Project
Loading...
Searching...
No Matches
GPUQA.cxx
Go to the documentation of this file.
1// Copyright 2019-2020 CERN and copyright holders of ALICE O2.
2// See https://alice-o2.web.cern.ch/copyright for details of the copyright holders.
3// All rights not expressly granted are reserved.
4//
5// This software is distributed under the terms of the GNU General Public
6// License v3 (GPL Version 3), copied verbatim in the file "COPYING".
7//
8// In applying this license CERN does not waive the privileges and immunities
9// granted to it by virtue of its status as an Intergovernmental Organization
10// or submit itself to any jurisdiction.
11
14
15#define QA_DEBUG 0
16#define QA_TIMING 0
17
18#include "Rtypes.h" // Include ROOT header first, to use ROOT and disable replacements
19
20#include "TH1F.h"
21#include "TH2F.h"
22#include "TH1D.h"
23#include "TGraphAsymmErrors.h"
24#include "TCanvas.h"
25#include "TPad.h"
26#include "TLegend.h"
27#include "TColor.h"
28#include "TPaveText.h"
29#include "TF1.h"
30#include "TFile.h"
31#include "TTree.h"
32#include "TStyle.h"
33#include "TLatex.h"
34#include "TObjArray.h"
35#include <sys/stat.h>
36
37#include "GPUQA.h"
38#include "GPUTPCDef.h"
39#include "GPUTPCTrackingData.h"
40#include "GPUChainTracking.h"
41#include "GPUChainTrackingGetters.inc"
42#include "GPUTPCTrack.h"
43#include "GPUTPCTracker.h"
44#include "GPUTPCGMMergedTrack.h"
45#include "GPUTPCGMPropagator.h"
46#include "AliHLTTPCClusterMCData.h"
47#include "GPUTPCMCInfo.h"
48#include "GPUO2DataTypes.h"
49#include "GPUParam.inc"
50#include "GPUTPCClusterRejection.h"
51#include "GPUTPCConvertImpl.h"
52#include "TPCFastTransformPOD.h"
53#include "GPUROOTDump.h"
54#include "SimulationDataFormat/ConstMCTruthContainer.h"
55#include "SimulationDataFormat/MCCompLabel.h"
56#include "GPUSettings.h"
57#include "GPUDefMacros.h"
58#ifndef GPUCA_STANDALONE
59#include "DetectorsRaw/HBFUtils.h"
60#include "DataFormatsTPC/TrackTPC.h"
61#include "DataFormatsTPC/Constants.h"
62#include "SimulationDataFormat/MCTrack.h"
63#include "SimulationDataFormat/TrackReference.h"
64#include "SimulationDataFormat/DigitizationContext.h"
65#include "DetectorsCommonDataFormats/DetID.h"
66#include "CommonUtils/NameConf.h"
67#include "Steer/MCKinematicsReader.h"
68#include "TPDGCode.h"
69#include "TParticlePDG.h"
70#include "TDatabasePDG.h"
71#endif
72#include "GPUQAHelper.h"
73#include <algorithm>
74#include <cstdio>
75#include <cinttypes>
76#include <fstream>
77
78#include "utils/timer.h"
79
80#include <oneapi/tbb.h>
81
82using namespace o2::gpu;
83
84namespace o2::gpu
85{
86struct checkClusterStateResult {
87 bool unattached = false;
88 float qpt = 0.f;
89 bool lowPt = false;
90 bool mev200 = false;
91 bool mergedLooperUnconnected = false;
92 bool mergedLooperConnected = false;
93 int32_t id = 0;
94 bool physics = false, protect = false;
95};
96} // namespace o2::gpu
97
98template <bool COUNT, class T>
99inline checkClusterStateResult GPUQA::checkClusterState(uint32_t attach, T* counts) const
100{
101 checkClusterStateResult r;
102 r.unattached = attach == 0;
103 r.id = attach & gputpcgmmergertypes::attachTrackMask;
104 if (!r.unattached && !(attach & gputpcgmmergertypes::attachProtect)) {
105 r.qpt = fabsf(mTracking->mIOPtrs.mergedTracks[r.id].GetParam().GetQPt());
106 r.lowPt = r.qpt * mTracking->GetParam().qptB5Scaler > mTracking->GetParam().rec.tpc.rejectQPtB5;
107 r.mev200 = r.qpt > 5;
108 r.mergedLooperUnconnected = mTracking->mIOPtrs.mergedTracks[r.id].MergedLooperUnconnected();
109 r.mergedLooperConnected = mTracking->mIOPtrs.mergedTracks[r.id].MergedLooperConnected();
110 }
111 if (r.mev200) {
112 if constexpr (COUNT) {
113 counts->n200MeV++;
114 }
115 }
116 if (r.lowPt) {
117 if constexpr (COUNT) {
118 counts->nLowPt++;
119 }
120 } else if (r.mergedLooperUnconnected) {
121 if constexpr (COUNT) {
122 counts->nMergedLooperUnconnected++;
123 }
124 } else if (r.mergedLooperConnected) {
125 if constexpr (COUNT) {
126 counts->nMergedLooperConnected++;
127 }
128 } else if (attach) {
129 r.protect = !GPUTPCClusterRejection::GetRejectionStatus<COUNT>(attach, r.physics, counts, &r.mev200) && ((attach & gputpcgmmergertypes::attachProtect) || !GPUTPCClusterRejection::IsTrackRejected(mTracking->mIOPtrs.mergedTracks[r.id], mTracking->GetParam()));
130 }
131 return r;
132}
133
134static const GPUSettingsQA& GPUQA_GetConfig(GPUChainTracking* chain)
135{
136 static GPUSettingsQA defaultConfig;
137 if (chain && chain->mConfigQA) {
138 return *chain->mConfigQA;
139 } else {
140 return defaultConfig;
141 }
142}
143
144static const constexpr float LOG_PT_MIN = -1.;
145
146static constexpr float Y_MAX = 40;
147static constexpr float Z_MAX = 100;
148static constexpr float PT_MIN = 0.01; // TODO: Take from Param
149static constexpr float PT_MIN_PRIM = 0.1;
150static constexpr float PT_MIN_CLUST = 0.01;
151static constexpr float PT_MAX = 20;
152static constexpr float ETA_MAX = 1.5;
153static constexpr float ETA_MAX2 = 0.9;
154static constexpr int32_t PADROW_CHECK_MINCLS = 50;
155
156static constexpr bool CLUST_HIST_INT_SUM = false;
157
158static constexpr const int32_t COLORCOUNT = 12;
159
160static const constexpr char* EFF_TYPES[6] = {"Rec", "Clone", "Fake", "All", "RecAndClone", "MC"};
161static const constexpr char* FINDABLE_NAMES[2] = {"All", "Findable"};
162static const constexpr char* PRIM_NAMES[2] = {"Prim", "Sec"};
163static const constexpr char* PARAMETER_NAMES[5] = {"Y", "Z", "#Phi", "#lambda", "Relative #it{p}_{T}"};
164static const constexpr char* PARAMETER_NAMES_NATIVE[5] = {"Y", "Z", "sin(#Phi)", "tan(#lambda)", "q/#it{p}_{T} (curvature)"};
165static const constexpr char* VSPARAMETER_NAMES[6] = {"Y", "Z", "Phi", "Eta", "Pt", "Pt_log"};
166static const constexpr char* EFF_NAMES[3] = {"Efficiency", "Clone Rate", "Fake Rate"};
167static const constexpr char* EFFICIENCY_TITLES[4] = {"Efficiency (Primary Tracks, Findable)", "Efficiency (Secondary Tracks, Findable)", "Efficiency (Primary Tracks)", "Efficiency (Secondary Tracks)"};
168static const constexpr double SCALE[5] = {10., 10., 1000., 1000., 100.};
169static const constexpr double SCALE_NATIVE[5] = {10., 10., 1000., 1000., 1.};
170static const constexpr char* XAXIS_TITLES[5] = {"#it{y}_{mc} (cm)", "#it{z}_{mc} (cm)", "#Phi_{mc} (rad)", "#eta_{mc}", "#it{p}_{Tmc} (GeV/#it{c})"};
171static const constexpr char* AXIS_TITLES[5] = {"#it{y}-#it{y}_{mc} (mm) (Resolution)", "#it{z}-#it{z}_{mc} (mm) (Resolution)", "#phi-#phi_{mc} (mrad) (Resolution)", "#lambda-#lambda_{mc} (mrad) (Resolution)", "(#it{p}_{T} - #it{p}_{Tmc}) / #it{p}_{Tmc} (%) (Resolution)"};
172static const constexpr char* AXIS_TITLES_NATIVE[5] = {"#it{y}-#it{y}_{mc} (mm) (Resolution)", "#it{z}-#it{z}_{mc} (mm) (Resolution)", "sin(#phi)-sin(#phi_{mc}) (Resolution)", "tan(#lambda)-tan(#lambda_{mc}) (Resolution)", "q*(q/#it{p}_{T} - q/#it{p}_{Tmc}) (Resolution)"};
173static const constexpr char* AXIS_TITLES_PULL[5] = {"#it{y}-#it{y}_{mc}/#sigma_{y} (Pull)", "#it{z}-#it{z}_{mc}/#sigma_{z} (Pull)", "sin(#phi)-sin(#phi_{mc})/#sigma_{sin(#phi)} (Pull)", "tan(#lambda)-tan(#lambda_{mc})/#sigma_{tan(#lambda)} (Pull)",
174 "q*(q/#it{p}_{T} - q/#it{p}_{Tmc})/#sigma_{q/#it{p}_{T}} (Pull)"};
175static const constexpr char* CLUSTER_NAMES[GPUQA::N_CLS_HIST] = {"Correctly attached clusters", "Fake attached clusters", "Attached + adjacent clusters", "Fake adjacent clusters", "Clusters of reconstructed tracks", "Used in Physics", "Protected", "All clusters"};
176static const constexpr char* CLUSTER_TITLES[GPUQA::N_CLS_TYPE] = {"Clusters Pt Distribution / Attachment", "Clusters Pt Distribution / Attachment (relative to all clusters)", "Clusters Pt Distribution / Attachment (integrated)"};
177static const constexpr char* CLUSTER_NAMES_SHORT[GPUQA::N_CLS_HIST] = {"Attached", "Fake", "AttachAdjacent", "FakeAdjacent", "FoundTracks", "Physics", "Protected", "All"};
178static const constexpr char* CLUSTER_TYPES[GPUQA::N_CLS_TYPE] = {"", "Ratio", "Integral"};
179static const constexpr char* REJECTED_NAMES[3] = {"All", "Rejected", "Fraction"};
180static const constexpr int32_t COLORS_HEX[COLORCOUNT] = {0xB03030, 0x00A000, 0x0000C0, 0x9400D3, 0x19BBBF, 0xF25900, 0x7F7F7F, 0xFFD700, 0x07F707, 0x07F7F7, 0xF08080, 0x000000};
181
182static const constexpr int32_t CONFIG_DASHED_MARKERS = 0;
183
184static const constexpr float AXES_MIN[5] = {-Y_MAX, -Z_MAX, 0.f, -ETA_MAX, PT_MIN};
185static const constexpr float AXES_MAX[5] = {Y_MAX, Z_MAX, 2.f * M_PI, ETA_MAX, PT_MAX};
186static const constexpr int32_t AXIS_BINS[5] = {51, 51, 144, 31, 50};
187static const constexpr int32_t RES_AXIS_BINS[] = {1017, 113}; // Consecutive bin sizes, histograms are binned down until the maximum entry is 50, each bin size should evenly divide its predecessor.
