source: BAORadio/AmasNancay/trunk/mergeRawOnOff.cc @ 552

// Utilisation de SOPHYA pour faciliter les tests ...
#include <regex.h>
//#include <regexp.h>
#include <stdio.h>

#include "sopnamsp.h"
#include "machdefs.h"

#include <stdlib.h>
#include <dirent.h>
#include <matharr.h>

// include standard c/c++
#include <iostream>
#include <fstream>
#include <string>
#include <vector>
#include <map>
#include <functional>
#include <algorithm>
#include <numeric>
#include <list>
#include <exception>

// include sophya mesure ressource CPU/memoire ...
#include "resusage.h"
#include "ctimer.h"
#include "timing.h"
#include "timestamp.h"
#include "strutilxx.h"
#include "ntuple.h"
#include "fioarr.h"
#include "tarrinit.h"
#include "histinit.h"
#include "fitsioserver.h"
#include "fiosinit.h"
#include "ppersist.h"

//-----------------------------------------------
const sa_size_t NUMBER_OF_CHANNELS = 2;
const sa_size_t NUMBER_OF_FREQ     = 8192;
const sa_size_t TOTAL_NUM_CYCLES   = 10000;
const r_4    LOWER_FREQUENCY = 1250.0; //MHz
const r_4    TOTAL_BANDWIDTH = 250.0; //MHz
//-----------------------------------------------
//Input parameters
struct Param {
  int debuglev_;      //debug
  string inPath_;     //root directory of the input files
  string outPath_;    //output files are located here
  string sourceName_; //source name & subdirectory of the input files
  string ppfFile_;    //generic name of the input files
  int nSliceInFreq_;  //used by reduceSpectra() fnc
} para;
//--------------------------------------------------------------
//Utility functions

  class median_of_empty_list_exception:public std::exception{
    virtual const char* what() const throw() {
    return "Attempt to take the median of an empty list of numbers.  "
      "The median of an empty list is undefined.";
  }
  };
  template<class RandAccessIter>
    double median(RandAccessIter begin, RandAccessIter end) 
    throw(median_of_empty_list_exception){
    if(begin == end){ throw median_of_empty_list_exception(); }
    std::size_t size = end - begin;
    std::size_t middleIdx = size/2;
    RandAccessIter target = begin + middleIdx;
    std::nth_element(begin, target, end);
    
    if(size % 2 != 0){ //Odd number of elements
      return *target;
    }else{            //Even number of elements
      double a = *target;
      RandAccessIter targetNeighbor= target-1;
      std::nth_element(begin, targetNeighbor, end);
      return (a+*targetNeighbor)/2.0;
    }
  }

//--------------------------------------------------------------
char *regexp (const char *string, const char *patrn, int *begin, int *end) {    
        int i, w=0, len;                 
        char *word = NULL;
        regex_t rgT;
        regmatch_t match;
        regcomp(&rgT,patrn,REG_EXTENDED);
        if ((regexec(&rgT,string,1,&match,0)) == 0) {
                *begin = (int)match.rm_so;
                *end = (int)match.rm_eo;
                len = *end-*begin;
                word=(char*)malloc(len+1);
                for (i=*begin; i<*end; i++) {
                        word[w] = string[i];
                        w++; }
                word[w]=0;
        }
        regfree(&rgT);
        return word;
}
//------- 
sa_size_t round_sa(r_4 r) {
  return static_cast<sa_size_t>((r > 0.0) ? (r + 0.5) : (r - 0.5));
}
//-----
string StringToLower(string strToConvert){
  //change each element of the string to lower case
  for(unsigned int i=0;i<strToConvert.length();i++) {
    strToConvert[i] = tolower(strToConvert[i]);
  }
  return strToConvert;//return the converted string
}
//-----
bool stringCompare( const string &left, const string &right ){
   if( left.size() < right.size() )
      return true;
   for( string::const_iterator lit = left.begin(), rit = right.begin(); lit != left.end() && rit != right.end(); ++lit, ++rit )
      if( tolower( *lit ) < tolower( *rit ) )
         return true;
      else if( tolower( *lit ) > tolower( *rit ) )
         return false;
   return false;
}
//-----
list<string> ListOfFileInDir(string dir, string filePettern) throw(string) {
  list<string> theList;


