| 1 |
#if !defined(__CINT__) && !defined(__MAKECINT__)
|
| 2 |
#include "DataFormats/FWLite/interface/Handle.h"
|
| 3 |
#include "DataFormats/FWLite/interface/Event.h"
|
| 4 |
#include "DataFormats/FWLite/interface/ChainEvent.h"
|
| 5 |
//Headers for the data items
|
| 6 |
#include "DataFormats/VertexReco/interface/Vertex.h"
|
| 7 |
#include "DataFormats/BeamSpot/interface/BeamSpot.h"
|
| 8 |
|
| 9 |
#endif
|
| 10 |
|
| 11 |
#include "TROOT.h"
|
| 12 |
#include "TH2F.h"
|
| 13 |
#include "TFile.h"
|
| 14 |
#include "TF1.h"
|
| 15 |
#include "TCanvas.h"
|
| 16 |
#include "TMath.h"
|
| 17 |
#include "TBox.h"
|
| 18 |
#include <vector>
|
| 19 |
#include <map>
|
| 20 |
#include <string>
|
| 21 |
|
| 22 |
using namespace std;
|
| 23 |
typedef map<unsigned int, unsigned int> RunNumberMap;
|
| 24 |
|
| 25 |
//
|
| 26 |
// iterative fit of +- 2 sigma Gaussian core on a 1d histogram
|
| 27 |
//
|
| 28 |
bool fitHistogram (TH1* slice, TF1& fg, double* pars, double* epars, bool dbg = false) {
|
| 29 |
// initialization of parameters
|
| 30 |
pars[0] = slice->GetMaximum();
|
| 31 |
pars[1] = slice->GetMean();
|
| 32 |
pars[2] = slice->GetRMS();
|
| 33 |
epars[0] = 0;
|
| 34 |
epars[1] = 0;
|
| 35 |
epars[2] = 0;
|
| 36 |
// some iterations
|
| 37 |
bool converged(false);
|
| 38 |
for ( unsigned int it=0; it<10; ++it) {
|
| 39 |
const double nSigRange = 2.;
|
| 40 |
fg.SetParameters(pars);
|
| 41 |
double oldMean = pars[1];
|
| 42 |
double oldWidth = pars[2];
|
| 43 |
if ( it>0 ) {
|
| 44 |
fg.SetRange(pars[1]-nSigRange*pars[2],
|
| 45 |
pars[1]+nSigRange*pars[2]);
|
| 46 |
// if ( dbg ) {
|
| 47 |
// cout << "Setting range to " << pars[1]-nSigRange*pars[2]
|
| 48 |
// << " " << pars[1]+nSigRange*pars[2] << endl;
|
| 49 |
// }
|
| 50 |
}
|
| 51 |
else {
|
| 52 |
fg.SetRange(slice->GetXaxis()->GetXmin(),
|
| 53 |
slice->GetXaxis()->GetXmax());
|
| 54 |
// if ( dbg ) {
|
| 55 |
// cout << "Setting range to " << slice->GetXaxis()->GetXmin()
|
| 56 |
// << " " << slice->GetXaxis()->GetXmax() << endl;
|
| 57 |
// }
|
| 58 |
}
|
| 59 |
if ( dbg )
|
| 60 |
slice->Fit(&fg,"QLR+");
|
| 61 |
else
|
| 62 |
slice->Fit(&fg,"QLR");
|
| 63 |
fg.GetParameters(pars);
|
| 64 |
epars[0] = fg.GetParError(0);
|
| 65 |
epars[1] = fg.GetParError(1);
|
| 66 |
epars[2] = fg.GetParError(2);
|
| 67 |
if ( converged ) {
|
| 68 |
cout << "terminating at iteration " << it << " "
|
| 69 |
<< pars[1] << " " << oldMean << " " << epars[1] << " "
|
| 70 |
<< pars[2] << " " << oldWidth << " " << epars[2] << endl;
|
| 71 |
break;
|
| 72 |
}
|
| 73 |
else if ( dbg ) {
|
| 74 |
cout << " at iteration " << it << " "
|
| 75 |