188static const constexpr float RES_AXES[5] = {1., 1., 0.03, 0.03, 1.0};
189static const constexpr float RES_AXES_NATIVE[5] = {1., 1., 0.1, 0.1, 5.0};
190static const constexpr float PULL_AXIS = 10.f;
191
192std::vector<TColor*> GPUQA::mColors;
193int32_t GPUQA::initColors()
194{
195 mColors.reserve(COLORCOUNT);
196 for (int32_t i = 0; i < COLORCOUNT; i++) {
197 float f1 = (float)((COLORS_HEX[i] >> 16) & 0xFF) / (float)0xFF;
198 float f2 = (float)((COLORS_HEX[i] >> 8) & 0xFF) / (float)0xFF;
199 float f3 = (float)((COLORS_HEX[i] >> 0) & 0xFF) / (float)0xFF;
200 mColors.emplace_back(new TColor(10000 + i, f1, f2, f3));
201 }
202 return 0;
203}
204static constexpr Color_t defaultColorNums[COLORCOUNT] = {kRed, kBlue, kGreen, kMagenta, kOrange, kAzure, kBlack, kYellow, kGray, kTeal, kSpring, kPink};
205
206#define TRACK_EXPECTED_REFERENCE_X_DEFAULT 81
207#ifndef GPUCA_RUN2 // Run 3 implementation
208static inline int32_t GPUQA_O2_ConvertFakeLabel(int32_t label) { return label >= 0x7FFFFFFE ? -1 : label; }
209inline uint32_t GPUQA::GetNMCCollissions() const { return mMCInfosCol.size(); }
210inline uint32_t GPUQA::GetNMCTracks(int32_t iCol) const { return mMCInfosCol[iCol].num; }
211inline uint32_t GPUQA::GetNMCTracks(const mcLabelI_t& label) const { return mMCInfosCol[mMCEventOffset[label.getSourceID()] + label.getEventID()].num; }
212inline uint32_t GPUQA::GetNMCLabels() const { return mClNative->clustersMCTruth ? mClNative->clustersMCTruth->getIndexedSize() : 0; }
213inline const GPUQA::mcInfo_t& GPUQA::GetMCTrack(uint32_t iTrk, uint32_t iCol) { return mMCInfos[mMCInfosCol[iCol].first + iTrk]; }
214inline const GPUQA::mcInfo_t& GPUQA::GetMCTrack(const mcLabel_t& label) { return mMCInfos[mMCInfosCol[mMCEventOffset[label.getSourceID()] + label.getEventID()].first + label.getTrackID()]; }
215inline GPUQA::mcLabels_t GPUQA::GetMCLabel(uint32_t i) { return mClNative->clustersMCTruth->getLabels(i); }
216inline int32_t GPUQA::GetMCLabelNID(const mcLabels_t& label) { return label.size(); }
217inline int32_t GPUQA::GetMCLabelNID(uint32_t i) { return mClNative->clustersMCTruth->getLabels(i).size(); }
218inline GPUQA::mcLabel_t GPUQA::GetMCLabel(uint32_t i, uint32_t j) { return mClNative->clustersMCTruth->getLabels(i)[j]; }
219inline int32_t GPUQA::GetMCLabelID(uint32_t i, uint32_t j) { return GPUQA_O2_ConvertFakeLabel(mClNative->clustersMCTruth->getLabels(i)[j].getTrackID()); }
220inline int32_t GPUQA::GetMCLabelID(const mcLabels_t& label, uint32_t j) { return GPUQA_O2_ConvertFakeLabel(label[j].getTrackID()); }
221inline int32_t GPUQA::GetMCLabelID(const mcLabel_t& label) { return GPUQA_O2_ConvertFakeLabel(label.getTrackID()); }
222inline uint32_t GPUQA::GetMCLabelCol(uint32_t i, uint32_t j) { return mMCEventOffset[mClNative->clustersMCTruth->getLabels(i)[j].getSourceID()] + mClNative->clustersMCTruth->getLabels(i)[j].getEventID(); }
223inline const auto& GPUQA::GetClusterLabels() { return mClNative->clustersMCTruth; }
224inline float GPUQA::GetMCLabelWeight(uint32_t i, uint32_t j) { return 1; }
225inline float GPUQA::GetMCLabelWeight(const mcLabels_t& label, uint32_t j) { return 1; }
226inline float GPUQA::GetMCLabelWeight(const mcLabel_t& label) { return 1; }
227inline bool GPUQA::mcPresent() { return !mConfig.noMC && mTracking && mClNative && mClNative->clustersMCTruth && mMCInfos.size(); }
228uint32_t GPUQA::GetMCLabelCol(const mcLabel_t& label) const { return !label.isValid() ? 0 : (mMCEventOffset[label.getSourceID()] + label.getEventID()); }
229GPUQA::mcLabelI_t GPUQA::GetMCTrackLabel(uint32_t trackId) const { return trackId >= mTrackMCLabels.size() ? MCCompLabel() : mTrackMCLabels[trackId]; }
230bool GPUQA::CompareIgnoreFake(const mcLabelI_t& l1, const mcLabelI_t& l2) { return l1.compare(l2) >= 0; }
231#define TRACK_EXPECTED_REFERENCE_X 78
232#else // Run 2 implementation
233inline GPUQA::mcLabelI_t::mcLabelI_t(const GPUQA::mcLabel_t& l) : track(l.fMCID) {}
234inline bool GPUQA::mcLabelI_t::operator==(const GPUQA::mcLabel_t& l) { return AbsLabelID(track) == l.fMCID; }
235inline uint32_t GPUQA::GetNMCCollissions() const { return 1; }
236inline uint32_t GPUQA::GetNMCTracks(int32_t iCol) const { return mTracking->mIOPtrs.nMCInfosTPC; }
237inline uint32_t GPUQA::GetNMCTracks(const mcLabelI_t& label) const { return mTracking->mIOPtrs.nMCInfosTPC; }
238inline uint32_t GPUQA::GetNMCLabels() const { return mTracking->mIOPtrs.nMCLabelsTPC; }
239inline const GPUQA::mcInfo_t& GPUQA::GetMCTrack(uint32_t iTrk, uint32_t iCol) { return mTracking->mIOPtrs.mcInfosTPC[AbsLabelID(iTrk)]; }
240inline const GPUQA::mcInfo_t& GPUQA::GetMCTrack(const mcLabel_t& label) { return GetMCTrack(label.fMCID, 0); }
241inline const GPUQA::mcInfo_t& GPUQA::GetMCTrack(const mcLabelI_t& label) { return GetMCTrack(label.track, 0); }
242inline const GPUQA::mcLabels_t& GPUQA::GetMCLabel(uint32_t i) { return mTracking->mIOPtrs.mcLabelsTPC[i]; }
243inline const GPUQA::mcLabel_t& GPUQA::GetMCLabel(uint32_t i, uint32_t j) { return mTracking->mIOPtrs.mcLabelsTPC[i].fClusterID[j]; }
244inline int32_t GPUQA::GetMCLabelNID(const mcLabels_t& label) { return 3; }
245inline int32_t GPUQA::GetMCLabelNID(uint32_t i) { return 3; }
246inline int32_t GPUQA::GetMCLabelID(uint32_t i, uint32_t j) { return mTracking->mIOPtrs.mcLabelsTPC[i].fClusterID[j].fMCID; }
247inline int32_t GPUQA::GetMCLabelID(const mcLabels_t& label, uint32_t j) { return label.fClusterID[j].fMCID; }
248inline int32_t GPUQA::GetMCLabelID(const mcLabel_t& label) { return label.fMCID; }
249inline uint32_t GPUQA::GetMCLabelCol(uint32_t i, uint32_t j) { return 0; }
250inline const auto& GPUQA::GetClusterLabels() { return mTracking->mIOPtrs.mcLabelsTPC; }
251inline float GPUQA::GetMCLabelWeight(uint32_t i, uint32_t j) { return mTracking->mIOPtrs.mcLabelsTPC[i].fClusterID[j].fWeight; }
252inline float GPUQA::GetMCLabelWeight(const mcLabels_t& label, uint32_t j) { return label.fClusterID[j].fWeight; }
253inline float GPUQA::GetMCLabelWeight(const mcLabel_t& label) { return label.fWeight; }
254inline int32_t GPUQA::FakeLabelID(int32_t id) { return id < 0 ? id : (-2 - id); }
255inline int32_t GPUQA::AbsLabelID(int32_t id) { return id >= 0 ? id : (-id - 2); }
256inline bool GPUQA::mcPresent() { return !mConfig.noMC && mTracking && GetNMCLabels() && GetNMCTracks(0); }
257uint32_t GPUQA::GetMCLabelCol(const mcLabel_t& label) const { return 0; }
258GPUQA::mcLabelI_t GPUQA::GetMCTrackLabel(uint32_t trackId) const { return trackId >= mTrackMCLabels.size() ? mcLabelI_t() : mTrackMCLabels[trackId]; }
259bool GPUQA::CompareIgnoreFake(const mcLabelI_t& l1, const mcLabelI_t& l2) { return AbsLabelID(l1.track) == AbsLabelID(l2.track); }
260#define TRACK_EXPECTED_REFERENCE_X TRACK_EXPECTED_REFERENCE_X_DEFAULT
261#endif
262template <class T>
263inline auto& GPUQA::GetMCTrackObj(T& obj, const GPUQA::mcLabelI_t& l)
264{
265 return obj[mMCEventOffset[l.getSourceID()] + l.getEventID()][l.getTrackID()];
266}
267
268template <>
269auto GPUQA::getHistArray<TH1F>()
270{
271 return std::make_pair(mHist1D, &mHist1D_pos);
272}
273template <>
274auto GPUQA::getHistArray<TH2F>()
275{
276 return std::make_pair(mHist2D, &mHist2D_pos);
277}
278template <>
279auto GPUQA::getHistArray<TH1D>()
280{
281 return std::make_pair(mHist1Dd, &mHist1Dd_pos);
282}
283template <>
284auto GPUQA::getHistArray<TGraphAsymmErrors>()
285{
286 return std::make_pair(mHistGraph, &mHistGraph_pos);
287}
288template <class T, typename... Args>
289void GPUQA::createHist(T*& h, const char* name, Args... args)
290{
291 const auto& p = getHistArray<T>();
292 if (mHaveExternalHists) {
293 if (p.first->size() <= p.second->size()) {
294 GPUError("Array sizes mismatch: Histograms %lu <= Positions %lu", p.first->size(), p.second->size());
295 throw std::runtime_error("Incoming histogram array incomplete");
296 }
297 if (strcmp((*p.first)[p.second->size()].GetName(), name)) {
298 GPUError("Histogram name mismatch: in array %s, trying to create %s", (*p.first)[p.second->size()].GetName(), name);
299 throw std::runtime_error("Incoming histogram has incorrect name");
300 }
301 } else {
302 if constexpr (std::is_same_v<T, TGraphAsymmErrors>) {
303 p.first->emplace_back();
304 p.first->back().SetName(name);
305 } else {
306 p.first->emplace_back(name, args...);
307 }
308 }
309 h = &((*p.first)[p.second->size()]);
310 p.second->emplace_back(&h);
311}
312
313namespace o2::gpu::internal
314{
315struct GPUQAGarbageCollection {
316 std::tuple<std::vector<std::unique_ptr<TCanvas>>, std::vector<std::unique_ptr<TLegend>>, std::vector<std::unique_ptr<TPad>>, std::vector<std::unique_ptr<TLatex>>, std::vector<std::unique_ptr<TH1D>>> v;
317};
318} // namespace o2::gpu::internal
319
320template <class T, typename... Args>
321T* GPUQA::createGarbageCollected(Args... args)
322{
323 auto& v = std::get<std::vector<std::unique_ptr<T>>>(mGarbageCollector->v);
324 v.emplace_back(std::make_unique<T>(args...));
325 return v.back().get();
326}
327void GPUQA::clearGarbagageCollector()
328{
329 std::get<std::vector<std::unique_ptr<TPad>>>(mGarbageCollector->v).clear(); // Make sure to delete TPad first due to ROOT ownership (std::tuple has no defined order in its destructor)
330 std::apply([](auto&&... args) { ((args.clear()), ...); }, mGarbageCollector->v);
331}
332
333GPUQA::GPUQA(GPUChainTracking* chain, const GPUSettingsQA* config, const GPUParam* param) : mTracking(chain), mConfig(config ? *config : GPUQA_GetConfig(chain)), mParam(param ? param : &chain->GetParam()), mGarbageCollector(std::make_unique<internal::GPUQAGarbageCollection>())
334{
335 mMCEventOffset.resize(1, 0);
336}
337
338GPUQA::~GPUQA()
339{
340 if (mQAInitialized && !mHaveExternalHists) {