  DIR *dip;
  struct dirent *dit;
  string msg;  string fileName;
  string fullFileName;
  size_t found;

  if ((dip=opendir(dir.c_str())) == NULL ) {
    msg = "opendir failed on directory "+dir;
    throw msg;
  }
  while ( (dit = readdir(dip)) != NULL ) {
    fileName = dit->d_name;
    found=fileName.find(filePettern);
    if (found != string::npos) {
      fullFileName = dir + "/";
      fullFileName += fileName;
      theList.push_back(fullFileName);
    }
  }//eo while
  if (closedir(dip) == -1) {
    msg = "closedir failed on directory "+dir;
    throw msg;
  }
  
  theList.sort(stringCompare);

  return theList;

}
//
class  StringMatch : public unary_function<string,bool> {
public:
  StringMatch(const string& pattern): pattern_(pattern){}
  bool operator()(const string& aStr) const {


    int b,e;
    regexp(aStr.c_str(),pattern_.c_str(),&b,&e);

//     cout << "investigate " << aStr << " to find " << pattern_ 
//       << "[" <<b<<","<<e<<"]" 
//       << endl;

    
    if (b != 0) return false;
    if (e != (int)aStr.size()) return false;
    return true;

  }
private:
  string pattern_;
};
//-------------------------------------------------------
//Rebin in frequence + compute mean and sigma
void reduceSpectra(const TMatrix<r_4>& specMtxInPut, 
                    TMatrix<r_4>& meanMtx, 
                    TMatrix<r_4>& sigmaMtx) {
  sa_size_t nSliceFreq = para.nSliceInFreq_;
  sa_size_t deltaFreq =  NUMBER_OF_FREQ/nSliceFreq;
  for (sa_size_t iSlice=0; iSlice<nSliceFreq; iSlice++){
    sa_size_t freqLow= iSlice*deltaFreq;
    sa_size_t freqHigh= freqLow + deltaFreq -1;
    for (sa_size_t iCh=0; iCh<NUMBER_OF_CHANNELS; ++iCh){
      TVector<r_4> reducedRow;
      reducedRow = specMtxInPut.SubMatrix(Range(iCh),Range(freqLow,freqHigh)).CompactAllDimensions();
      double mean; 
      double sigma;
      MeanSigma(reducedRow,mean,sigma);
      meanMtx(iCh,iSlice) = mean;
      sigmaMtx(iCh,iSlice) = sigma;
    }//eo loop on channels
  }//eo loop on slices
}
//-------------------------------------------------------
//Compute the mean of Raw spectra and also the mean/sigma of rebinned spectra 
//Used like:
//
void meanOnCycles() throw(string) {
  list<string> listOfFiles;
  string directoryName;
  directoryName = para.inPath_ + "/" + para.sourceName_;

  //Make the listing of the directory
  listOfFiles = ListOfFileInDir(directoryName,para.ppfFile_);
  
  list<string>::const_iterator iFile, iFileEnd, iSpec, iSpecEnd;
  iFileEnd = listOfFiles.end();
  
  StringMatch match("specONOFFRaw[0-9]+"); //Tag of the PPF objects
  TMatrix<r_4> meanOfSpectra(NUMBER_OF_CHANNELS,NUMBER_OF_FREQ);
  uint_4 nSpectra=0;
  //Loop on files
  for (iFile = listOfFiles.begin(); iFile != iFileEnd; ++iFile) {
    if (para.debuglev_>90){
      cout << "load file <" << *iFile << ">" << endl;
    }
    PInPersist fin(*iFile);
    vector<string> vec = fin.GetNameTags();
    list<string> listOfSpectra;
    //Keep only required PPF objects
    std::remove_copy_if(
                        vec.begin(), vec.end(), back_inserter(listOfSpectra),
                        not1(match)
                        );
   
    iSpecEnd = listOfSpectra.end();
    listOfSpectra.sort(stringCompare);
    //Loop of spectra matrix
    for (iSpec = listOfSpectra.begin(); iSpec !=iSpecEnd;  ++iSpec){
      if (para.debuglev_>90){
        cout << " spactra <" << *iSpec << ">" << endl;
      }
      TMatrix<r_4> aSpec(NUMBER_OF_CHANNELS,NUMBER_OF_FREQ);
      fin.GetObject(aSpec,*iSpec);
      //How to see if the GetObject is ok?? Ask Reza
      nSpectra++;
      meanOfSpectra+=aSpec;
    }//eo loop on spectra in a file
  }//eo loop on files
  
  //Normalisation
  if(nSpectra>0)meanOfSpectra/=(r_4)(nSpectra);

  //Compute the reduced version of the mean and sigma
  TMatrix<r_4> meanRedMtx(NUMBER_OF_CHANNELS,para.nSliceInFreq_);
  TMatrix<r_4> sigmaRedMtx(NUMBER_OF_CHANNELS,para.nSliceInFreq_);
  reduceSpectra(meanOfSpectra,meanRedMtx,sigmaRedMtx);