<< pars[1] << " " << oldMean << " " << epars[1] << " "
|
| 76 |
<< pars[2] << " " << oldWidth << " " << epars[2] << endl;
|
| 77 |
}
|
| 78 |
// convergence criterion
|
| 79 |
if ( it>2 &&
|
| 80 |
fabs((pars[1]-oldMean))<max(fabs(0.01*epars[1]),0.0010) &&
|
| 81 |
fabs((pars[2]-oldWidth))<max(fabs(0.01*epars[2]),0.0010) ) {
|
| 82 |
converged = true;
|
| 83 |
}
|
| 84 |
}
|
| 85 |
|
| 86 |
return converged;
|
| 87 |
}
|
| 88 |
|
| 89 |
//
|
| 90 |
// fit Gaussian to slices of the 2d "run" histogram
|
| 91 |
//
|
| 92 |
void fitRunHistogram (TH2* histo, TH1*& hmean, TH1*& hwidth) {
|
| 93 |
//
|
| 94 |
// create histograms for mean and width vs. run
|
| 95 |
//
|
| 96 |
TDirectory* saveDir = gDirectory;
|
| 97 |
gROOT->cd();
|
| 98 |
int nbin = histo->GetNbinsX();
|
| 99 |
string title = histo->GetName();
|
| 100 |
TAxis* xhisto = histo->GetXaxis();
|
| 101 |
hmean = new TH1F((title+"Mean").c_str(),title.c_str(),nbin,-0.5,nbin-0.5);
|
| 102 |
TAxis* xhmean = hmean->GetXaxis();
|
| 103 |
hwidth = new TH1F((title+"Width").c_str(),title.c_str(),nbin,-0.5,nbin-0.5);
|
| 104 |
TAxis* xhwidth = hwidth->GetXaxis();
|
| 105 |
//
|
| 106 |
// fit function
|
| 107 |
//
|
| 108 |
TF1 fg("fg","gaus");
|
| 109 |
//
|
| 110 |
// loop over runs
|
| 111 |
//
|
| 112 |
bool converged;
|
| 113 |
double pars[3];
|
| 114 |
double epars[3];
|
| 115 |
for ( unsigned int ib=1; ib<=nbin; ++ib ) {
|
| 116 |
//
|
| 117 |
// initialize mean and width to 0 error
|
| 118 |
//
|
| 119 |
hmean->SetBinError(ib,0.);
|
| 120 |
hwidth->SetBinError(ib,0.);
|
| 121 |
//
|
| 122 |
// get slice
|
| 123 |
//
|
| 124 |
TH1* slice =
|
| 125 |
histo->ProjectionY((title+xhisto->GetBinLabel(ib)).c_str(),ib,ib);
|
| 126 |
// drop histograms with too few entries
|
| 127 |
if ( slice->GetEntries()<100 ) {
|
| 128 |
delete slice;
|
| 129 |
continue;
|
| 130 |
}
|
| 131 |
//
|
| 132 |
// fit slice
|
| 133 |
//
|
| 134 |
converged = fitHistogram(slice,fg,pars,epars);
|
| 135 |
if ( converged ) {
|
| 136 |
// copy parameters and errors
|
| 137 |
hmean->SetBinContent(ib,pars[1]);
|
| 138 |
hmean->SetBinError(ib,epars[1]);
|
| 139 |
hwidth->SetBinContent(ib,pars[2]);
|
| 140 |
hwidth->SetBinError(ib,epars[2]);
|
| 141 |
}
|
| 142 |
else {
|
| 143 |
cout << "No convergence for " << histo->GetName() << " / run "
|
| 144 |
<< histo->GetXaxis()->GetBinLabel(ib) << endl;
|
| 145 |
// refit with debug flag
|
| 146 |
// fitHistogram(slice,fg,pars,epars,true);
|
| 147 |
hmean->SetBinContent(ib,pars[1]);
|
| 148 |
hmean->SetBinError(ib,epars[1]);
|
| 149 |
hwidth->SetBinContent(ib,pars[2]);
|
| 150 |
hwidth->SetBinError(ib,epars[2]);
|
| 151 |
}
|
| 152 |