341 delete mHist1D;
342 delete mHist2D;
343 delete mHist1Dd;
344 delete mHistGraph;
345 }
346 clearGarbagageCollector(); // Needed to guarantee correct order for ROOT ownership
347}
348
349bool GPUQA::clusterRemovable(int32_t attach, bool prot) const
350{
351 const auto& r = checkClusterState<false>(attach);
352 if (prot) {
353 return r.protect || r.physics;
354 }
355 return (!r.unattached && !r.physics && !r.protect);
356}
357
358template <class T>
359void GPUQA::SetAxisSize(T* e)
360{
361 e->GetYaxis()->SetTitleOffset(1.0);
362 e->GetYaxis()->SetTitleSize(0.045);
363 e->GetYaxis()->SetLabelSize(0.045);
364 e->GetXaxis()->SetTitleOffset(1.03);
365 e->GetXaxis()->SetTitleSize(0.045);
366 e->GetXaxis()->SetLabelOffset(-0.005);
367 e->GetXaxis()->SetLabelSize(0.045);
368}
369
370void GPUQA::SetLegend(TLegend* l, bool bigText)
371{
372 l->SetTextFont(72);
373 l->SetTextSize(bigText ? 0.03 : 0.016);
374 l->SetFillColor(0);
375}
376
377double* GPUQA::CreateLogAxis(int32_t nbins, float xmin, float xmax)
378{
379 float logxmin = std::log10(xmin);
380 float logxmax = std::log10(xmax);
381 float binwidth = (logxmax - logxmin) / nbins;
382
383 double* xbins = new double[nbins + 1];
384
385 xbins[0] = xmin;
386 for (int32_t i = 1; i <= nbins; i++) {
387 xbins[i] = std::pow(10, logxmin + i * binwidth);
388 }
389 return xbins;
390}
391
392void GPUQA::ChangePadTitleSize(TPad* p, float size)
393{
394 p->Update();
395 TPaveText* pt = (TPaveText*)(p->GetPrimitive("title"));
396 if (pt == nullptr) {
397 GPUError("Error changing title");
398 } else {
399 pt->SetTextSize(size);
400 p->Modified();
401 }
402}
403
404void GPUQA::DrawHisto(TH1* histo, char* filename, char* options)
405{
406 TCanvas tmp;
407 tmp.cd();
408 histo->Draw(options);
409 tmp.Print(filename);
410}
411
412void GPUQA::doPerfFigure(float x, float y, float size)
413{
414 if (mConfig.perfFigure == "") {
415 return;
416 }
417 static constexpr const char* str_perf_figure_1 = "ALICE Performance";
418 static constexpr const char* str_perf_figure_2_mc = "MC, Pb#minusPb, #sqrt{s_{NN}} = 5.36 TeV";
419 static constexpr const char* str_perf_figure_2_data = "Pb#minusPb, #sqrt{s_{NN}} = 5.36 TeV";
420 const char* str_perf_figure_2 = (mConfig.perfFigure == "mc" || mConfig.perfFigure == "MC") ? str_perf_figure_2_mc : (mConfig.perfFigure == "data" ? str_perf_figure_2_data : mConfig.perfFigure.c_str());
421
422 TLatex* t = createGarbageCollected<TLatex>(); // TODO: We could perhaps put everything in a legend, to get a white background if there is a grid
423 t->SetNDC(kTRUE);
424 t->SetTextColor(1);
425 t->SetTextSize(size);
426 t->DrawLatex(x, y, str_perf_figure_1);
427 t->SetTextSize(size * 0.8);
428 t->DrawLatex(x, y - 0.01 - size, str_perf_figure_2);
429}
430
431void GPUQA::SetMCTrackRange(int32_t min, int32_t max)
432{
433 mMCTrackMin = min;
434 mMCTrackMax = max;
435}
436
437int32_t GPUQA::InitQACreateHistograms()
438{
439 char name[2048], fname[1024];
440 if (mQATasks & taskTrackingEff) {
441 // Create Efficiency Histograms
442 for (int32_t i = 0; i < 6; i++) {
443 for (int32_t j = 0; j < 2; j++) {
444 for (int32_t k = 0; k < 2; k++) {
445 for (int32_t l = 0; l < 5; l++) {
446 snprintf(name, 2048, "%s%s%s%sVs%s", "tracks", EFF_TYPES[i], FINDABLE_NAMES[j], PRIM_NAMES[k], VSPARAMETER_NAMES[l]);
447 if (l == 4) {
448 std::unique_ptr<double[]> binsPt{CreateLogAxis(AXIS_BINS[4], k == 0 ? PT_MIN_PRIM : AXES_MIN[4], AXES_MAX[4])};
449 createHist(mEff[i][j][k][l], name, name, AXIS_BINS[l], binsPt.get());
450 } else {
451 createHist(mEff[i][j][k][l], name, name, AXIS_BINS[l], AXES_MIN[l], AXES_MAX[l]);
452 }
453 if (!mHaveExternalHists) {
454 mEff[i][j][k][l]->Sumw2();
455 }
456 strcat(name, "_eff");
457 if (i < 4) {
458 createHist(mEffResult[i][j][k][l], name);
459 }
460 }
461 }
462 }
463 }
464 }
465
466 // Create Resolution Histograms
467 if (mQATasks & taskTrackingRes) {
468 for (int32_t i = 0; i < 5; i++) {
469 for (int32_t j = 0; j < 5; j++) {
470 snprintf(name, 2048, "rms_%s_vs_%s", VSPARAMETER_NAMES[i], VSPARAMETER_NAMES[j]);
471 snprintf(fname, 1024, "mean_%s_vs_%s", VSPARAMETER_NAMES[i], VSPARAMETER_NAMES[j]);
472 if (j == 4) {
473 std::unique_ptr<double[]> binsPt{CreateLogAxis(AXIS_BINS[4], mConfig.resPrimaries == 1 ? PT_MIN_PRIM : AXES_MIN[4], AXES_MAX[4])};
474 createHist(mRes[i][j][0], name, name, AXIS_BINS[j], binsPt.get());
475 createHist(mRes[i][j][1], fname, fname, AXIS_BINS[j], binsPt.get());
476 } else {
477 createHist(mRes[i][j][0], name, name, AXIS_BINS[j], AXES_MIN[j], AXES_MAX[j]);
478 createHist(mRes[i][j][1], fname, fname, AXIS_BINS[j], AXES_MIN[j], AXES_MAX[j]);
479 }
480 snprintf(name, 2048, "res_%s_vs_%s", VSPARAMETER_NAMES[i], VSPARAMETER_NAMES[j]);
481 const float* axis = mConfig.nativeFitResolutions ? RES_AXES_NATIVE : RES_AXES;
482 const int32_t nbins = i == 4 && mConfig.nativeFitResolutions ? (10 * RES_AXIS_BINS[0]) : RES_AXIS_BINS[0];
483 if (j == 4) {
484 std::unique_ptr<double[]> binsPt{CreateLogAxis(AXIS_BINS[4], mConfig.resPrimaries == 1 ? PT_MIN_PRIM : AXES_MIN[4], AXES_MAX[4])};
485 createHist(mRes2[i][j], name, name, nbins, -axis[i], axis[i], AXIS_BINS[j], binsPt.get());
486 } else {
487 createHist(mRes2[i][j], name, name, nbins, -axis[i], axis[i], AXIS_BINS[j], AXES_MIN[j], AXES_MAX[j]);
488 }
489 }
490 }
491 }
492
493 // Create Pull Histograms
494 if (mQATasks & taskTrackingResPull) {
495 for (int32_t i = 0; i < 5; i++) {
496 for (int32_t j = 0; j < 5; j++) {
497 snprintf(name, 2048, "pull_rms_%s_vs_%s", VSPARAMETER_NAMES[i], VSPARAMETER_NAMES[j]);
498 snprintf(fname, 1024, "pull_mean_%s_vs_%s", VSPARAMETER_NAMES[i], VSPARAMETER_NAMES[j]);
499 if (j == 4) {
500 std::unique_ptr<double[]> binsPt{CreateLogAxis(AXIS_BINS[4], AXES_MIN[4], AXES_MAX[4])};
501 createHist(mPull[i][j][0], name, name, AXIS_BINS[j], binsPt.get());
502 createHist(mPull[i][j][1], fname, fname, AXIS_BINS[j], binsPt.get());
503 } else {
504 createHist(mPull[i][j][0], name, name, AXIS_BINS[j], AXES_MIN[j], AXES_MAX[j]);
505 createHist(mPull[i][j][1], fname, fname, AXIS_BINS[j], AXES_MIN[j], AXES_MAX[j]);
506 }
507 snprintf(name, 2048, "pull_%s_vs_%s", VSPARAMETER_NAMES[i], VSPARAMETER_NAMES[j]);
508 if (j == 4) {
509 std::unique_ptr<double[]> binsPt{CreateLogAxis(AXIS_BINS[4], AXES_MIN[4], AXES_MAX[4])};
510 createHist(mPull2[i][j], name, name, RES_AXIS_BINS[0], -PULL_AXIS, PULL_AXIS, AXIS_BINS[j], binsPt.get());
511 } else {
512 createHist(mPull2[i][j], name, name, RES_AXIS_BINS[0], -PULL_AXIS, PULL_AXIS, AXIS_BINS[j], AXES_MIN[j], AXES_MAX[j]);
513 }
514 }
515 }
516 }
517
518 // Create Cluster Histograms
519 if (mQATasks & taskClusterAttach) {
520 for (int32_t i = 0; i < N_CLS_TYPE * N_CLS_HIST - 1; i++) {
521 int32_t ioffset = i >= (2 * N_CLS_HIST - 1) ? (2 * N_CLS_HIST - 1) : i >= N_CLS_HIST ? N_CLS_HIST : 0;
522 int32_t itype = i >= (2 * N_CLS_HIST - 1) ? 2 : i >= N_CLS_HIST ? 1 : 0;
523 snprintf(name, 2048, "clusters%s%s", CLUSTER_NAMES_SHORT[i - ioffset], CLUSTER_TYPES[itype]);
524 std::unique_ptr<double[]> binsPt{CreateLogAxis(AXIS_BINS[4], PT_MIN_CLUST, PT_MAX)};
525 createHist(mClusters[i], name, name, AXIS_BINS[4], binsPt.get());
526 }
527
528 createHist(mPadRow[0], "padrow0", "padrow0", GPUTPCGeometry::NROWS - PADROW_CHECK_MINCLS, 0, GPUTPCGeometry::NROWS - 1 - PADROW_CHECK_MINCLS, GPUTPCGeometry::NROWS - PADROW_CHECK_MINCLS, 0, GPUTPCGeometry::NROWS - 1 - PADROW_CHECK_MINCLS);
529 createHist(mPadRow[1], "padrow1", "padrow1", 100.f, -0.2f, 0.2f, GPUTPCGeometry::NROWS - PADROW_CHECK_MINCLS, 0, GPUTPCGeometry::NROWS - 1 - PADROW_CHECK_MINCLS);
530 createHist(mPadRow[2], "padrow2", "padrow2", 100.f, -0.2f, 0.2f, GPUTPCGeometry::NROWS - PADROW_CHECK_MINCLS, 0, GPUTPCGeometry::NROWS - 1 - PADROW_CHECK_MINCLS);
531 createHist(mPadRow[3], "padrow3", "padrow3", 100.f, 0, 300000, GPUTPCGeometry::NROWS - PADROW_CHECK_MINCLS, 0, GPUTPCGeometry::NROWS - 1 - PADROW_CHECK_MINCLS);
532 }
533
534 if (mQATasks & taskTrackStatistics) {
535 // Create Tracks Histograms
536 for (int32_t i = 0; i < 2; i++) {
537 snprintf(name, 2048, i ? "nrows_with_cluster" : "nclusters");
538 createHist(mNCl[i], name, name, GPUTPCGeometry::NROWS + 1, 0, GPUTPCGeometry::NROWS);
539 }
540 std::unique_ptr<double[]> binsPt{CreateLogAxis(AXIS_BINS[4], PT_MIN_CLUST, PT_MAX)};
541 createHist(mTrackPt, "tracks_pt", "tracks_pt", AXIS_BINS[4], binsPt.get());
542 for (int32_t i = 0; i < 2; i++) {
543 snprintf(name, 2048, i ? "tracks_dedx_max" : "tracks_dedx_tot");
544 createHist(mTrackdEdx[i], name, name, 200, -3, 1, 1000, 0, i ? 1500 : 5000);
545 }
546 const uint32_t maxTime = (mTracking && mTracking->GetParam().continuousMaxTimeBin > 0) ? mTracking->GetParam().continuousMaxTimeBin : constants::TPC_MAX_TIME_BIN_TRIGGERED;
547 createHist(mT0[0], "tracks_t0", "tracks_t0", (maxTime + 1) / 10, 0, maxTime);
548 createHist(mT0[1], "tracks_t0_res", "tracks_t0_res", 1000, -100, 100);
549 createHist(mClXY, "clXY", "clXY", 1000, -250, 250, 1000, -250, 250); // TODO: Pass name only once
550 }
551 if (mQATasks & taskClusterRejection) {
552 const int padCount = GPUTPCGeometry::NPads(GPUTPCGeometry::NROWS - 1);
553 for (int32_t i = 0; i < 3; i++) {
554 snprintf(name, 2048, "clrej_%d", i);
555 createHist(mClRej[i], name, name, 2 * padCount, -padCount / 2 + 0.5f, padCount / 2 - 0.5f, GPUTPCGeometry::NROWS, 0, GPUTPCGeometry::NROWS - 1);
556 }
557 createHist(mClRejP, "clrejp", "clrejp", GPUTPCGeometry::NROWS, 0, GPUTPCGeometry::NROWS - 1);
558 }
559
560 if ((mQATasks & taskClusterCounts) && mConfig.clusterRejectionHistograms) {
561 int32_t num = DoClusterCounts(nullptr, 2);
562 mHistClusterCount.resize(num);
563 DoClusterCounts(nullptr, 1);
