CMS 3D CMS Logo

 All Classes Namespaces Files Functions Variables Typedefs Enumerations Enumerator Properties Friends Macros Groups Pages
EcalUncalibRecHitFixedAlphaBetaAlgo.h
Go to the documentation of this file.
1 #ifndef RecoLocalCalo_EcalRecAlgos_EcalUncalibRecHitFixedAlphaBetaAlgo_HH
2 #define RecoLocalCalo_EcalRecAlgos_EcalUncalibRecHitFixedAlphaBetaAlgo_HH
3 
14 #include "CLHEP/Matrix/Vector.h"
15 #include "CLHEP/Matrix/SymMatrix.h"
16 //#include "CLHEP/Matrix/Matrix.h"
18 
21 
22 #include <vector>
23 #include <string>
24 
25 //#include "TROOT.h"
26 //#include "TMinuit.h"
27 //#include "TGraph.h"
28 //#include "TF1.h"
29 //#include "TMatrixD.h"
30 //#include "TVectorD.h"
31 
32 template <class C>
34 private:
35  double MinAmpl_;
37  double fAlpha_; //parameter of the shape
38  double fBeta_; //parameter of the shape
39  double fAmp_max_; // peak amplitude
40  double fTim_max_; // time of the peak (in 25ns units)
41  double fPed_max_; // pedestal value
42  double alfabeta_;
43 
44  int fNb_iter_;
47 
48  float fSigma_ped;
49  double un_sur_sigma;
50  //temporary solution for different alpha and beta
51  float alpha_table_[36][1701];
52  float beta_table_[36][1701];
53 
54  bool doFit_;
55 
56  double pulseShapeFunction(double t);
57  float PerformAnalyticFit(double* samples, int max_sample);
58  void InitFitParameters(double* samples, int max_sample);
59  CLHEP::HepSymMatrix DM1_;
60  CLHEP::HepVector temp_;
61 
62 public:
64  : fAlpha_(0.),
65  fBeta_(0.),
66  fAmp_max_(-1.),
67  fTim_max_(-1),
68  fPed_max_(0),
69  alfabeta_(0),
70  fNb_iter_(4),
73  fSigma_ped(1.1),
74  DM1_(3),
75  temp_(3) {
76  un_sur_sigma = 1. / double(fSigma_ped);
77  for (int i = 0; i < 36; i++) {
78  for (int j = 0; j < 1701; j++) {
79  alpha_table_[i][j] = 1.2;
80  beta_table_[i][j] = 1.7;
81  }
82  }
83  doFit_ = false;
84  MinAmpl_ = 16;
85  dyn_pedestal = true;
86  }
87  EcalUncalibRecHitFixedAlphaBetaAlgo<C>(int n_iter, int n_bef_max = 1, int n_aft_max = 3, float sigma_ped = 1.1)
88  : fAlpha_(0.), fBeta_(0.), fAmp_max_(-1.), fTim_max_(-1), fPed_max_(0), alfabeta_(0), DM1_(3), temp_(3) {
89  fNb_iter_ = n_iter;
90  fNum_samp_bef_max_ = n_bef_max;
91  fNum_samp_after_max_ = n_aft_max;
92  fSigma_ped = sigma_ped;
93  un_sur_sigma = 1. / double(fSigma_ped);
94  for (int i = 0; i < 36; i++) {
95  for (int j = 0; j < 1701; j++) {
96  alpha_table_[i][j] = 1.2;
97  beta_table_[i][j] = 1.7;
98  }
99  }
100  doFit_ = false;
101  MinAmpl_ = 16;
102  dyn_pedestal = true;
103  };
104 
106  EcalUncalibratedRecHit makeRecHit(const C& dataFrame,
107  const double* pedestals,
108  const double* gainRatios,
110  const EcalWeightSet::EcalChi2WeightMatrix** chi2Matrix) override;
111  void SetAlphaBeta(double alpha, double beta);
112  void SetMinAmpl(double ampl);
113  void SetDynamicPedestal(bool dyn_pede);
114 };
115 
117 template <class C>
119  const C& dataFrame,
120  const double* pedestals,
121  const double* gainRatios,
123  const EcalWeightSet::EcalChi2WeightMatrix** chi2Matrix) {
124  double chi2_(-1.);
125 
126  // double Gain12Equivalent[4]={0,1,2,12};
127  double frame[C::MAXSAMPLES]; // will contain the ADC values
128  double pedestal = 0; // carries pedestal for gain12 i.e. gainId==1
129 
130  int gainId0 = 1; // expected gainId at the beginning of dataFrame
131  int iGainSwitch = 0; // flags whether there's any gainId other than gainId0