// keep failing fits for debugging
|
| 153 |
// if ( converged ) delete slice;
|
| 154 |
xhmean->SetBinLabel(ib,xhisto->GetBinLabel(ib));
|
| 155 |
xhwidth->SetBinLabel(ib,xhisto->GetBinLabel(ib));
|
| 156 |
}
|
| 157 |
|
| 158 |
saveDir->cd();
|
| 159 |
}
|
| 160 |
|
| 161 |
//
|
| 162 |
// set x-axis labels to run number
|
| 163 |
//
|
| 164 |
void setRunNumberLabels (TH1* histo, RunNumberMap& runNumbers) {
|
| 165 |
char label[32];
|
| 166 |
TAxis* axis = histo->GetXaxis();
|
| 167 |
for ( RunNumberMap::const_iterator i=runNumbers.begin();
|
| 168 |
i!=runNumbers.end(); ++i ) {
|
| 169 |
sprintf(label,"%d",(*i).first);
|
| 170 |
axis->SetBinLabel((*i).second+1,label);
|
| 171 |
}
|
| 172 |
}
|
| 173 |
|
| 174 |
|
| 175 |
//
|
| 176 |
// plot PV distribution for a single run with indication of BS position
|
| 177 |
//
|
| 178 |
void plotRun (char coord, int run)
|
| 179 |
{
|
| 180 |
char title[64];
|
| 181 |
|
| 182 |
sprintf(title,"pv%cVsRun%d",coord,run);
|
| 183 |
TH1* hpv = (TH1*)gROOT->FindObject(title);
|
| 184 |
if ( hpv==0 ) return;
|
| 185 |
|
| 186 |
sprintf(title,"pv%cRun%d",coord,run);
|
| 187 |
TCanvas* c = new TCanvas(title);
|
| 188 |
|
| 189 |
hpv->Draw();
|
| 190 |
|
| 191 |
sprintf(title,"bs%cVsRunMean",coord);
|
| 192 |
TH1* hbs = (TH1*)gROOT->FindObject(title);
|
| 193 |
sprintf(title,"%d",run);
|
| 194 |
double vbs(0.);
|
| 195 |
double ebs(0.);
|
| 196 |
for ( unsigned int ib=1; ib<=hbs->GetNbinsX(); ++ib ) {
|
| 197 |
const char* label = hbs->GetXaxis()->GetBinLabel(ib);
|
| 198 |
if ( strcmp(title,label)==0 ) {
|
| 199 |
vbs = hbs->GetBinContent(ib);
|
| 200 |
ebs = hbs->GetBinError(ib);
|
| 201 |
break;
|
| 202 |
}
|
| 203 |
}
|
| 204 |
TBox* box = new TBox(vbs-ebs,0,vbs+ebs,hpv->GetMaximum());
|
| 205 |
box->SetFillStyle(1001);
|
| 206 |
box->SetFillColor(7);
|
| 207 |
box->Draw();
|
| 208 |
|
| 209 |
hpv->Draw("same");
|
| 210 |
|
| 211 |
sprintf(title,"pv%cRun%d.png",coord,run);
|
| 212 |
c->SaveAs(title);
|
| 213 |
delete c;
|
| 214 |
}
|
| 215 |
//
|
| 216 |
// summary plots (PV + BS) vs run
|
| 217 |
//
|
| 218 |
void plotPvVsBs () {
|
| 219 |
char coord[] = { "xyz" };
|
| 220 |
for ( int ic=0; ic<3; ++ic ) {
|
| 221 |
string canvasTitle("pvVsBs");
|
| 222 |
canvasTitle += coord[ic];
|
| 223 |
TCanvas* c = new TCanvas(canvasTitle.c_str(),canvasTitle.c_str(),
|
| 224 |
700,500);
|
| 225 |
string bsTitle("bs.VsRunMean");
|
| 226 |
bsTitle.replace(2,1,1,coord[ic]);
|
| 227 |
TH1* bsVsRun = (TH1*)gROOT->FindObject(bsTitle.c_str());
|
| 228 |
string pvTitle("pv.VsRunMean");
|
| 229 |
pvTitle.replace(2,1,1,coord[ic]);
|
| 230 |
TH1* pvVsRun = (TH1*)gROOT->FindObject(pvTitle.c_str());
