source: trunk/MagicSoft/Mars/mjtrain/MJTrainEnergy.cc@ 8675

/* ======================================================================== *\
!
! *
! * This file is part of MARS, the MAGIC Analysis and Reconstruction
! * Software. It is distributed to you in the hope that it can be a useful
! * and timesaving tool in analysing Data of imaging Cerenkov telescopes.
! * It is distributed WITHOUT ANY WARRANTY.
! *
! * Permission to use, copy, modify and distribute this software and its
! * documentation for any purpose is hereby granted without fee,
! * provided that the above copyright notice appear in all copies and
! * that both that copyright notice and this permission notice appear
! * in supporting documentation. It is provided "as is" without express
! * or implied warranty.
! *
!
!
!   Author(s): Thomas Bretz 11/2005 <mailto:tbretz@astro.uni-wuerzburg.de>
!
!   Copyright: MAGIC Software Development, 2005
!
!
\* ======================================================================== */

/////////////////////////////////////////////////////////////////////////////
//
// MJTrainEnergy
//
//
// Example:
// --------
//
//    // SequencesOn are used for training, SequencesOff for Testing
//    MDataSet set("mctesttrain.txt");
//    set.SetNumAnalysis(1);  // Must have a number
//    MJTrainEnergy opt;
//    //opt.SetDebug();
//    opt.AddParameter("MHillas.fSize");
//    opt.AddParameter("MHillasSrc.fDist");
//    MStatusDisplay *d = new MStatusDisplay;
//    opt.SetDisplay(d);
//    opt.AddPreCut("MHillasSrc.fDCA*MGeomCam.fConvMm2Deg<0.3");
//    opt.Train("rf-energy.root", set, 30000); // Number of train events
//
// Random Numbers:
// ---------------
//   Use:
//         if(gRandom)
//             delete gRandom;
//         gRandom = new TRandom3();
//   in advance to change the random number generator.
//
////////////////////////////////////////////////////////////////////////////
#include "MJTrainEnergy.h"

#include "MHMatrix.h"

#include "MLog.h"
#include "MLogManip.h"

// tools
#include "MDataSet.h"
#include "MTFillMatrix.h"
#include "MStatusDisplay.h"

// eventloop
#include "MParList.h"
#include "MTaskList.h"
#include "MEvtLoop.h"

// tasks
#include "MReadMarsFile.h"
#include "MContinue.h"
#include "MFillH.h"
#include "MRanForestCalc.h"

// histograms
#include "MHEnergyEst.h"

// filter
#include "MFilterList.h"

ClassImp(MJTrainEnergy);

using namespace std;

Bool_t MJTrainEnergy::Train(const char *out, const MDataSet &set, Int_t num)
{
    SetTitle(Form("TrainEnergy: %s", out));

    if (fDisplay)
        fDisplay->SetTitle(fTitle);

    if (!set.IsValid())
    {
        *fLog << err << "ERROR - DataSet invalid!" << endl;
        return kFALSE;
    }

    if (!HasWritePermission(out))
        return kFALSE;

    *fLog << inf;
    fLog->Separator(GetDescriptor());

    // --------------------- Setup files --------------------
    MReadMarsFile readtrn("Events");
    MReadMarsFile readtst("Events");
    readtrn.DisableAutoScheme();
    readtst.DisableAutoScheme();

    if (!set.AddFilesOn(readtrn))
    {
        *fLog << err << "ERROR - Adding SequencesOn." << endl;
        return kFALSE;
    }
    if (!set.AddFilesOff(readtst))
    {
        *fLog << err << "ERROR - Adding SequencesOff." << endl;
        return kFALSE;
    }

    // ----------------------- Setup Matrix ------------------
    MHMatrix train("Train");
    train.AddColumns(fRules);
    if (fEnableWeights)
        train.AddColumn("MWeight.fVal");
    train.AddColumn("MMcEvt.fImpact/100");
    train.AddColumn("MMcEvt.fTelescopeTheta*TMath::RadToDeg()");
    train.AddColumn("MMcEvt.fEnergy");


    // ----------------------- Fill Matrix RF ----------------------
    MTFillMatrix fill(fTitle);
    fill.SetDisplay(fDisplay);
    fill.SetLogStream(fLog);
    fill.SetDestMatrix1(&train, num);
    fill.SetReader(&readtrn);
    fill.AddPreCuts(fPreCuts);
    fill.AddPreCuts(fTrainCuts);
    fill.AddPreTasks(fPreTasks);
    fill.AddPostTasks(fPostTasks);
    if (!fill.Process())
        return kFALSE;

    // ------------------------ Train RF --------------------------
    MRanForestCalc rf("TrainEnergy", fTitle);
    rf.SetNumTrees(fNumTrees);
    rf.SetNdSize(fNdSize);
    rf.SetNumTry(fNumTry);
    rf.SetNumObsoleteVariables(3);
    rf.SetLastDataColumnHasWeights(fEnableWeights);
    rf.SetDisplay(fDisplay);
    rf.SetLogStream(fLog);
    rf.SetFileName(out);
    rf.SetDebug(fDebug>1);
    rf.SetNameOutput("MEnergyEst");

    /*
    MBinning b(32, 10, 100000, "BinningEnergyEst", "log");

    if (!rf.TrainMultiRF(train, b.GetEdgesD()))    // classification with one tree per bin
        return;

    if (!rf.TrainSingleRF(train, b.GetEdgesD()))   // classification into different bins
        return;
    */
    if (!rf.TrainRegression(train))                  // regression (best choice)
        return kFALSE;

    // --------------------- Display result ----------------------

    gLog.Separator("Test");

    MParList  plist;
    MTaskList tlist;
    plist.AddToList(this);
    plist.AddToList(&tlist);
    //plist.AddToList(&b);

    MFilterList list;
    if (!list.AddToList(fPreCuts))
        *fLog << err << "ERROR - Calling MFilterList::AddToList for fPreCuts failed!" << endl;
    if (!list.AddToList(fTestCuts))
        *fLog << err << "ERROR - Calling MFilterList::AddToList for fTestCuts failed!" << endl;

    MContinue cont(&list);
    cont.SetInverted();

    MHEnergyEst hist;
    MFillH fillh(&hist);
    if (fEnableWeights)
        fillh.SetWeight();

    tlist.AddToList(&readtst);
    tlist.AddToList(fPreTasks);
    tlist.AddToList(&cont);
    tlist.AddToList(&rf);
    tlist.AddToList(fPostTasks);
    tlist.AddToList(&fillh);

    MEvtLoop loop(fTitle);
    loop.SetLogStream(fLog);
    loop.SetDisplay(fDisplay);
    loop.SetParList(&plist);
    //if (!SetupEnv(loop))
    //   return kFALSE;

    if (!loop.Eventloop())
        return kFALSE;

    SetPathOut(out);
    if (!WriteDisplay(0, "UPDATE"))
        return kFALSE;

    return kTRUE;
}