132  int GainId = 0; // stores gainId at every sample
133  double maxsample(-1); // ADC value of maximal ped-subtracted sample
134  int imax(-1); // sample number of maximal ped-subtracted sample
135  bool external_pede = false;
136  bool isSaturated = false; // flag reporting whether gain0 has been found
137 
138  // Get time samples checking for Gain Switch and pedestals
139  if (pedestals) {
140  external_pede = true;
141  if (dyn_pedestal) {
142  pedestal = (double(dataFrame.sample(0).adc()) + double(dataFrame.sample(1).adc())) / 2.;
143  } else {
144  pedestal = pedestals[0];
145  }
146  for (int iSample = 0; iSample < C::MAXSAMPLES; iSample++) {
147  //create frame in adc gain 12 equivalent
148  GainId = dataFrame.sample(iSample).gainId();
149 
150  // FIX-ME: warning: the vector pedestal is supposed to have in the order G12, G6 and G1
151  // if GainId is zero treat it as 3 temporarily to protect against undefined
152  // frame will be set to ~max of gain1
153  if (GainId == 0) {
154  GainId = 3;
155  isSaturated = true;
156  }
157 
158  if (GainId != gainId0)
159  iGainSwitch = 1;
160 
161  if (GainId == gainId0) {
162  frame[iSample] = double(dataFrame.sample(iSample).adc()) - pedestal;
163  } else {
164  frame[iSample] = (double(dataFrame.sample(iSample).adc()) - pedestals[GainId - 1]) * gainRatios[GainId - 1];
165  }
166 
167  if (frame[iSample] > maxsample) {
168  maxsample = frame[iSample];
169  imax = iSample;
170  }
171  }
172  } else { // pedestal from pre-sample
173  external_pede = false;
174  pedestal = (double(dataFrame.sample(0).adc()) + double(dataFrame.sample(1).adc())) / 2.;
175 
176  for (int iSample = 0; iSample < C::MAXSAMPLES; iSample++) {
177  //create frame in adc gain 12 equivalent
178  GainId = dataFrame.sample(iSample).gainId();
179  //no gain switch forseen if there is no external pedestal
180  if (GainId == 0) {
181  GainId = 3;
182  isSaturated = true;
183  }
184 
185  frame[iSample] = double(dataFrame.sample(iSample).adc()) - pedestal;
186  // if gain has switched but no pedestals are available, no much good you can do...
187  if (GainId > gainId0)
188  iGainSwitch = 1;
189  if (frame[iSample] > maxsample) {
190  maxsample = frame[iSample];
191  imax = iSample;
192  }
193  }
194  }
195 
196  if ((iGainSwitch == 1 && external_pede == false) || // ... thus you return dummy rechit
197  imax == -1) { // protect against all frames being <-1
198  return EcalUncalibratedRecHit(dataFrame.id(), -1., -100., -1., -1.);
199  }
200 
201  InitFitParameters(frame, imax);
202  chi2_ = PerformAnalyticFit(frame, imax);
203  uint32_t flags = 0;
204  if (isSaturated)
206 
207  /* std::cout << "separate fits\nA: " << fAmp_max_ << ", ResidualPed: " << fPed_max_
208  <<", pedestal: "<<pedestal << ", tPeak " << fTim_max_ << std::endl;
209  */
210  return EcalUncalibratedRecHit(dataFrame.id(), fAmp_max_, pedestal + fPed_max_, fTim_max_ - 5, chi2_, flags);
211 }
212 
213 template <class C>
215  if (alfabeta_ <= 0)
216  return ((double)0.);
217  double dtsbeta, variable, puiss;
218  double dt = t - fTim_max_;
219  if (dt > -alfabeta_) {
220  dtsbeta = dt / fBeta_;
221  variable = 1. + dt / alfabeta_;
222  puiss = pow(variable, fAlpha_);
223  return fAmp_max_ * puiss * exp(-dtsbeta) + fPed_max_;
224  }
225  return fPed_max_;
226 }
227 
228 template <class C>