|
| 231 |
pvVsRun->SetMarkerStyle(20);
|
| 232 |
pvVsRun->Draw();
|
| 233 |
bsVsRun->SetFillStyle(1001);
|
| 234 |
bsVsRun->SetFillColor(7);
|
| 235 |
bsVsRun->Draw("E2 same");
|
| 236 |
pvVsRun->Draw("same");
|
| 237 |
}
|
| 238 |
}
|
| 239 |
|
| 240 |
//
|
| 241 |
// main entry point
|
| 242 |
//
|
| 243 |
void pvHistory (const vector<string>& fileNames) {
|
| 244 |
|
| 245 |
fwlite::ChainEvent ev(fileNames);
|
| 246 |
|
| 247 |
fwlite::Handle< reco::BeamSpot > bsHandle;
|
| 248 |
fwlite::Handle< vector<reco::Vertex> > pvHandle;
|
| 249 |
//
|
| 250 |
// create list of runs with > 100 PVs
|
| 251 |
//
|
| 252 |
RunNumberMap runNumberCounts;
|
| 253 |
int lastRun(0);
|
| 254 |
for( ev.toBegin(); !ev.atEnd(); ++ev) {
|
| 255 |
unsigned int run = ev.id().run();
|
| 256 |
pvHandle.getByLabel(ev,"offlinePrimaryVertices");
|
| 257 |
|
| 258 |
const reco::Vertex& pv = (*pvHandle)[0];
|
| 259 |
if ( pv.tracksSize()>0 && pv.ndof()>2 &&
|
| 260 |
(pv.ndof()+3.)/2./pv.tracksSize()>0.5 ) {
|
| 261 |
++runNumberCounts[run];
|
| 262 |
}
|
| 263 |
}
|
| 264 |
// create list of bin indices
|
| 265 |
unsigned int nRun(0);
|
| 266 |
RunNumberMap runNumberIndices;
|
| 267 |
for ( RunNumberMap::iterator i=runNumberCounts.begin();
|
| 268 |
i!=runNumberCounts.end(); ++i ) {
|
| 269 |
if ( (*i).second>100 ) runNumberIndices[(*i).first] = nRun++;
|
| 270 |
}
|
| 271 |
cout << "Found " << nRun << " run numbers" << endl;
|
| 272 |
//
|
| 273 |
// create histograms
|
| 274 |
//
|
| 275 |
TDirectory* saveDir = gDirectory;
|
| 276 |
gROOT->cd();
|
| 277 |
TH2* pvxVsRun = new TH2F("pvxVsRun","pvxVsRun",nRun,-0.5,nRun-0.5,80,-0.1,0.3);
|
| 278 |
TH2* pvyVsRun = new TH2F("pvyVsRun","pvyVsRun",nRun,-0.5,nRun-0.5,80,-0.2,0.2);
|
| 279 |
TH2* pvzVsRun = new TH2F("pvzVsRun","pvzVsRun",nRun,-0.5,nRun-0.5,80,-20,20);
|
| 280 |
TH2* bsxVsRun = new TH2F("bsxVsRun","bsxVsRun",nRun,-0.5,nRun-0.5,80,0,0.4);
|
| 281 |
TH2* bsyVsRun = new TH2F("bsyVsRun","bsyVsRun",nRun,-0.5,nRun-0.5,80,0,0.4);
|
| 282 |
TH2* bszVsRun = new TH2F("bszVsRun","bszVsRun",nRun,-0.5,nRun-0.5,80,-20,20);
|
| 283 |
TH1* bsxVsRunMean = new TH1F("bsxVsRunMean","bsxVsRunMean",nRun,-0.5,nRun-0.5);
|
| 284 |
TH1* bsyVsRunMean = new TH1F("bsyVsRunMean","bsyVsRunMean",nRun,-0.5,nRun-0.5);
|
| 285 |
TH1* bszVsRunMean = new TH1F("bszVsRunMean","bszVsRunMean",nRun,-0.5,nRun-0.5);
|
| 286 |
saveDir->cd();
|
| 287 |
|
| 288 |
//
|
| 289 |
// fill histograms
|
| 290 |
//
|
| 291 |
lastRun = -1;
|
| 292 |
int idx(-1);
|
| 293 |
int nev(0);
|
| 294 |
for( ev.toBegin(); !ev.atEnd(); ++ev) {
|
| 295 |
int run = ev.id().run();
|
| 296 |