564 }
565
566 for (uint32_t i = 0; i < mHist1D->size(); i++) {
567 *mHist1D_pos[i] = &(*mHist1D)[i];
568 }
569 for (uint32_t i = 0; i < mHist2D->size(); i++) {
570 *mHist2D_pos[i] = &(*mHist2D)[i];
571 }
572 for (uint32_t i = 0; i < mHist1Dd->size(); i++) {
573 *mHist1Dd_pos[i] = &(*mHist1Dd)[i];
574 }
575 for (uint32_t i = 0; i < mHistGraph->size(); i++) {
576 *mHistGraph_pos[i] = &(*mHistGraph)[i];
577 }
578
579 return 0;
580}
581
582int32_t GPUQA::loadHistograms(std::vector<TH1F>& i1, std::vector<TH2F>& i2, std::vector<TH1D>& i3, std::vector<TGraphAsymmErrors>& i4, int32_t tasks)
583{
584 if (tasks == tasksAutomatic) {
585 tasks = tasksDefaultPostprocess;
586 }
587 if (mQAInitialized && (!mHaveExternalHists || tasks != mQATasks)) {
588 throw std::runtime_error("QA not initialized or initialized with different task array");
589 }
590 mHist1D = &i1;
591 mHist2D = &i2;
592 mHist1Dd = &i3;
593 mHistGraph = &i4;
594 mHist1D_pos.clear();
595 mHist2D_pos.clear();
596 mHist1Dd_pos.clear();
597 mHistGraph_pos.clear();
598 mHaveExternalHists = true;
599 if (mConfig.noMC) {
600 tasks &= tasksAllNoQC;
601 }
602 mQATasks = tasks;
603 if (InitQACreateHistograms()) {
604 return 1;
605 }
606 mQAInitialized = true;
607 return 0;
608}
609
610void GPUQA::DumpO2MCData(const char* filename) const
611{
612 FILE* fp = fopen(filename, "w+b");
613 if (fp == nullptr) {
614 return;
615 }
616 uint32_t n = mMCInfos.size();
617 fwrite(&n, sizeof(n), 1, fp);
618 fwrite(mMCInfos.data(), sizeof(mMCInfos[0]), n, fp);
619 n = mMCInfosCol.size();
620 fwrite(&n, sizeof(n), 1, fp);
621 fwrite(mMCInfosCol.data(), sizeof(mMCInfosCol[0]), n, fp);
622 n = mMCEventOffset.size();
623 fwrite(&n, sizeof(n), 1, fp);
624 fwrite(mMCEventOffset.data(), sizeof(mMCEventOffset[0]), n, fp);
625 fclose(fp);
626}
627
628int32_t GPUQA::ReadO2MCData(const char* filename)
629{
630 FILE* fp = fopen(filename, "rb");
631 if (fp == nullptr) {
632 return 1;
633 }
634 uint32_t n;
635 uint32_t x;
636 if ((x = fread(&n, sizeof(n), 1, fp)) != 1) {
637 fclose(fp);
638 return 1;
639 }
640 mMCInfos.resize(n);
641 if (fread(mMCInfos.data(), sizeof(mMCInfos[0]), n, fp) != n) {
642 fclose(fp);
643 return 1;
644 }
645 if ((x = fread(&n, sizeof(n), 1, fp)) != 1) {
646 fclose(fp);
647 return 1;
648 }
649 mMCInfosCol.resize(n);
650 if (fread(mMCInfosCol.data(), sizeof(mMCInfosCol[0]), n, fp) != n) {
651 fclose(fp);
652 return 1;
653 }
654 if ((x = fread(&n, sizeof(n), 1, fp)) != 1) {
655 fclose(fp);
656 return 1;
657 }
658 mMCEventOffset.resize(n);
659 if (fread(mMCEventOffset.data(), sizeof(mMCEventOffset[0]), n, fp) != n) {
660 fclose(fp);
661 return 1;
662 }
663 if (mTracking && mTracking->GetProcessingSettings().debugLevel >= 2) {
664 printf("Read %ld bytes MC Infos\n", ftell(fp));
665 }
666 fclose(fp);
667 if (mTracking) {
668 CopyO2MCtoIOPtr(&mTracking->mIOPtrs);
669 }
670 return 0;
671}
672
673void GPUQA::CopyO2MCtoIOPtr(GPUTrackingInOutPointers* ptr)
674{
675 ptr->mcInfosTPC = mMCInfos.data();
676 ptr->nMCInfosTPC = mMCInfos.size();
677 ptr->mcInfosTPCCol = mMCInfosCol.data();
678 ptr->nMCInfosTPCCol = mMCInfosCol.size();
679}
680
681void GPUQA::InitO2MCData(GPUTrackingInOutPointers* updateIOPtr)
682{
683#ifndef GPUCA_STANDALONE
684 if (!mO2MCDataLoaded) {
685 HighResTimer timer(mTracking && mTracking->GetProcessingSettings().debugLevel);
686 if (mTracking && mTracking->GetProcessingSettings().debugLevel) {
687 GPUInfo("Start reading O2 Track MC information");
688 }
689 static constexpr float PRIM_MAX_T = 0.01f;
690
691 o2::steer::MCKinematicsReader mcReader("collisioncontext.root");
692 std::vector<int32_t> refId;
693
694 auto dc = o2::steer::DigitizationContext::loadFromFile("collisioncontext.root");
695 const auto& evrec = dc->getEventRecords();
696 const auto& evparts = dc->getEventParts();
697 std::vector<std::vector<float>> evTimeBins(mcReader.getNSources());
698 for (uint32_t i = 0; i < evTimeBins.size(); i++) {
699 evTimeBins[i].resize(mcReader.getNEvents(i), -100.f);
700 }
701 for (uint32_t i = 0; i < evrec.size(); i++) {
702 const auto& ir = evrec[i];
703 for (uint32_t j = 0; j < evparts[i].size(); j++) {
704 const int iSim = evparts[i][j].sourceID;
705 const int iEv = evparts[i][j].entryID;
706 if (iSim == o2::steer::QEDSOURCEID || ir.differenceInBC(o2::raw::HBFUtils::Instance().getFirstIR()) >= 0) {
707 auto ir0 = o2::raw::HBFUtils::Instance().getFirstIRofTF(ir);
708 float timebin = (float)ir.differenceInBC(ir0) / o2::tpc::constants::LHCBCPERTIMEBIN;
709 if (evTimeBins[iSim][iEv] >= 0) {
710 throw std::runtime_error("Multiple time bins for same MC collision found");
711 }
712 evTimeBins[iSim][iEv] = timebin;
713 }
714 }
715 }
716
717 uint32_t nSimSources = mcReader.getNSources();
718 mMCEventOffset.resize(nSimSources);
719 uint32_t nSimTotalEvents = 0;
720 uint32_t nSimTotalTracks = 0;
721 for (uint32_t i = 0; i < nSimSources; i++) {
722 mMCEventOffset[i] = nSimTotalEvents;
723 nSimTotalEvents += mcReader.getNEvents(i);
724 }
725
726 mMCInfosCol.resize(nSimTotalEvents);
727 for (int32_t iSim = 0; iSim < mcReader.getNSources(); iSim++) {
728 for (int32_t i = 0; i < mcReader.getNEvents(iSim); i++) {
729 const float timebin = evTimeBins[iSim][i];
730
731 const std::vector<o2::MCTrack>& tracks = mcReader.getTracks(iSim, i);
732 const std::vector<o2::TrackReference>& trackRefs = mcReader.getTrackRefsByEvent(iSim, i);
733
734 refId.resize(tracks.size());
735 std::fill(refId.begin(), refId.end(), -1);
736 for (uint32_t j = 0; j < trackRefs.size(); j++) {
737 if (trackRefs[j].getDetectorId() == o2::detectors::DetID::TPC) {
738 int32_t trkId = trackRefs[j].getTrackID();
739 if (refId[trkId] == -1) {
740 refId[trkId] = j;
741 }
742 }
743 }
744 mMCInfosCol[mMCEventOffset[iSim] + i].first = mMCInfos.size();
745 mMCInfosCol[mMCEventOffset[iSim] + i].num = tracks.size();
746 mMCInfos.resize(mMCInfos.size() + tracks.size());
747 for (uint32_t j = 0; j < tracks.size(); j++) {
748 auto& info = mMCInfos[mMCInfosCol[mMCEventOffset[iSim] + i].first + j];
749 const auto& trk = tracks[j];
750 TParticlePDG* particle = TDatabasePDG::Instance()->GetParticle(trk.GetPdgCode());
751 Int_t pid = -1;
752 if (abs(trk.GetPdgCode()) == kElectron) {
753 pid = 0;
754 }
755 if (abs(trk.GetPdgCode()) == kMuonMinus) {
756 pid = 1;
757 }
758 if (abs(trk.GetPdgCode()) == kPiPlus) {
759 pid = 2;
760 }
761 if (abs(trk.GetPdgCode()) == kKPlus) {
762 pid = 3;
763 }
764 if (abs(trk.GetPdgCode()) == kProton) {
765 pid = 4;
766 }
767
768 info.charge = particle ? particle->Charge() : 0;
769 info.prim = trk.T() < PRIM_MAX_T;
770 info.primDaughters = 0;
771 if (trk.getFirstDaughterTrackId() != -1) {
772 for (int32_t k = trk.getFirstDaughterTrackId(); k <= trk.getLastDaughterTrackId(); k++) {
773 if (tracks[k].T() < PRIM_MAX_T) {
774 info.primDaughters = 1;
775 break;
776 }
777 }
778 }
779 info.pid = pid;
780 info.t0 = timebin;
781 if (refId[j] >= 0) {
782 const auto& trkRef = trackRefs[refId[j]];
783 info.x = trkRef.X();
784 info.y = trkRef.Y();
785 info.z = trkRef.Z();
786 info.pX = trkRef.Px();
787 info.pY = trkRef.Py();
788 info.pZ = trkRef.Pz();
789 info.genRadius = std::sqrt(trk.GetStartVertexCoordinatesX() * trk.GetStartVertexCoordinatesX() + trk.GetStartVertexCoordinatesY() * trk.GetStartVertexCoordinatesY() + trk.GetStartVertexCoordinatesZ() * trk.GetStartVertexCoordinatesZ());
790 } else {
791 info.x = info.y = info.z = info.pX = info.pY = info.pZ = 0;
792 info.genRadius = 0;
793 }
794 }
795 }
796 }
797 if (timer.IsRunning()) {
798 GPUInfo("Finished reading O2 Track MC information (%f seconds)", timer.GetCurrentElapsedTime());
799 }
800 mO2MCDataLoaded = true;
801 }
802 if (updateIOPtr) {
803 CopyO2MCtoIOPtr(updateIOPtr);
804 }
805#endif
806}
807
808int32_t GPUQA::InitQA(int32_t tasks)
809{
810 if (mQAInitialized) {
811 throw std::runtime_error("QA already initialized");
812 }
813 if (tasks == tasksAutomatic) {
814 tasks = tasksDefault;
815 }
816
817 mHist1D = new std::vector<TH1F>;
818 mHist2D = new std::vector<TH2F>;
819 mHist1Dd = new std::vector<TH1D>;
820 mHistGraph = new std::vector<TGraphAsymmErrors>;
821 if (mConfig.noMC) {
822 tasks &= tasksAllNoQC;
823 }
824 mQATasks = tasks;
825
826 if (mTracking->GetProcessingSettings().qcRunFraction != 100.f && mQATasks != taskClusterCounts) {
827 throw std::runtime_error("QA with qcRunFraction only supported for taskClusterCounts");
828 }
829
830 if (mTracking) {
831 mClNative = mTracking->mIOPtrs.clustersNative;
832 }
833
834 if (InitQACreateHistograms()) {
835 return 1;
836 }
837
838 if (mConfig.enableLocalOutput) {
839 mkdir(mConfig.plotsDir.c_str(), S_IRWXU | S_IRWXG | S_IROTH | S_IXOTH);
840 }
841
842#ifndef GPUCA_STANDALONE
843 if (!mConfig.noMC) {
844 InitO2MCData(mTracking ? &mTracking->mIOPtrs : nullptr);
845 }
846#endif
847
848 if (mConfig.matchMCLabels.size()) {
849 uint32_t nFiles = mConfig.matchMCLabels.size();
850 std::vector<std::unique_ptr<TFile>> files;
851 std::vector<std::vector<std::vector<int32_t>>*> labelsBuffer(nFiles);
852 std::vector<std::vector<std::vector<int32_t>>*> effBuffer(nFiles);
853 for (uint32_t i = 0; i < nFiles; i++) {
854 files.emplace_back(std::make_unique<TFile>(mConfig.matchMCLabels[i].c_str()));
855 labelsBuffer[i] = (std::vector<std::vector<int32_t>>*)files[i]->Get("mcLabelBuffer");
856 effBuffer[i] = (std::vector<std::vector<int32_t>>*)files[i]->Get("mcEffBuffer");
857 if (labelsBuffer[i] == nullptr || effBuffer[i] == nullptr) {
858 GPUError("Error opening / reading from labels file %u/%s: %p %p", i, mConfig.matchMCLabels[i].c_str(), (void*)labelsBuffer[i], (void*)effBuffer[i]);
859 exit(1);
860 }
861 }
862
863 mGoodTracks.resize(labelsBuffer[0]->size());