229 void EcalUncalibRecHitFixedAlphaBetaAlgo<C>::InitFitParameters(double* samples, int max_sample) {
230  // in a first attempt just use the value of the maximum sample
231  fAmp_max_ = samples[max_sample];
232  fTim_max_ = max_sample;
233  fPed_max_ = 0;
234 
235  // amplitude too low for fit to converge
236  // timing set correctly is assumed
237  if (fAmp_max_ < MinAmpl_) {
238  fAmp_max_ = samples[5];
239  double sumA = samples[5] + samples[4] + samples[6];
240  if (sumA != 0) {
241  fTim_max_ = 5 + (samples[6] - samples[4]) / sumA;
242  } else {
243  fTim_max_ = -993;
244  } //-999+6
245  doFit_ = false;
246  }
247  // if timing very badly off, that just use max sample
248  else if (max_sample < 1 || max_sample > 7) {
249  doFit_ = false;
250  } else {
251  //y=a*(x-xM)^2+b*(x-xM)+c
252  doFit_ = true;
253  float a = float(samples[max_sample - 1] + samples[max_sample + 1] - 2 * samples[max_sample]) / 2.;
254  if (a == 0) {
255  doFit_ = false;
256  return;
257  }
258  float b = float(samples[max_sample + 1] - samples[max_sample - 1]) / 2.;
259 
260  fTim_max_ = max_sample - b / (2 * a);
261  fAmp_max_ = samples[max_sample] - b * b / (4 * a);
262  }
263 }
264 
265 template <class C>
266 float EcalUncalibRecHitFixedAlphaBetaAlgo<C>::PerformAnalyticFit(double* samples, int max_sample) {
267  //int fValue_tim_max = max_sample;
272  //| the pedestal (fPed_max_)
273 
274  double chi2 = -1, db[3];
275 
276  //HepSymMatrix DM1(3) ; CLHEP::HepVector temp(3) ;
277 
278  int num_fit_min = (int)(max_sample - fNum_samp_bef_max_);
279  int num_fit_max = (int)(max_sample + fNum_samp_after_max_);
280 
281  if (num_fit_min < 0)
282  num_fit_min = 0;
283  //if (num_fit_max>=fNsamples-1) num_fit_max = fNsamples-2 ;
284  if (num_fit_max >= C::MAXSAMPLES) {
285  num_fit_max = C::MAXSAMPLES - 1;
286  }
287 
288  if (!doFit_) {
289  LogDebug("EcalUncalibRecHitFixedAlphaBetaAlgo") << "No fit performed. The amplitude of sample 5 will be used";
290  return -1;
291  }
292 
293  double func, delta;
294  double variation_func_max = 0.;
295  double variation_tim_max = 0.;
296  double variation_ped_max = 0.;
298  for (int iter = 0; iter < fNb_iter_; iter++) {
300  chi2 = 0.; //PROD.Zero() ; DM1.Zero() ;
301 
302  for (int i1 = 0; i1 < 3; i1++) {
303  temp_[i1] = 0;
304  for (int j1 = i1; j1 < 3; j1++) {
305  DM1_.fast(j1 + 1, i1 + 1) = 0;
306  }
307  }
308 
309  fAmp_max_ += variation_func_max;
310  fTim_max_ += variation_tim_max;
311  fPed_max_ += variation_ped_max;
312 
314  for (int i = num_fit_min; i <= num_fit_max; i++) {
315  //if(i>fsamp_edge_fit && i<num_fit_min) continue ; // remove front edge samples
317  func = pulseShapeFunction((double)i);
319  double dt = (double)i - fTim_max_;
320  if (dt > -alfabeta_) {
321  double dt_sur_beta = dt / fBeta_;
322  double variable = (double)1. + dt / alfabeta_;
323  double expo = exp(-dt_sur_beta);
324  double puissance = pow(variable, fAlpha_);
325 
326  db[0] = un_sur_sigma * puissance * expo;
327  db[1] = fAmp_max_ * db[0] * dt_sur_beta / (alfabeta_ * variable);
328  } else {
329  db[0] = 0.;
330  db[1] = 0.;
331  }
332  db[2] = un_sur_sigma;
334  for (int i1 = 0; i1 < 3; i1++) {
335  for (int j1 = i1; j1 < 3; j1++) {
336  //double & fast(int row, int col);
337  DM1_.fast(j1 + 1, i1 + 1) += db[i1] * db[j1];
338  }
339  }
341  delta = (samples[i] - func) * un_sur_sigma;
343  for (int ii = 0; ii < 3; ii++) {