// if ( (++nev)%1000 == 0 ) cout << "processing record " << nev << " ( run "
|
| 297 |
// << ev.id().run() << " " << " evt "
|
| 298 |
// << ev.id().event() << " )" << endl;
|
| 299 |
//
|
| 300 |
// update bin index if run changed
|
| 301 |
//
|
| 302 |
if ( run!=lastRun ) {
|
| 303 |
RunNumberMap::const_iterator irun = runNumberIndices.find(run);
|
| 304 |
if ( irun == runNumberIndices.end() ) {
|
| 305 |
idx = -1;
|
| 306 |
}
|
| 307 |
else {
|
| 308 |
idx = (*irun).second;
|
| 309 |
}
|
| 310 |
lastRun = run;
|
| 311 |
// cout << "Run " << run << " idx " << idx << endl;
|
| 312 |
}
|
| 313 |
if ( idx<0 ) continue;
|
| 314 |
|
| 315 |
//
|
| 316 |
// retrieve beam spot and PV
|
| 317 |
//
|
| 318 |
bsHandle.getByLabel(ev,"offlineBeamSpot");
|
| 319 |
pvHandle.getByLabel(ev,"offlinePrimaryVertices");
|
| 320 |
//
|
| 321 |
// fill PV histograms (after quality cut)
|
| 322 |
//
|
| 323 |
const reco::Vertex& pv = (*pvHandle)[0];
|
| 324 |
if ( pv.tracksSize()>0 && pv.ndof()>2 &&
|
| 325 |
(pv.ndof()+3.)/2./pv.tracksSize()>0.5 ) {
|
| 326 |
pvxVsRun->Fill(idx,pv.x());
|
| 327 |
pvyVsRun->Fill(idx,pv.y());
|
| 328 |
pvzVsRun->Fill(idx,pv.z());
|
| 329 |
}
|
| 330 |
//
|
| 331 |
// fill BS histograms
|
| 332 |
//
|
| 333 |
bsxVsRun->Fill(idx,bsHandle->x0());
|
| 334 |
bsyVsRun->Fill(idx,bsHandle->y0());
|
| 335 |
bszVsRun->Fill(idx,bsHandle->z0());
|
| 336 |
if ( idx>=0 && idx<bsxVsRun->GetNbinsX() ) {
|
| 337 |
// assume BS is constant / run: fill current value into 1d histogram
|
| 338 |
bsxVsRunMean->SetBinContent(idx+1,bsHandle->x0());
|
| 339 |
bsxVsRunMean->SetBinError(idx+1,bsHandle->x0Error());
|
| 340 |
bsyVsRunMean->SetBinContent(idx+1,bsHandle->y0());
|
| 341 |
bsyVsRunMean->SetBinError(idx+1,bsHandle->y0Error());
|
| 342 |
bszVsRunMean->SetBinContent(idx+1,bsHandle->z0());
|
| 343 |
bszVsRunMean->SetBinError(idx+1,bsHandle->z0Error());
|
| 344 |
}
|
| 345 |
|
| 346 |
}
|
| 347 |
saveDir = gDirectory;
|
| 348 |
//
|
| 349 |
// number of entries / run
|
| 350 |
//
|
| 351 |
gROOT->cd();
|
| 352 |
TH1* npvVsRun = pvzVsRun->ProjectionX("npvVsRun");
|
| 353 |
npvVsRun->SetTitle("npvVsRun");
|
| 354 |
TH1* nbsVsRun = bszVsRun->ProjectionX("nbsVsRun");
|
| 355 |
npvVsRun->SetTitle("nbsVsRun");
|
| 356 |
saveDir->cd();
|
| 357 |
//
|
| 358 |
// set run numbers on x-axis labels
|
| 359 |
//
|
| 360 |
setRunNumberLabels(npvVsRun,runNumberIndices);
|
| 361 |
setRunNumberLabels(pvxVsRun,runNumberIndices);
|
| 362 |
setRunNumberLabels(pvyVsRun,runNumberIndices);
|
| 363 |
setRunNumberLabels(pvzVsRun,runNumberIndices);
|
| 364 |
setRunNumberLabels(nbsVsRun,runNumberIndices);
|
| 365 |
setRunNumberLabels(bsxVsRun,runNumberIndices);