864 mGoodHits.resize(labelsBuffer[0]->size());
865 for (uint32_t iEvent = 0; iEvent < labelsBuffer[0]->size(); iEvent++) {
866 std::vector<bool> labelsOK((*effBuffer[0])[iEvent].size());
867 for (uint32_t k = 0; k < (*effBuffer[0])[iEvent].size(); k++) {
868 labelsOK[k] = false;
869 for (uint32_t l = 0; l < nFiles; l++) {
870 if ((*effBuffer[0])[iEvent][k] != (*effBuffer[l])[iEvent][k]) {
871 labelsOK[k] = true;
872 break;
873 }
874 }
875 }
876 mGoodTracks[iEvent].resize((*labelsBuffer[0])[iEvent].size());
877 for (uint32_t k = 0; k < (*labelsBuffer[0])[iEvent].size(); k++) {
878 if ((*labelsBuffer[0])[iEvent][k] == MC_LABEL_INVALID) {
879 continue;
880 }
881 mGoodTracks[iEvent][k] = labelsOK[abs((*labelsBuffer[0])[iEvent][k])];
882 }
883 }
884 }
885 mQAInitialized = true;
886 return 0;
887}
888
889void GPUQA::RunQA(bool matchOnly, const std::vector<o2::tpc::TrackTPC>* tracksExternal, const std::vector<o2::MCCompLabel>* tracksExtMC, const o2::tpc::ClusterNativeAccess* clNative)
890{
891 if (!mQAInitialized) {
892 throw std::runtime_error("QA not initialized");
893 }
894 if (mTracking && mTracking->GetProcessingSettings().debugLevel >= 2) {
895 GPUInfo("Running QA - Mask %d, Efficiency %d, Resolution %d, Pulls %d, Cluster Attachment %d, Track Statistics %d, Cluster Counts %d", mQATasks, (int32_t)(mQATasks & taskTrackingEff), (int32_t)(mQATasks & taskTrackingRes), (int32_t)(mQATasks & taskTrackingResPull), (int32_t)(mQATasks & taskClusterAttach), (int32_t)(mQATasks & taskTrackStatistics), (int32_t)(mQATasks & taskClusterCounts));
896 }
897 if (!clNative && mTracking) {
898 clNative = mTracking->mIOPtrs.clustersNative;
899 }
900 mClNative = clNative;
901
902#ifndef GPUCA_RUN2
903 uint32_t nSimEvents = GetNMCCollissions();
904 if (mTrackMCLabelsReverse.size() < nSimEvents) {
905 mTrackMCLabelsReverse.resize(nSimEvents);
906 }
907 if (mRecTracks.size() < nSimEvents) {
908 mRecTracks.resize(nSimEvents);
909 }
910 if (mFakeTracks.size() < nSimEvents) {
911 mFakeTracks.resize(nSimEvents);
912 }
913 if (mMCParam.size() < nSimEvents) {
914 mMCParam.resize(nSimEvents);
915 }
916#endif
917
918 // Initialize Arrays
919 uint32_t nReconstructedTracks = 0;
920 if (tracksExternal) {
921#ifndef GPUCA_STANDALONE
922 nReconstructedTracks = tracksExternal->size();
923#endif
924 } else {
925 nReconstructedTracks = mTracking->mIOPtrs.nMergedTracks;
926 }
927 mTrackMCLabels.resize(nReconstructedTracks);
928 for (uint32_t iCol = 0; iCol < GetNMCCollissions(); iCol++) {
929 mTrackMCLabelsReverse[iCol].resize(GetNMCTracks(iCol));
930 mRecTracks[iCol].resize(GetNMCTracks(iCol));
931 mFakeTracks[iCol].resize(GetNMCTracks(iCol));
932 mMCParam[iCol].resize(GetNMCTracks(iCol));
933 memset(mRecTracks[iCol].data(), 0, mRecTracks[iCol].size() * sizeof(mRecTracks[iCol][0]));
934 memset(mFakeTracks[iCol].data(), 0, mFakeTracks[iCol].size() * sizeof(mFakeTracks[iCol][0]));
935 for (size_t i = 0; i < mTrackMCLabelsReverse[iCol].size(); i++) {
936 mTrackMCLabelsReverse[iCol][i] = -1;
937 }
938 }
939 if (mQATasks & taskClusterAttach && GetNMCLabels()) {
940 mClusterParam.resize(GetNMCLabels());
941 memset(mClusterParam.data(), 0, mClusterParam.size() * sizeof(mClusterParam[0]));
942 }
943 HighResTimer timer(QA_TIMING || (mTracking && mTracking->GetProcessingSettings().debugLevel >= 2));
944
945 mNEvents++;
946 if (mConfig.writeMCLabels) {
947 mcEffBuffer.resize(mNEvents);
948 mcLabelBuffer.resize(mNEvents);
949 mcEffBuffer[mNEvents - 1].resize(GetNMCTracks(0));
950 mcLabelBuffer[mNEvents - 1].resize(nReconstructedTracks);
951 }
952
953 bool mcAvail = mcPresent() || tracksExtMC;
954
955 if (mcAvail) { // Assign Track MC Labels
956 if (tracksExternal) {
957#ifndef GPUCA_STANDALONE
958 for (uint32_t i = 0; i < tracksExternal->size(); i++) {
959 mTrackMCLabels[i] = (*tracksExtMC)[i];
960 }
961#endif
962 } else {
963 tbb::parallel_for(tbb::blocked_range<uint32_t>(0, nReconstructedTracks, (QA_DEBUG == 0) ? 32 : nReconstructedTracks), [&](const tbb::blocked_range<uint32_t>& range) {
964 auto acc = GPUTPCTrkLbl<true, mcLabelI_t>(GetClusterLabels(), 1.f - mConfig.recThreshold);
965 for (auto i = range.begin(); i < range.end(); i++) {
966 acc.reset();
967 int32_t nClusters = 0;
968 const GPUTPCGMMergedTrack& track = mTracking->mIOPtrs.mergedTracks[i];
969 std::vector<mcLabel_t> labels;
970 for (uint32_t k = 0; k < track.NClusters(); k++) {
971 if (mTracking->mIOPtrs.mergedTrackHits[track.FirstClusterRef() + k].state & GPUTPCGMMergedTrackHit::flagReject) {
972 continue;
973 }
974 nClusters++;
975 uint32_t hitId = mTracking->mIOPtrs.mergedTrackHits[track.FirstClusterRef() + k].num;
976 if (hitId >= GetNMCLabels()) {
977 GPUError("Invalid hit id %u > %d (nClusters %d)", hitId, GetNMCLabels(), clNative ? clNative->nClustersTotal : 0);
978 throw std::runtime_error("qa error");
979 }
980 acc.addLabel(hitId);
981 for (int32_t j = 0; j < GetMCLabelNID(hitId); j++) {
982 if (GetMCLabelID(hitId, j) >= (int32_t)GetNMCTracks(GetMCLabelCol(hitId, j))) {
983 GPUError("Invalid label %d > %d (hit %d, label %d, col %d)", GetMCLabelID(hitId, j), GetNMCTracks(GetMCLabelCol(hitId, j)), hitId, j, (int32_t)GetMCLabelCol(hitId, j));
984 throw std::runtime_error("qa error");
985 }
986 if (GetMCLabelID(hitId, j) >= 0) {
987 if (QA_DEBUG >= 3 && track.OK()) {
988 GPUInfo("Track %d Cluster %u Label %d: %d (%f)", i, k, j, GetMCLabelID(hitId, j), GetMCLabelWeight(hitId, j));
989 }
990 }
991 }
992 }
993
994 float maxweight, sumweight;
995 int32_t maxcount;
996 auto maxLabel = acc.computeLabel(&maxweight, &sumweight, &maxcount);
997 mTrackMCLabels[i] = maxLabel;
998 if (QA_DEBUG && track.OK() && GetNMCTracks(maxLabel) > (uint32_t)maxLabel.getTrackID()) {
999 const mcInfo_t& mc = GetMCTrack(maxLabel);
1000 GPUInfo("Track %d label %d (fake %d) weight %f clusters %d (fitted %d) (%f%% %f%%) Pt %f", i, maxLabel.getTrackID(), (int32_t)(maxLabel.isFake()), maxweight, nClusters, track.NClustersFitted(), 100.f * maxweight / sumweight, 100.f * (float)maxcount / (float)nClusters,
1001 std::sqrt(mc.pX * mc.pX + mc.pY * mc.pY));
1002 }
1003 }
1004 });
1005 }
1006 if (timer.IsRunning()) {
1007 GPUInfo("QA Time: Assign Track Labels:\t\t%6.0f us", timer.GetCurrentElapsedTime(true) * 1e6);
1008 }
1009
1010 for (uint32_t i = 0; i < nReconstructedTracks; i++) {
1011 const GPUTPCGMMergedTrack* track = mTracking ? &mTracking->mIOPtrs.mergedTracks[i] : nullptr;
1012 mcLabelI_t label = mTrackMCLabels[i];
1013 if (mQATasks & taskClusterAttach) {
1014 // fill cluster attachment status
1015 if (!track->OK()) {
1016 continue;
1017 }
1018 if (!mTrackMCLabels[i].isValid()) {
1019 for (uint32_t k = 0; k < track->NClusters(); k++) {
1020 if (mTracking->mIOPtrs.mergedTrackHits[track->FirstClusterRef() + k].state & GPUTPCGMMergedTrackHit::flagReject) {
1021 continue;
1022 }
1023 mClusterParam[mTracking->mIOPtrs.mergedTrackHits[track->FirstClusterRef() + k].num].fakeAttached++;
1024 }
1025 continue;
1026 }
1027 if (mMCTrackMin == -1 || (label.getTrackID() >= mMCTrackMin && label.getTrackID() < mMCTrackMax)) {
1028 for (uint32_t k = 0; k < track->NClusters(); k++) {
1029 if (mTracking->mIOPtrs.mergedTrackHits[track->FirstClusterRef() + k].state & GPUTPCGMMergedTrackHit::flagReject) {
1030 continue;
1031 }
1032 int32_t hitId = mTracking->mIOPtrs.mergedTrackHits[track->FirstClusterRef() + k].num;
1033 bool correct = false;
1034 for (int32_t j = 0; j < GetMCLabelNID(hitId); j++) {
1035 if (label == GetMCLabel(hitId, j)) {
1036 correct = true;
1037 break;
1038 }
1039 }
1040 if (correct) {
1041 mClusterParam[hitId].attached++;
1042 } else {
1043 mClusterParam[hitId].fakeAttached++;
1044 }
1045 }
1046 }
1047 }
1048
1049 if (mTrackMCLabels[i].isFake()) {
1050 (GetMCTrackObj(mFakeTracks, label))++;
1051 } else if (tracksExternal || !track->MergedLooper()) {
1052 GetMCTrackObj(mRecTracks, label)++;
1053 if (mMCTrackMin == -1 || (label.getTrackID() >= mMCTrackMin && label.getTrackID() < mMCTrackMax)) {
1054 int32_t& revLabel = GetMCTrackObj(mTrackMCLabelsReverse, label);
1055 if (tracksExternal) {
1056#ifndef GPUCA_STANDALONE
1057 if (revLabel == -1 || fabsf((*tracksExternal)[i].getZ()) < fabsf((*tracksExternal)[revLabel].getZ())) {
1058 revLabel = i;
1059 }
1060#endif
1061 } else {
1062 const auto* trks = mTracking->mIOPtrs.mergedTracks;
1063 bool comp;
1064 if (revLabel == -1) {
1065 comp = true;
1066 } else {
1067 float shift1 = mTracking->GetTPCTransform()->convDeltaTimeToDeltaZinTimeFrame(trks[i].CSide() * GPUChainTracking::NSECTORS / 2, trks[i].GetParam().GetTOffset());
1068 float shift2 = mTracking->GetTPCTransform()->convDeltaTimeToDeltaZinTimeFrame(trks[revLabel].CSide() * GPUChainTracking::NSECTORS / 2, trks[revLabel].GetParam().GetTOffset());
1069 comp = fabsf(trks[i].GetParam().GetZ() + shift1) < fabsf(trks[revLabel].GetParam().GetZ() + shift2);
1070 }
1071 if (revLabel == -1 || !trks[revLabel].OK() || (trks[i].OK() && comp)) {
1072 revLabel = i;
1073 }
1074 }
1075 }
1076 }
1077 }
1078 if ((mQATasks & taskClusterAttach) && !tracksExternal) {
1079 std::vector<uint8_t> lowestPadRow(mTracking->mIOPtrs.nMergedTracks);
1080 // fill cluster adjacent status
1081 if (mTracking->mIOPtrs.mergedTrackHitAttachment) {
1082 for (uint32_t i = 0; i < GetNMCLabels(); i++) {
1083 if (mClusterParam[i].attached == 0 && mClusterParam[i].fakeAttached == 0) {
1084 int32_t attach = mTracking->mIOPtrs.mergedTrackHitAttachment[i];
1085 if (attach & gputpcgmmergertypes::attachFlagMask) {
1086 int32_t track = attach & gputpcgmmergertypes::attachTrackMask;
1087 mcLabelI_t trackL = mTrackMCLabels[track];
1088 bool fake = true;
1089 for (int32_t j = 0; j < GetMCLabelNID(i); j++) {