344  temp_[ii] += delta * db[ii];
345  }
346  chi2 += delta * delta;
347  }
348 
349  int fail = 0;
350  DM1_.invert(fail);
351  if (fail != 0.) {
352  //just a guess from the value of the parameters in the previous interaction;
353  //printf("wH4PulseFitWithFunction =====> determinant error --> No Fit Provided !\n") ;
354  InitFitParameters(samples, max_sample);
355  return -101;
356  }
357  /* for(int i1=0 ; i1<3 ; i1++) { */
358  /* for(int j1=0 ; j1<3 ; j1++) { */
359  /* //double & fast(int row, int col); */
360  /* std::cout<<"inverted: "<<DM1[j1][i1]<<std::endl;;} */
361  /* } */
362  /* std::cout<<"vector temp: "<< temp[0]<<" "<<temp[1]<<" "<<temp[2]<<std::endl; */
364  CLHEP::HepVector PROD = DM1_ * temp_;
365  // std::cout<<"vector PROD: "<< PROD[0]<<" "<<PROD[1]<<" "<<PROD[2]<<std::endl;
366 
367  // Probably the fastest way to protect against
368  // +-inf value in the matrix DM1_ after inversion
369  // (which is nevertheless flagged as successfull...)
370  if (edm::isNotFinite(PROD[0])) {
371  InitFitParameters(samples, max_sample);
372  return -103;
373  }
374 
375  variation_func_max = PROD[0];
376  variation_tim_max = PROD[1];
377  variation_ped_max = PROD[2];
378  //chi2 = chi2/((double)nsamp_used - 3.) ;
379  }
380 
382  if (variation_func_max > 2000. || variation_func_max < -1000.) {
383  InitFitParameters(samples, max_sample);
384  return -102;
385  }
386 
388  fAmp_max_ += variation_func_max;
389  fTim_max_ += variation_tim_max;
390  fPed_max_ += variation_ped_max;
391 
392  // protection against unphysical results:
393  // ampli mismatched to MaxSample, ampli largely negative, time off preselected range
394  if (fAmp_max_ > 2 * samples[max_sample] || fAmp_max_ < -100 || fTim_max_ < 0 || 9 < fTim_max_) {
395  InitFitParameters(samples, max_sample);
396  return -104;
397  }
398 
399  //std::cout <<"chi2: "<<chi2<<" ampl: "<<fAmp_max_<<" time: "<<fTim_max_<<" pede: "<<fPed_max_<<std::endl;
400  return chi2;
401 }
402 
403 template <class C>
405  fAlpha_ = alpha;
406  fBeta_ = beta;
407  alfabeta_ = fAlpha_ * fBeta_;
408 }
409 
410 template <class C>
412  MinAmpl_ = ampl;
413 }
414 template <class C>
416  dyn_pedestal = p;
417 }
418 #endif
void InitFitParameters(double *samples, int max_sample)
float dt
Definition: AMPTWrapper.h:136
float alpha
Definition: AMPTWrapper.h:105
constexpr bool isNotFinite(T x)
Definition: isFinite.h:9
tuple db
Definition: EcalCondDB.py:153
uint32_t T const *__restrict__ uint32_t const *__restrict__ int32_t int Histo::index_type cudaStream_t Func __host__ __device__ V int Func func
math::Matrix< 10, 10 >::type EcalChi2WeightMatrix
Definition: EcalWeightSet.h:20
math::Matrix< 3, 10 >::type EcalWeightMatrix
Definition: EcalWeightSet.h:19
EcalUncalibratedRecHit makeRecHit(const C &dataFrame, const double *pedestals, const double *gainRatios, const EcalWeightSet::EcalWeightMatrix **weights, const EcalWeightSet::EcalChi2WeightMatrix **chi2Matrix) override
Compute parameters.
Exp< T >::type exp(const T &t)
Definition: Exp.h:22
int ii
Definition: cuy.py:589
double b
Definition: hdecay.h:118
bool isSaturated(const Digi &digi, const int &maxADCvalue, int ifirst, int n)
double a
Definition: hdecay.h:119
#define PROD(A, B)
float PerformAnalyticFit(double *samples, int max_sample)
Power< A, B >::type pow(const A &a, const B &b)
Definition: Power.h:29
#define LogDebug(id)