|
| 366 |
setRunNumberLabels(bsyVsRun,runNumberIndices);
|
| 367 |
setRunNumberLabels(bszVsRun,runNumberIndices);
|
| 368 |
setRunNumberLabels(bsxVsRunMean,runNumberIndices);
|
| 369 |
setRunNumberLabels(bsyVsRunMean,runNumberIndices);
|
| 370 |
setRunNumberLabels(bszVsRunMean,runNumberIndices);
|
| 371 |
//
|
| 372 |
// fit histograms
|
| 373 |
//
|
| 374 |
TH1* pvxMean; TH1* pvxWidth;
|
| 375 |
fitRunHistogram(pvxVsRun,pvxMean,pvxWidth);
|
| 376 |
TH1* pvyMean; TH1* pvyWidth;
|
| 377 |
fitRunHistogram(pvyVsRun,pvyMean,pvyWidth);
|
| 378 |
TH1* pvzMean; TH1* pvzWidth;
|
| 379 |
fitRunHistogram(pvzVsRun,pvzMean,pvzWidth);
|
| 380 |
|
| 381 |
// cout << pvxMean << endl;
|
| 382 |
// cout << pvxMean->GetNbinsX() << " " << pvxMean->GetXaxis() << endl;
|
| 383 |
//
|
| 384 |
// summary table
|
| 385 |
//
|
| 386 |
for ( unsigned int i=1; i<=pvxMean->GetNbinsX(); ++i ) {
|
| 387 |
if ( pvxMean->GetBinError(i)>1.e-10 ||
|
| 388 |
pvxMean->GetBinError(i)>1.e-10 ||
|
| 389 |
pvxMean->GetBinError(i)>1.e-10 ) {
|
| 390 |
printf(" %-6s : %8.4f +- %6.4f %6.4f +- %6.4f %8.4f +- %6.4f %6.4f +- %6.4f %9.4f +- %7.4f %6.4f +- %7.4f\n",
|
| 391 |
pvxMean->GetXaxis()->GetBinLabel(i),
|
| 392 |
pvxMean->GetBinContent(i),pvxMean->GetBinError(i),
|
| 393 |
pvxWidth->GetBinContent(i),pvxWidth->GetBinError(i),
|
| 394 |
pvyMean->GetBinContent(i),pvyMean->GetBinError(i),
|
| 395 |
pvyWidth->GetBinContent(i),pvyWidth->GetBinError(i),
|
| 396 |
pvzMean->GetBinContent(i),pvzMean->GetBinError(i),
|
| 397 |
pvzWidth->GetBinContent(i),pvzWidth->GetBinError(i));
|
| 398 |
}
|
| 399 |
}
|
| 400 |
|
| 401 |
//
|
| 402 |
// save histograms
|
| 403 |
//
|
| 404 |
TFile* fo = new TFile("../skim_BSC0_AND_40OR41_MinBias_Dec19th_v1.root","recreate");
|
| 405 |
pvxVsRun->Write();
|
| 406 |
pvyVsRun->Write();
|
| 407 |
pvzVsRun->Write();
|
| 408 |
pvxMean->Write();
|
| 409 |
pvxWidth->Write();
|
| 410 |
pvyMean->Write();
|
| 411 |
pvyWidth->Write();
|
| 412 |
pvzMean->Write();
|
| 413 |
pvzWidth->Write();
|
| 414 |
fo->Write();
|
| 415 |
delete fo;
|
| 416 |
|
| 417 |
plotPvVsBs();
|
| 418 |
//
|
| 419 |
// plot inidividual histograms / run
|
| 420 |
//
|
| 421 |
// for ( RunNumberMap::iterator i=runNumberIndices.begin();
|
| 422 |
// i!=runNumberIndices.end(); ++i ) {
|
| 423 |
// plotRun('x',(*i).first);
|
| 424 |
// plotRun('y',(*i).first);
|
| 425 |
// plotRun('z',(*i).first);
|
| 426 |
// }
|
| 427 |
|
| 428 |
}
|
| 429 |
//
|
| 430 |
// call with single file name
|
| 431 |
//
|
| 432 |
void pvHistory (const char* fileName) {
|
| 433 |
vector<string> fileNames(1,fileName);
|
| 434 |
pvHistory(fileNames);
|
| 435 |
}
|