1090 // GPUInfo("Attach %x Track %d / %d:%d", attach, track, j, GetMCLabelID(i, j));
1091 if (trackL == GetMCLabel(i, j)) {
1092 fake = false;
1093 break;
1094 }
1095 }
1096 if (fake) {
1097 mClusterParam[i].fakeAdjacent++;
1098 } else {
1099 mClusterParam[i].adjacent++;
1100 }
1101 }
1102 }
1103 }
1104 }
1105 if (mTracking->mIOPtrs.nMergedTracks && clNative) {
1106 std::fill(lowestPadRow.begin(), lowestPadRow.end(), 255);
1107 for (uint32_t iSector = 0; iSector < GPUTPCGeometry::NSECTORS; iSector++) {
1108 for (uint32_t iRow = 0; iRow < GPUTPCGeometry::NROWS; iRow++) {
1109 for (uint32_t iCl = 0; iCl < clNative->nClusters[iSector][iRow]; iCl++) {
1110 int32_t i = clNative->clusterOffset[iSector][iRow] + iCl;
1111 for (int32_t j = 0; j < GetMCLabelNID(i); j++) {
1112 uint32_t trackId = GetMCTrackObj(mTrackMCLabelsReverse, GetMCLabel(i, j));
1113 if (trackId < lowestPadRow.size() && lowestPadRow[trackId] > iRow) {
1114 lowestPadRow[trackId] = iRow;
1115 }
1116 }
1117 }
1118 }
1119 }
1120 for (uint32_t i = 0; i < mTracking->mIOPtrs.nMergedTracks; i++) {
1121 const auto& trk = mTracking->mIOPtrs.mergedTracks[i];
1122 if (trk.OK() && lowestPadRow[i] != 255 && trk.NClustersFitted() >= PADROW_CHECK_MINCLS && CAMath::Abs(trk.GetParam().GetQPt()) < 1.0) {
1123 const auto& lowestCl = mTracking->mIOPtrs.mergedTrackHits[trk.FirstClusterRef()].row < mTracking->mIOPtrs.mergedTrackHits[trk.FirstClusterRef() + trk.NClusters() - 1].row ? mTracking->mIOPtrs.mergedTrackHits[trk.FirstClusterRef()] : mTracking->mIOPtrs.mergedTrackHits[trk.FirstClusterRef() + trk.NClusters() - 1];
1124 const int32_t lowestRow = lowestCl.row;
1125 mPadRow[0]->Fill(lowestPadRow[i], lowestRow, 1.f);
1126 mPadRow[1]->Fill(CAMath::ATan2(trk.GetParam().GetY(), trk.GetParam().GetX()), lowestRow, 1.f);
1127 if (lowestPadRow[i] < 10 && lowestRow > lowestPadRow[i] + 3) {
1128 const auto& cl = clNative->clustersLinear[lowestCl.num];
1129 float x, y, z;
1130 mTracking->GetTPCTransform()->Transform(lowestCl.sector, lowestCl.row, cl.getPad(), cl.getTime(), x, y, z, trk.GetParam().GetTOffset());
1131 float phi = CAMath::ATan2(y, x);
1132 mPadRow[2]->Fill(phi, lowestRow, 1.f);
1133 if (CAMath::Abs(phi) < 0.15) {
1134 const float time = cl.getTime();
1135 mPadRow[3]->Fill(mTracking->GetParam().GetUnscaledMult(time), lowestRow, 1.f);
1136 }
1137 }
1138 }
1139 }
1140 }
1141 }
1142
1143 if (mConfig.matchMCLabels.size()) {
1144 mGoodHits[mNEvents - 1].resize(GetNMCLabels());
1145 std::vector<bool> allowMCLabels(GetNMCTracks(0));
1146 for (uint32_t k = 0; k < GetNMCTracks(0); k++) {
1147 allowMCLabels[k] = false;
1148 }
1149 for (uint32_t i = 0; i < nReconstructedTracks; i++) {
1150 if (!mGoodTracks[mNEvents - 1][i]) {
1151 continue;
1152 }
1153 if (mConfig.matchDisplayMinPt > 0) {
1154 if (!mTrackMCLabels[i].isValid()) {
1155 continue;
1156 }
1157 const mcInfo_t& info = GetMCTrack(mTrackMCLabels[i]);
1158 if (info.pX * info.pX + info.pY * info.pY < mConfig.matchDisplayMinPt * mConfig.matchDisplayMinPt) {
1159 continue;
1160 }
1161 }
1162
1163 const GPUTPCGMMergedTrack& track = mTracking->mIOPtrs.mergedTracks[i];
1164 for (uint32_t j = 0; j < track.NClusters(); j++) {
1165 int32_t hitId = mTracking->mIOPtrs.mergedTrackHits[track.FirstClusterRef() + j].num;
1166 if (GetMCLabelNID(hitId)) {
1167 int32_t mcID = GetMCLabelID(hitId, 0);
1168 if (mcID >= 0) {
1169 allowMCLabels[mcID] = true;
1170 }
1171 }
1172 }
1173 }
1174 for (uint32_t i = 0; i < GetNMCLabels(); i++) {
1175 for (int32_t j = 0; j < GetMCLabelNID(i); j++) {
1176 int32_t mcID = GetMCLabelID(i, j);
1177 if (mcID >= 0 && allowMCLabels[mcID]) {
1178 mGoodHits[mNEvents - 1][i] = true;
1179 }
1180 }
1181 }
1182 }
1183 if (timer.IsRunning()) {
1184 GPUInfo("QA Time: Cluster attach status:\t\t%6.0f us", timer.GetCurrentElapsedTime(true) * 1e6);
1185 }
1186
1187 if (matchOnly) {
1188 return;
1189 }
1190
1191 // Recompute fNWeightCls (might have changed after merging events into timeframes)
1192 for (uint32_t iCol = 0; iCol < GetNMCCollissions(); iCol++) {
1193 for (uint32_t i = 0; i < GetNMCTracks(iCol); i++) {
1194 mMCParam[iCol][i].nWeightCls = 0.;
1195 }
1196 }
1197 for (uint32_t i = 0; i < GetNMCLabels(); i++) {
1198 float weightTotal = 0.f;
1199 for (int32_t j = 0; j < GetMCLabelNID(i); j++) {
1200 if (GetMCLabelID(i, j) >= 0) {
1201 weightTotal += GetMCLabelWeight(i, j);
1202 }
1203 }
1204 for (int32_t j = 0; j < GetMCLabelNID(i); j++) {
1205 if (GetMCLabelID(i, j) >= 0) {
1206 GetMCTrackObj(mMCParam, GetMCLabel(i, j)).nWeightCls += GetMCLabelWeight(i, j) / weightTotal;
1207 }
1208 }
1209 }
1210 if (timer.IsRunning()) {
1211 GPUInfo("QA Time: Compute cluster label weights:\t%6.0f us", timer.GetCurrentElapsedTime(true) * 1e6);
1212 }
1213
1214 // Compute MC Track Parameters for MC Tracks
1215 tbb::parallel_for<uint32_t>(0, GetNMCCollissions(), [&](auto iCol) {
1216 for (uint32_t i = 0; i < GetNMCTracks(iCol); i++) {
1217 const mcInfo_t& info = GetMCTrack(i, iCol);
1218 additionalMCParameters& mc2 = mMCParam[iCol][i];
1219 mc2.pt = std::sqrt(info.pX * info.pX + info.pY * info.pY);
1220 mc2.phi = M_PI + std::atan2(-info.pY, -info.pX);
1221 float p = info.pX * info.pX + info.pY * info.pY + info.pZ * info.pZ;
1222 if (p < 1e-18) {
1223 mc2.theta = mc2.eta = 0.f;
1224 } else {
1225 mc2.theta = info.pZ == 0 ? (M_PI / 2) : (std::acos(info.pZ / std::sqrt(p)));
1226 mc2.eta = -std::log(std::tan(0.5 * mc2.theta));
1227 }
1228 if (mConfig.writeMCLabels) {
1229 std::vector<int32_t>& effBuffer = mcEffBuffer[mNEvents - 1];
1230 effBuffer[i] = mRecTracks[iCol][i] * 1000 + mFakeTracks[iCol][i];
1231 }
1232 } // clang-format off
1233 }, tbb::simple_partitioner()); // clang-format on
1234 if (timer.IsRunning()) {
1235 GPUInfo("QA Time: Compute track mc parameters:\t%6.0f us", timer.GetCurrentElapsedTime(true) * 1e6);
1236 }
1237
1238 // Fill Efficiency Histograms
1239 if (mQATasks & taskTrackingEff) {
1240 for (uint32_t iCol = 0; iCol < GetNMCCollissions(); iCol++) {
1241 for (uint32_t i = 0; i < GetNMCTracks(iCol); i++) {
1242 if ((mMCTrackMin != -1 && (int32_t)i < mMCTrackMin) || (mMCTrackMax != -1 && (int32_t)i >= mMCTrackMax)) {
1243 continue;
1244 }
1245 const mcInfo_t& info = GetMCTrack(i, iCol);
1246 const additionalMCParameters& mc2 = mMCParam[iCol][i];
1247 if (mc2.nWeightCls == 0.f) {
1248 continue;
1249 }
1250 const float& mcpt = mc2.pt;
1251 const float& mcphi = mc2.phi;
1252 const float& mceta = mc2.eta;
1253
1254 if (info.primDaughters) {
1255 continue;
1256 }
1257 if (mc2.nWeightCls < mConfig.minNClEff) {
1258 continue;
1259 }
1260 int32_t findable = mc2.nWeightCls >= mConfig.minNClFindable;
1261 if (info.pid < 0) {
1262 continue;
1263 }
1264 if (info.charge == 0.f) {
1265 continue;
1266 }
1267 if (mConfig.filterCharge && info.charge * mConfig.filterCharge < 0) {
1268 continue;
1269 }
1270 if (mConfig.filterPID >= 0 && info.pid != mConfig.filterPID) {
1271 continue;
1272 }
1273
1274 if (fabsf(mceta) > ETA_MAX || mcpt < PT_MIN || mcpt > PT_MAX) {
1275 continue;
1276 }
1277
1278 float alpha = std::atan2(info.y, info.x);
1279 alpha /= M_PI / 9.f;
1280 alpha = std::floor(alpha);
1281 alpha *= M_PI / 9.f;
1282 alpha += M_PI / 18.f;
1283
1284 float c = std::cos(alpha);
1285 float s = std::sin(alpha);
1286 float localY = -info.x * s + info.y * c;
1287
1288 if (mConfig.dumpToROOTLevel >= 1) {
1289 static auto effdump = GPUROOTDump<TNtuple>::getNew("eff", "alpha:x:y:z:mcphi:mceta:mcpt:rec:fake:findable:prim:ncls");
1290 float localX = info.x * c + info.y * s;
1291 effdump.Fill(alpha, localX, localY, info.z, mcphi, mceta, mcpt, mRecTracks[iCol][i], mFakeTracks[iCol][i], findable, info.prim, mc2.nWeightCls);
1292 }
1293
1294 for (int32_t j = 0; j < 6; j++) {
1295 if (j == 3 || j == 4) {
1296 continue;
1297 }
1298 for (int32_t k = 0; k < 2; k++) {
1299 if (k == 0 && findable == 0) {
1300 continue;
1301 }
1302
1303 int32_t val = (j == 0) ? (mRecTracks[iCol][i] ? 1 : 0) : (j == 1) ? (mRecTracks[iCol][i] ? mRecTracks[iCol][i] - 1 : 0) : (j == 2) ? mFakeTracks[iCol][i] : 1;
1304 if (val == 0) {
1305 continue;
1306 }
1307
1308 for (int32_t l = 0; l < 5; l++) {
1309 if (info.prim && mcpt < PT_MIN_PRIM) {
1310 continue;
1311 }
1312 if (l != 3 && fabsf(mceta) > ETA_MAX2) {
1313 continue;
1314 }
1315 if (l < 4 && mcpt < 1.f / mConfig.qpt) {
1316 continue;
1317 }
1318
1319 float pos = l == 0 ? localY : l == 1 ? info.z : l == 2 ? mcphi : l == 3 ? mceta : mcpt;
1320
1321 mEff[j][k][!info.prim][l]->Fill(pos, val);
1322 }
1323 }
1324 }
1325 }
1326 }
1327 if (timer.IsRunning()) {
1328 GPUInfo("QA Time: Fill efficiency histograms:\t%6.0f us", timer.GetCurrentElapsedTime(true) * 1e6);
1329 }
1330 }
1331
1332 // Fill Resolution Histograms
1333 if (mQATasks & (taskTrackingRes | taskTrackingResPull)) {
1334 GPUTPCGMPropagator prop;
1335 prop.SetMaxSinPhi(.999);
1336 prop.SetMaterialTPC();
1337 prop.SetPolynomialField(&mParam->polynomialField);
1338
1339 for (uint32_t i = 0; i < mTrackMCLabels.size(); i++) {
1340 if (mConfig.writeMCLabels) {
1341 std::vector<int32_t>& labelBuffer = mcLabelBuffer[mNEvents - 1];
1342 labelBuffer[i] = mTrackMCLabels[i].getTrackID();
1343 }
1344 if (mTrackMCLabels[i].isFake()) {
1345 continue;
1346 }
1347 const mcInfo_t& mc1 = GetMCTrack(mTrackMCLabels[i]);
1348 const additionalMCParameters& mc2 = GetMCTrackObj(mMCParam, mTrackMCLabels[i]);
1349
1350 if (mc1.primDaughters) {
1351 continue;
1352 }
1353 if (!tracksExternal) {
1354 if (!mTracking->mIOPtrs.mergedTracks[i].OK()) {
1355 continue;
1356 }
1357 if (mTracking->mIOPtrs.mergedTracks[i].MergedLooper()) {
1358 continue;
1359 }
1360 }
