                }
        } catch ( SettingException e ) {
                it_error ( "error: "+string ( e.getPath() ) );
        }

        //DBG
        Cfg.writeFile ( "estimator.cfg" );

#else
int main ( int argc, char* argv[] ) {
        const char *fname;
        if ( argc>1 ) {
                fname = argv[1];
        } else {
                fname="estimator.cfg";
        }
        UIFile Cfg ( fname );
#endif

        shared_ptr<DS> Ds = UI::build<DS> ( Cfg, "system" );
        Array<shared_ptr<BM> > Es;
        UI::get ( Es,Cfg, "estimators" );
        int Ndat=10;
        if ( Cfg.exists ( "experiment" ) ) {
                if (  UI::get(Ndat, Cfg.getRoot()["experiment"],"ndat" ) ) {
                        bdm_assert ( Ndat<=Ds->max_length(), "Data source has less data then required" );
                };
        } else {
                if ( Ds->max_length() < std::numeric_limits< int >::max() ) {
                        Ndat=Ds->max_length();
                }
                ;// else Ndat=10;
        }
        shared_ptr<logger> L = UI::build<logger> ( Cfg, "logger",UI::optional );
        if ( !L ) {
#ifdef MEX
                //mex logger has only from_setting constructor - we  have to fill it...
                L=new mexlog ( Ndat );
#else
                L=new stdlog();
#endif
        }

        Ds->log_register ( *L, "DS" );
        string Ename;
        Setting &S=Cfg;
        for ( int i=0; i<Es.length(); i++ ) {
                if (!UI::get ( Ename, S["estimators"][i], "name" ,UI::optional)){
                        Ename="Est"+num2str ( i );
                }
                if (!Es(i)->posterior().log_level[epdf::logmean]){
                        const_cast<epdf&>(Es (i) ->posterior()).log_level[epdf::logmean] =true;
                }               
                Es ( i )->log_register ( *L,Ename ); // estimate
        }
        L->init();

        vec dt=zeros ( Ds->_drv()._dsize() );   //data variable
        Array<datalink*> Dls ( Es.length() );
        Array<datalink*> Dlsc ( Es.length() );
        Array<datalink_buffered*> Dls_buf (0);
        for ( int i=0; i<Es.length(); i++ ) {
                //check if they are properly named
                bdm_assert(Es(i)->_yrv()._dsize() == Es(i)->dimensiony(), "Estimator["+num2str(i)+"] does not name yrv properly."
                "size(yrv)="+num2str(Es(i)->_yrv()._dsize() ) + ", declared dimension of y="+num2str(Es(i)->dimensiony()));
                bdm_assert(Es(i)->_rvc()._dsize() == Es(i)->dimensionc(), "Estimator["+num2str(i)+"] does not name rvc properly."
                "size(rvc)="+num2str(Es(i)->_rvc()._dsize() ) + ", declared dimension of rvc="+num2str(Es(i)->dimensionc()));
                //connect actual data
                Dls ( i ) = new datalink ( Es ( i )->_yrv(),Ds->_drv() ); //datalink between a datasource and estimator
                //connect data in condition
                if (Es ( i )->_rvc().mint()<0){ 
                        //delayed values are required
                        
                        //create delayed dl
                        int ith_buf=Dls_buf.size();
                        Dls_buf.set_size( ith_buf + 1, true);
                        Dls_buf(ith_buf) = new datalink_buffered(); 
                        //add dl to list of buffered DS
                        Dlsc(i) = Dls_buf(ith_buf);
                        Dlsc(i)->set_connection ( Es ( i )->_rvc(),Ds->_drv() ); //datalink between a datasource and estimator
                        
                        bdm_assert_debug(Dlsc(i)->_downsize() == Es ( i )->_rvc()._dsize(), "Data required by Est[" + num2str(i) + "], " + 
                        Es(i)->_rvc().to_string() + ", are not available in DS drv:" + Ds->_drv().to_string(););
                } else {
                        Dlsc ( i ) = new datalink ( Es ( i )->_rvc(),Ds->_drv() ); //datalink between a datasource and estimator
                }
        }

        for ( int tK=0;tK<Ndat;tK++ ) {
                Ds->getdata ( dt );                                     // read data
                Ds->log_write ( );

                for ( int i=0; i<Es.length(); i++ ) {
                        if (tK + Es ( i )->_rvc().mint() > 0 ) {
                                Es ( i )->bayes ( Dls ( i )->pushdown ( dt ), Dlsc(i) ->pushdown(dt) );         // update estimates
                        }
                        Es ( i )->log_write ();
                }
                
                L->step();
                Ds->step();                                                     // simulator step
                //update buffered fdat links
                for (int i=0; i<Dls_buf.length(); i++){
                        Dls_buf(i)->store_data(dt);
                }
                        
        }

        L->finalize();
        // ------------------ End of routine -----------------------------

#ifdef MEX
        mexlog* mL=dynamic_cast<mexlog*> ( L.get() );

        if ( mL ) { // user wants output!!
                if ( n_output<1 ) mexErrMsgTxt ( "Wrong number of output variables!" );
                output[0] = mL->toCell();
                if (n_output>1){ // write estimators
                        UImxArray Ests;
                        UI::save(Es, Ests,"estimators");
                        output[1]=UImxArray::create_mxArray(Ests);
                }
                if (n_output>2) {
                        mL->_setting_conf().setAutoConvert(true);
                        output[2]= UImxArray::create_mxArray(mL->_setting_conf().getRoot());
                }
        }
#endif
        for (int i=0;i<Dls.length(); i++){delete Dls(i); delete Dlsc(i);}
        UIFile F;
        UI::save(Es, F.getRoot(),"estimators");
        F.writeFile("estim_out.cfg");
}