  {//Save the result
    stringstream tmp;
    tmp << nSpectra;
    string fileName = para.outPath_+"/meanDiffOnOffRaw_"+StringToLower(para.sourceName_)+"-"+tmp.str()+"Cycles.ppf";
    cout << "Save mean based on " <<  nSpectra << " cycles " << endl;
    POutPersist fos(fileName);

    string tag = "mean";
    fos << PPFNameTag(tag) << meanOfSpectra;
    tag = "meanred";
    fos << PPFNameTag(tag) << meanRedMtx;
    tag = "sigmared";
    fos << PPFNameTag(tag) << sigmaRedMtx;
  }
}
//-------------------------------------------------------
//Compute the median of Raw spectra and also the mean/sigma of rebinned spectra 
//Used like:
//
void medianOnCycles() throw(string) {
  list<string> listOfFiles;
  string directoryName;
  directoryName = para.inPath_ + "/" + para.sourceName_;

  //Make the listing of the directory
  listOfFiles = ListOfFileInDir(directoryName,para.ppfFile_);
  
  list<string>::const_iterator iFile, iFileEnd, iSpec, iSpecEnd;
  iFileEnd = listOfFiles.end();
  

  TArray<r_4> tableOfSpectra(NUMBER_OF_FREQ,NUMBER_OF_CHANNELS,TOTAL_NUM_CYCLES); //TOTAL_NUM_CYCLES should be large enough...

  StringMatch match("specONOFFRaw[0-9]+"); //Tag of the PPF objects
  uint_4 nSpectra=0;
  //Loop on files
  for (iFile = listOfFiles.begin(); iFile != iFileEnd; ++iFile) {
    if (para.debuglev_>90){
      cout << "load file <" << *iFile << ">" << endl;
    }
    PInPersist fin(*iFile);
    vector<string> vec = fin.GetNameTags();
    list<string> listOfSpectra;
    //Keep only required PPF objects
    std::remove_copy_if(
                        vec.begin(), vec.end(), back_inserter(listOfSpectra),
                        not1(match)
                        );
   
    iSpecEnd = listOfSpectra.end();
    listOfSpectra.sort(stringCompare);
    //Loop of spectra matrix
    for (iSpec = listOfSpectra.begin(); iSpec !=iSpecEnd;  ++iSpec){
      if (para.debuglev_>90){
        cout << " spactra <" << *iSpec << ">" << endl;
      }
      TMatrix<r_4> aSpec(NUMBER_OF_CHANNELS,NUMBER_OF_FREQ);
      fin.GetObject(aSpec,*iSpec);

      if ((sa_size_t)nSpectra < TOTAL_NUM_CYCLES ) {
        tableOfSpectra(Range::all(),Range::all(),Range(nSpectra)) = aSpec;
      } else {
        stringstream tmp;
        tmp<<nSpectra;
        throw "FATAL: SIZE ERROR: too many cycles:"+tmp.str();
      }
      nSpectra++;
    }//eo loop on spectra in a file
  }//eo loop on files
  
  //Resize

 
  TMatrix<r_4> medianOfSpectra(NUMBER_OF_CHANNELS,NUMBER_OF_FREQ);
  //Compute the median for each freq. and Channel
  for (sa_size_t iCh=0; iCh<NUMBER_OF_CHANNELS; iCh++){
    for (sa_size_t freq =0; freq<NUMBER_OF_FREQ; freq++){
      TVector<r_4> tmp0(tableOfSpectra(Range(freq),Range(iCh),Range(0,nSpectra-1)).CompactAllDimensions());
      vector<r_4> tmp;
      tmp0.FillTo(tmp);
      medianOfSpectra(iCh,freq) = median(tmp.begin(),tmp.end());
    }
  }


  //Compute the reduced version of the mean and sigma
  TMatrix<r_4> meanRedMtx(NUMBER_OF_CHANNELS,para.nSliceInFreq_);
  TMatrix<r_4> sigmaRedMtx(NUMBER_OF_CHANNELS,para.nSliceInFreq_);
  reduceSpectra(medianOfSpectra,meanRedMtx,sigmaRedMtx);