1361 if ((mMCTrackMin != -1 && mTrackMCLabels[i].getTrackID() < mMCTrackMin) || (mMCTrackMax != -1 && mTrackMCLabels[i].getTrackID() >= mMCTrackMax)) {
1362 continue;
1363 }
1364 if (fabsf(mc2.eta) > ETA_MAX || mc2.pt < PT_MIN || mc2.pt > PT_MAX) {
1365 continue;
1366 }
1367 if (mc1.charge == 0.f) {
1368 continue;
1369 }
1370 if (mc1.pid < 0) {
1371 continue;
1372 }
1373#ifndef GPUCA_RUN2
1374 if (mc1.t0 == -100.f) {
1375 continue;
1376 }
1377#endif
1378 if (mConfig.filterCharge && mc1.charge * mConfig.filterCharge < 0) {
1379 continue;
1380 }
1381 if (mConfig.filterPID >= 0 && mc1.pid != mConfig.filterPID) {
1382 continue;
1383 }
1384 if (mc2.nWeightCls < mConfig.minNClRes) {
1385 continue;
1386 }
1387 if (mConfig.resPrimaries == 1 && !mc1.prim) {
1388 continue;
1389 } else if (mConfig.resPrimaries == 2 && mc1.prim) {
1390 continue;
1391 }
1392 if (GetMCTrackObj(mTrackMCLabelsReverse, mTrackMCLabels[i]) != (int32_t)i) {
1393 continue;
1394 }
1395
1396 GPUTPCGMTrackParam param;
1397 float alpha = 0.f;
1398 int32_t side;
1399 if (tracksExternal) {
1400#ifndef GPUCA_STANDALONE
1401 for (int32_t k = 0; k < 5; k++) {
1402 param.Par()[k] = (*tracksExternal)[i].getParams()[k];
1403 }
1404 for (int32_t k = 0; k < 15; k++) {
1405 param.Cov()[k] = (*tracksExternal)[i].getCov()[k];
1406 }
1407 param.X() = (*tracksExternal)[i].getX();
1408 param.TOffset() = (*tracksExternal)[i].getTime0();
1409 alpha = (*tracksExternal)[i].getAlpha();
1410 side = (*tracksExternal)[i].hasBothSidesClusters() ? 2 : ((*tracksExternal)[i].hasCSideClusters() ? 1 : 0);
1411#endif
1412 } else {
1413 param = mTracking->mIOPtrs.mergedTracks[i].GetParam();
1414 alpha = mTracking->mIOPtrs.mergedTracks[i].GetAlpha();
1415 side = mTracking->mIOPtrs.mergedTracks[i].CCE() ? 2 : (mTracking->mIOPtrs.mergedTracks[i].CSide() ? 1 : 0);
1416 }
1417
1418 float mclocal[4]; // Rotated x,y,Px,Py mc-coordinates - the MC data should be rotated since the track is propagated best along x
1419 float c = std::cos(alpha);
1420 float s = std::sin(alpha);
1421 float x = mc1.x;
1422 float y = mc1.y;
1423 mclocal[0] = x * c + y * s;
1424 mclocal[1] = -x * s + y * c;
1425 float px = mc1.pX;
1426 float py = mc1.pY;
1427 mclocal[2] = px * c + py * s;
1428 mclocal[3] = -px * s + py * c;
1429
1430 if (mclocal[0] < TRACK_EXPECTED_REFERENCE_X - 3) {
1431 continue;
1432 }
1433 if (mclocal[0] > param.GetX() + 20) {
1434 continue;
1435 }
1436 if (param.GetX() > mConfig.maxResX) {
1437 continue;
1438 }
1439
1440 auto getdz = [this, &param, &mc1, &side, tracksExternal]() {
1441 if (tracksExternal) {
1442 return param.GetZ();
1443 }
1444 if (!mParam->continuousMaxTimeBin) {
1445 return param.GetZ() - mc1.z;
1446 }
1447 float shift = 0;
1448#ifndef GPUCA_RUN2
1449 if (side != 2) {
1450 shift = mTracking->GetTPCTransform()->convDeltaTimeToDeltaZinTimeFrame(side * GPUChainTracking::NSECTORS / 2, param.GetTOffset() - mc1.t0);
1451 }
1452#endif
1453 return param.GetZ() + shift - mc1.z;
1454 };
1455
1456 prop.SetTrack(&param, alpha);
1457 bool inFlyDirection = 0;
1458 if (mConfig.strict) {
1459 const float dx = param.X() - std::max<float>(mclocal[0], TRACK_EXPECTED_REFERENCE_X_DEFAULT); // Limit distance check
1460 const float dy = param.Y() - mclocal[1];
1461 const float dz = getdz();
1462 if (dx * dx + dy * dy + dz * dz > 5.f * 5.f) {
1463 continue;
1464 }
1465 }
1466
1467 if (prop.PropagateToXAlpha(mclocal[0], alpha, inFlyDirection)) {
1468 continue;
1469 }
1470 if (fabsf(param.Y() - mclocal[1]) > (mConfig.strict ? 1.f : 4.f) || fabsf(getdz()) > (mConfig.strict ? 1.f : 4.f)) {
1471 continue;
1472 }
1473 float charge = mc1.charge > 0 ? 1.f : -1.f;
1474
1475 float deltaY = param.GetY() - mclocal[1];
1476 float deltaZ = getdz();
1477 float deltaPhiNative = param.GetSinPhi() - mclocal[3] / mc2.pt;
1478 float deltaPhi = std::asin(param.GetSinPhi()) - std::atan2(mclocal[3], mclocal[2]);
1479 float deltaLambdaNative = param.GetDzDs() - mc1.pZ / mc2.pt;
1480 float deltaLambda = std::atan(param.GetDzDs()) - std::atan2(mc1.pZ, mc2.pt);
1481 float deltaPtNative = (param.GetQPt() - charge / mc2.pt) * charge;
1482 float deltaPt = (fabsf(1.f / param.GetQPt()) - mc2.pt) / mc2.pt;
1483
1484 float paramval[5] = {mclocal[1], mc1.z, mc2.phi, mc2.eta, mc2.pt};
1485 float resval[5] = {deltaY, deltaZ, mConfig.nativeFitResolutions ? deltaPhiNative : deltaPhi, mConfig.nativeFitResolutions ? deltaLambdaNative : deltaLambda, mConfig.nativeFitResolutions ? deltaPtNative : deltaPt};
1486 float pullval[5] = {deltaY / std::sqrt(param.GetErr2Y()), deltaZ / std::sqrt(param.GetErr2Z()), deltaPhiNative / std::sqrt(param.GetErr2SinPhi()), deltaLambdaNative / std::sqrt(param.GetErr2DzDs()), deltaPtNative / std::sqrt(param.GetErr2QPt())};
1487
1488 for (int32_t j = 0; j < 5; j++) {
1489 for (int32_t k = 0; k < 5; k++) {
1490 if (k != 3 && fabsf(mc2.eta) > ETA_MAX2) {
1491 continue;
1492 }
1493 if (k < 4 && mc2.pt < 1.f / mConfig.qpt) {
1494 continue;
1495 }
1496 if (mQATasks & taskTrackingRes) {
1497 mRes2[j][k]->Fill(resval[j], paramval[k]);
1498 }
1499 if (mQATasks & taskTrackingResPull) {
1500 mPull2[j][k]->Fill(pullval[j], paramval[k]);
1501 }
1502 }
1503 }
1504 }
1505 if (timer.IsRunning()) {
1506 GPUInfo("QA Time: Fill resolution histograms:\t%6.0f us", timer.GetCurrentElapsedTime(true) * 1e6);
1507 }
1508 }
1509
1510 if ((mQATasks & taskClusterAttach) && !tracksExternal) {
1511 // Fill cluster histograms
1512 for (uint32_t iTrk = 0; iTrk < nReconstructedTracks; iTrk++) {
1513 const GPUTPCGMMergedTrack& track = mTracking->mIOPtrs.mergedTracks[iTrk];
1514 if (!track.OK()) {
1515 continue;
1516 }
1517 if (!mTrackMCLabels[iTrk].isValid()) {
1518 for (uint32_t k = 0; k < track.NClusters(); k++) {
1519 if (mTracking->mIOPtrs.mergedTrackHits[track.FirstClusterRef() + k].state & GPUTPCGMMergedTrackHit::flagReject) {
1520 continue;
1521 }
1522 int32_t hitId = mTracking->mIOPtrs.mergedTrackHits[track.FirstClusterRef() + k].num;
1523 float totalWeight = 0.;
1524 for (int32_t j = 0; j < GetMCLabelNID(hitId); j++) {
1525 if (GetMCLabelID(hitId, j) >= 0 && GetMCTrackObj(mMCParam, GetMCLabel(hitId, j)).pt > 1.f / mTracking->GetParam().rec.maxTrackQPtB5) {
1526 totalWeight += GetMCLabelWeight(hitId, j);
1527 }
1528 }
1529 int32_t attach = mTracking->mIOPtrs.mergedTrackHitAttachment[hitId];
1530 const auto& r = checkClusterState<false>(attach);
1531 if (totalWeight > 0) {
1532 float weight = 1.f / (totalWeight * (mClusterParam[hitId].attached + mClusterParam[hitId].fakeAttached));
1533 for (int32_t j = 0; j < GetMCLabelNID(hitId); j++) {
1534 mcLabelI_t label = GetMCLabel(hitId, j);
1535 if (!label.isFake() && GetMCTrackObj(mMCParam, label).pt > 1.f / mTracking->GetParam().rec.maxTrackQPtB5) {
1536 float pt = GetMCTrackObj(mMCParam, label).pt;
1537 if (pt < PT_MIN_CLUST) {
1538 pt = PT_MIN_CLUST;
1539 }
1540 mClusters[CL_fake]->Fill(pt, GetMCLabelWeight(hitId, j) * weight);
1541 mClusters[CL_att_adj]->Fill(pt, GetMCLabelWeight(hitId, j) * weight);
1542 if (GetMCTrackObj(mRecTracks, label)) {
1543 mClusters[CL_tracks]->Fill(pt, GetMCLabelWeight(hitId, j) * weight);
1544 }
1545 mClusters[CL_all]->Fill(pt, GetMCLabelWeight(hitId, j) * weight);
1546 if (r.protect || r.physics) {
1547 mClusters[CL_prot]->Fill(pt, GetMCLabelWeight(hitId, j) * weight);
1548 }
1549 if (r.physics) {
1550 mClusters[CL_physics]->Fill(pt, GetMCLabelWeight(hitId, j) * weight);
1551 }
1552 }
1553 }
1554 } else {
1555 float weight = 1.f / (mClusterParam[hitId].attached + mClusterParam[hitId].fakeAttached);
1556 mClusters[CL_fake]->Fill(0.f, weight);
1557 mClusters[CL_att_adj]->Fill(0.f, weight);
1558 mClusters[CL_all]->Fill(0.f, weight);
1559 mClusterCounts.nUnaccessible += weight;
1560 if (r.protect || r.physics) {
1561 mClusters[CL_prot]->Fill(0.f, weight);
1562 }
1563 if (r.physics) {
1564 mClusters[CL_physics]->Fill(0.f, weight);
1565 }
1566 }
1567 }
1568 continue;
1569 }
1570 mcLabelI_t label = mTrackMCLabels[iTrk];
1571 if (mMCTrackMin != -1 && (label.getTrackID() < mMCTrackMin || label.getTrackID() >= mMCTrackMax)) {
1572 continue;
1573 }
1574 for (uint32_t k = 0; k < track.NClusters(); k++) {
1575 if (mTracking->mIOPtrs.mergedTrackHits[track.FirstClusterRef() + k].state & GPUTPCGMMergedTrackHit::flagReject) {
1576 continue;
1577 }
1578 int32_t hitId = mTracking->mIOPtrs.mergedTrackHits[track.FirstClusterRef() + k].num;
1579 float pt = GetMCTrackObj(mMCParam, label).pt;
1580 if (pt < PT_MIN_CLUST) {
1581 pt = PT_MIN_CLUST;
1582 }
1583 float weight = 1.f / (mClusterParam[hitId].attached + mClusterParam[hitId].fakeAttached);
1584 bool correct = false;
1585 for (int32_t j = 0; j < GetMCLabelNID(hitId); j++) {
1586 if (label == GetMCLabel(hitId, j)) {
1587 correct = true;
1588 break;
1589 }
1590 }
1591 if (correct) {
1592 mClusters[CL_attached]->Fill(pt, weight);
1593 mClusters[CL_tracks]->Fill(pt, weight);
1594 } else {
1595 mClusters[CL_fake]->Fill(pt, weight);
1596 }
1597 mClusters[CL_att_adj]->Fill(pt, weight);
1598 mClusters[CL_all]->Fill(pt, weight);
1599 int32_t attach = mTracking->mIOPtrs.mergedTrackHitAttachment[hitId];
1600 const auto& r = checkClusterState<false>(attach);
1601 if (r.protect || r.physics) {
1602 mClusters[CL_prot]->Fill(pt, weight);
1603 }
1604 if (r.physics) {
1605 mClusters[CL_physics]->Fill(pt, weight);
1606 }
1607 }
1608 }
1609 for (uint32_t i = 0; i < GetNMCLabels(); i++) {
1610 if ((mMCTrackMin != -1 && GetMCLabelID(i, 0) < mMCTrackMin) || (mMCTrackMax != -1 && GetMCLabelID(i, 0) >= mMCTrackMax)) {
1611 continue;
1612 }
1613 if (mClusterParam[i].attached || mClusterParam[i].fakeAttached) {
1614 continue;
1615 }
1616 int32_t attach = mTracking->mIOPtrs.mergedTrackHitAttachment[i];