  {//Save the result
    stringstream tmp;
    tmp << nSpectra;
    string fileName = para.outPath_+"/medianDiffOnOffRaw_"+StringToLower(para.sourceName_)+"-"+tmp.str()+"Cycles.ppf";
    cout << "Save median based on " <<  nSpectra << " cycles " << endl;
    POutPersist fos(fileName);

    string tag = "median";
    fos << PPFNameTag(tag) << medianOfSpectra;
    tag = "meanmedred";
    fos << PPFNameTag(tag) << meanRedMtx;
    tag = "sigmamedred";
    fos << PPFNameTag(tag) << sigmaRedMtx;
    tag = "cycleVsfreq";
    
    TArray<r_4> tarr(tableOfSpectra(Range::all(),Range::all(),Range(0,nSpectra-1)));
    fos << PPFNameTag(tag) << tarr;
  }
}

//-------------------------------------------------------
int main(int narg, char* arg[]) {
  
  int rc = 0; //return code
  string msg; //message used in Exceptions



  //default value for initial parameters (see Para structure on top of the file)
  string debuglev = "0";
  string action = "meanDiffOnOff";
  string inputPath = "."; 
  string outputPath = "."; 
  string sourceName = "Abell85";
  string ppfFile;
  string nSliceInFreq = "32";
 
  
  //  bool okarg=false;
  int ka=1;
  while (ka<(narg-1)) {
    if (strcmp(arg[ka],"-h")==0) {
      cout << "Usage:  -act [meanDiffOnOff]|medianDiffOnOff\n"
           << " -src [Abell85]\n -inPath [.]|<top_root_dir of the ppf file>\n" 
           << " (ex. /sps/baoradio/AmasNancay/JEC/\n " 
           << " -outPath [.]|<dir of the output> \n"
           << " -nSliceInFreq [32]|<number of bin in freq. to cumulate>\n"
           << " -ppfFile <generic name of the input ppf files> (ex. diffOnOffRaw)"
           << " -debug <level> "
           << endl;
      return 0;
    }
    else if (strcmp(arg[ka],"-debug")==0) {
      debuglev=arg[ka+1];
      ka+=2;
    }
    else if (strcmp(arg[ka],"-act")==0) {
      action=arg[ka+1];
      ka+=2;
    }
    else if (strcmp(arg[ka],"-inPath")==0) {
      inputPath=arg[ka+1];
      ka+=2;
    }
    else if (strcmp(arg[ka],"-outPath")==0) {
      outputPath=arg[ka+1];
      ka+=2;
    }
    else if (strcmp(arg[ka],"-src")==0) {
      sourceName=arg[ka+1];
      ka+=2;
    }
    else if (strcmp(arg[ka],"-ppfFile")==0) {
      ppfFile=arg[ka+1];
      ka+=2;
    }
    else if (strcmp(arg[ka],"-nSliceInFreq")==0) {
      nSliceInFreq=arg[ka+1];
      ka+=2;
    }
    else ka++;
  }//eo while

  para.debuglev_   = atoi(debuglev.c_str());
  para.inPath_     = inputPath;
  para.outPath_    = outputPath;
  para.sourceName_ = sourceName;
  para.ppfFile_    = ppfFile;
  para.nSliceInFreq_ = atoi(nSliceInFreq.c_str());

  cout << "Dump Initial parameters ............" << endl;
  cout << " action = " << action << "\n"
       << " inputPath = " << para.inPath_  << "\n" 
       << " outputPath = " <<  para.outPath_ << "\n"
       << " sourceName = " << para.sourceName_ << "\n"
       << " ppfFile = " <<  para.ppfFile_ << "\n"
       << " nSliceInFreq = " << para.nSliceInFreq_  << "\n"
       << " debuglev = "  << para.debuglev_   << "\n";
  cout << "...................................." << endl;

  if ( "" == ppfFile ) {
    cerr << "mergeRawOnOff.cc: you have forgotten ppfFile option"
         << endl;
    return 999;
  }


  try {

//     int b,e;
//     char *match=regexp("truc0machin","[a-z]+[0-9]*",&b,&e);
//     printf("->%s<-\n(b=%d e=%d)\n",match,b,e);

    if ( action == "meanDiffOnOff" ) {
      meanOnCycles();
    } else if (action == "medianDiffOnOff") {
      medianOnCycles();
    } else {
      msg = "Unknown action " + action;
      throw msg;
    }


  }  catch (std::exception& sex) {
    cerr << "mergeRawOnOff.cc std::exception :"  << (string)typeid(sex).name() 
         << "\n msg= " << sex.what() << endl;
    rc = 78;
  }
  catch ( string str ) {
    cerr << "mergeRawOnOff.cc Exception raised: " << str << endl;
  }
  catch (...) {
    cerr << "mergeRawOnOff.cc catched unknown (...) exception  " << endl; 
    rc = 79; 
  } 

  return 0;

}