1617 const auto& r = checkClusterState<false>(attach);
1618 if (mClusterParam[i].adjacent) {
1619 int32_t label = mTracking->mIOPtrs.mergedTrackHitAttachment[i] & gputpcgmmergertypes::attachTrackMask;
1620 if (!mTrackMCLabels[label].isValid()) {
1621 float totalWeight = 0.;
1622 for (int32_t j = 0; j < GetMCLabelNID(i); j++) {
1623 mcLabelI_t labelT = GetMCLabel(i, j);
1624 if (!labelT.isFake() && GetMCTrackObj(mMCParam, labelT).pt > 1.f / mTracking->GetParam().rec.maxTrackQPtB5) {
1625 totalWeight += GetMCLabelWeight(i, j);
1626 }
1627 }
1628 float weight = 1.f / totalWeight;
1629 if (totalWeight > 0) {
1630 for (int32_t j = 0; j < GetMCLabelNID(i); j++) {
1631 mcLabelI_t labelT = GetMCLabel(i, j);
1632 if (!labelT.isFake() && GetMCTrackObj(mMCParam, labelT).pt > 1.f / mTracking->GetParam().rec.maxTrackQPtB5) {
1633 float pt = GetMCTrackObj(mMCParam, labelT).pt;
1634 if (pt < PT_MIN_CLUST) {
1635 pt = PT_MIN_CLUST;
1636 }
1637 if (GetMCTrackObj(mRecTracks, labelT)) {
1638 mClusters[CL_tracks]->Fill(pt, GetMCLabelWeight(i, j) * weight);
1639 }
1640 mClusters[CL_att_adj]->Fill(pt, GetMCLabelWeight(i, j) * weight);
1641 mClusters[CL_fakeAdj]->Fill(pt, GetMCLabelWeight(i, j) * weight);
1642 mClusters[CL_all]->Fill(pt, GetMCLabelWeight(i, j) * weight);
1643 if (r.protect || r.physics) {
1644 mClusters[CL_prot]->Fill(pt, GetMCLabelWeight(i, j) * weight);
1645 }
1646 if (r.physics) {
1647 mClusters[CL_physics]->Fill(pt, GetMCLabelWeight(i, j) * weight);
1648 }
1649 }
1650 }
1651 } else {
1652 mClusters[CL_att_adj]->Fill(0.f, 1.f);
1653 mClusters[CL_fakeAdj]->Fill(0.f, 1.f);
1654 mClusters[CL_all]->Fill(0.f, 1.f);
1655 mClusterCounts.nUnaccessible++;
1656 if (r.protect || r.physics) {
1657 mClusters[CL_prot]->Fill(0.f, 1.f);
1658 }
1659 if (r.physics) {
1660 mClusters[CL_physics]->Fill(0.f, 1.f);
1661 }
1662 }
1663 } else {
1664 float pt = GetMCTrackObj(mMCParam, mTrackMCLabels[label]).pt;
1665 if (pt < PT_MIN_CLUST) {
1666 pt = PT_MIN_CLUST;
1667 }
1668 mClusters[CL_att_adj]->Fill(pt, 1.f);
1669 mClusters[CL_tracks]->Fill(pt, 1.f);
1670 mClusters[CL_all]->Fill(pt, 1.f);
1671 if (r.protect || r.physics) {
1672 mClusters[CL_prot]->Fill(pt, 1.f);
1673 }
1674 if (r.physics) {
1675 mClusters[CL_physics]->Fill(pt, 1.f);
1676 }
1677 }
1678 } else {
1679 float totalWeight = 0.;
1680 for (int32_t j = 0; j < GetMCLabelNID(i); j++) {
1681 mcLabelI_t labelT = GetMCLabel(i, j);
1682 if (!labelT.isFake() && GetMCTrackObj(mMCParam, labelT).pt > 1.f / mTracking->GetParam().rec.maxTrackQPtB5) {
1683 totalWeight += GetMCLabelWeight(i, j);
1684 }
1685 }
1686 if (totalWeight > 0) {
1687 for (int32_t j = 0; j < GetMCLabelNID(i); j++) {
1688 mcLabelI_t label = GetMCLabel(i, j);
1689 if (!label.isFake() && GetMCTrackObj(mMCParam, label).pt > 1.f / mTracking->GetParam().rec.maxTrackQPtB5) {
1690 float pt = GetMCTrackObj(mMCParam, label).pt;
1691 if (pt < PT_MIN_CLUST) {
1692 pt = PT_MIN_CLUST;
1693 }
1694 float weight = GetMCLabelWeight(i, j) / totalWeight;
1695 if (mClusterParam[i].fakeAdjacent) {
1696 mClusters[CL_fakeAdj]->Fill(pt, weight);
1697 }
1698 if (mClusterParam[i].fakeAdjacent) {
1699 mClusters[CL_att_adj]->Fill(pt, weight);
1700 }
1701 if (GetMCTrackObj(mRecTracks, label)) {
1702 mClusters[CL_tracks]->Fill(pt, weight);
1703 }
1704 mClusters[CL_all]->Fill(pt, weight);
1705 if (r.protect || r.physics) {
1706 mClusters[CL_prot]->Fill(pt, weight);
1707 }
1708 if (r.physics) {
1709 mClusters[CL_physics]->Fill(pt, weight);
1710 }
1711 }
1712 }
1713 } else {
1714 if (mClusterParam[i].fakeAdjacent) {
1715 mClusters[CL_fakeAdj]->Fill(0.f, 1.f);
1716 }
1717 if (mClusterParam[i].fakeAdjacent) {
1718 mClusters[CL_att_adj]->Fill(0.f, 1.f);
1719 }
1720 mClusters[CL_all]->Fill(0.f, 1.f);
1721 mClusterCounts.nUnaccessible++;
1722 if (r.protect || r.physics) {
1723 mClusters[CL_prot]->Fill(0.f, 1.f);
1724 }
1725 if (r.physics) {
1726 mClusters[CL_physics]->Fill(0.f, 1.f);
1727 }
1728 }
1729 }
1730 }
1731
1732 if (timer.IsRunning()) {
1733 GPUInfo("QA Time: Fill cluster histograms:\t%6.0f us", timer.GetCurrentElapsedTime(true) * 1e6);
1734 }
1735 }
1736 } else if (!mConfig.inputHistogramsOnly && !mConfig.noMC && (mQATasks & (taskTrackingEff | taskTrackingRes | taskTrackingResPull | taskClusterAttach))) {
1737 GPUWarning("No MC information available, only running partial TPC QA!");
1738 } // mcAvail
1739
1740 if ((mQATasks & taskTrackStatistics) && !tracksExternal) {
1741 // Fill track statistic histograms
1742 std::vector<std::array<float, 3>> clusterAttachCounts;
1743 if (mcAvail) {
1744 clusterAttachCounts.resize(GetNMCLabels(), {0.f, 0.f});
1745 }
1746 for (uint32_t i = 0; i < nReconstructedTracks; i++) {
1747 const GPUTPCGMMergedTrack& track = mTracking->mIOPtrs.mergedTracks[i];
1748 if (!track.OK()) {
1749 continue;
1750 }
1751 mTrackPt->Fill(1.f / fabsf(track.GetParam().GetQPt()));
1752 if (mParam->par.dodEdx && mParam->dodEdxEnabled && track.NClusters() >= 60) {
1753 const GPUdEdxInfo& trackdEdx = mTracking->GetProcessors()->tpcMerger.MergedTracksdEdx()[i];
1754 const float logp = logf(1.f / fabsf(track.GetParam().GetQPt()) * sqrtf(1.f + track.GetParam().GetDzDs() * track.GetParam().GetDzDs()));
1755 mTrackdEdx[0]->Fill(logp, trackdEdx.dEdxTotTPC);
1756 mTrackdEdx[1]->Fill(logp, trackdEdx.dEdxMaxTPC);
1757 }
1758 mNCl[0]->Fill(track.NClustersFitted());
1759 uint32_t nClCorrected = 0;
1760 const auto& trackClusters = mTracking->mIOPtrs.mergedTrackHits;
1761 uint32_t jNext = 0;
1762 for (uint32_t j = 0; j < track.NClusters(); j = jNext) {
1763 uint32_t rowClCount = !(trackClusters[track.FirstClusterRef() + j].state & GPUTPCGMMergedTrackHit::flagReject);
1764 for (jNext = j + 1; j < track.NClusters(); jNext++) {
1765 if (trackClusters[track.FirstClusterRef() + j].sector != trackClusters[track.FirstClusterRef() + jNext].sector || trackClusters[track.FirstClusterRef() + j].row != trackClusters[track.FirstClusterRef() + jNext].row) {
1766 break;
1767 }
1768 rowClCount += !(trackClusters[track.FirstClusterRef() + jNext].state & GPUTPCGMMergedTrackHit::flagReject);
1769 }
1770 if (!track.MergedLooper() && rowClCount) {
1771 nClCorrected++;
1772 }
1773 if (mcAvail && rowClCount) {
1774 for (uint32_t k = j; k < jNext; k++) {
1775 const auto& cl = trackClusters[track.FirstClusterRef() + k];
1776 if (cl.state & GPUTPCGMMergedTrackHit::flagReject) {
1777 continue;
1778 }
1779 bool labelOk = false, labelOkNonFake = false;
1780 const mcLabelI_t& trkLabel = mTrackMCLabels[i];
1781 if (trkLabel.isValid() && !trkLabel.isNoise()) {
1782 for (int32_t l = 0; l < GetMCLabelNID(cl.num); l++) {
1783 const mcLabelI_t& clLabel = GetMCLabel(cl.num, l);
1784 if (clLabel.isValid() && !clLabel.isNoise() && CompareIgnoreFake(trkLabel, clLabel)) {
1785 labelOk = true;
1786 if (!trkLabel.isFake()) {
1787 labelOkNonFake = true;
1788 }
1789 break;
1790 }
1791 }
1792 }
1793 clusterAttachCounts[cl.num][0] += 1.0f;
1794 clusterAttachCounts[cl.num][1] += (float)labelOk / rowClCount;
1795 clusterAttachCounts[cl.num][2] += (float)labelOkNonFake / rowClCount;
1796 }
1797 }
1798 }
1799 if (nClCorrected) {
1800 mNCl[1]->Fill(nClCorrected);
1801 }
1802 mT0[0]->Fill(track.GetParam().GetTOffset());
1803#ifndef GPUCA_RUN2
1804 if (mTrackMCLabels.size() && !mTrackMCLabels[i].isFake() && !track.MergedLooper() && !track.CCE()) {
1805 const auto& info = GetMCTrack(mTrackMCLabels[i]);
1806 if (info.t0 != -100.f) {
1807 mT0[1]->Fill(track.GetParam().GetTOffset() - info.t0);
1808 }
1809 }
1810#endif
1811 }
1812 if (mClNative && mTracking && mTracking->GetTPCTransform()) {
1813 for (uint32_t i = 0; i < GPUChainTracking::NSECTORS; i++) {
1814 for (uint32_t j = 0; j < GPUTPCGeometry::NROWS; j++) {
1815 for (uint32_t k = 0; k < mClNative->nClusters[i][j]; k++) {
1816 const auto& cl = mClNative->clusters[i][j][k];
1817 float x, y, z;
1818 GPUTPCConvertImpl::convert(*mTracking->GetTPCTransform(), mTracking->GetParam(), i, j, cl.getPad(), cl.getTime(), x, y, z);
1819 mTracking->GetParam().Sector2Global(i, x, y, z, &x, &y, &z);
1820 mClXY->Fill(x, y);
1821 }
1822 }
1823 }
1824 }
1825 if (mcAvail) {
1826 double clusterAttachNormalizedCount = 0, clusterAttachNormalizedCountNonFake = 0;
1827 for (uint32_t i = 0; i < clusterAttachCounts.size(); i++) {
1828 if (clusterAttachCounts[i][0]) {
1829 clusterAttachNormalizedCount += clusterAttachCounts[i][1] / clusterAttachCounts[i][0];
1830 clusterAttachNormalizedCountNonFake += clusterAttachCounts[i][2] / clusterAttachCounts[i][0];
1831 }
1832 }
1833 mClusterCounts.nCorrectlyAttachedNormalized = clusterAttachNormalizedCount;
1834 mClusterCounts.nCorrectlyAttachedNormalizedNonFake = clusterAttachNormalizedCountNonFake;
1835 clusterAttachCounts.clear();
1836 }
1837
1838 if (timer.IsRunning()) {
1839 GPUInfo("QA Time: Fill track statistics:\t%6.0f us", timer.GetCurrentElapsedTime(true) * 1e6);
1840 }
1841 }
1842
1843 uint32_t nCl = clNative ? clNative->nClustersTotal : mTracking->GetProcessors()->tpcMerger.NMaxClusters();
1844 mClusterCounts.nTotal += nCl;
1845 if (mQATasks & (taskClusterCounts | taskClusterRejection)) {
1846 for (uint32_t iSector = 0; iSector < GPUTPCGeometry::NSECTORS; iSector++) {
1847 for (uint32_t iRow = 0; iRow < GPUTPCGeometry::NROWS; iRow++) {
1848 for (uint32_t iCl = 0; iCl < clNative->nClusters[iSector][iRow]; iCl++) {
1849 uint32_t i = clNative->clusterOffset[iSector][iRow] + iCl;
1850 int32_t attach = mTracking->mIOPtrs.mergedTrackHitAttachment[i];
1851 const auto& r = checkClusterState<true>(attach, &mClusterCounts);
1852
1853 if (mQATasks & taskClusterRejection) {
1854 if (mcAvail) {
1855 float totalWeight = 0, weight400 = 0, weight40 = 0;
1856 for (int32_t j = 0; j < GetMCLabelNID(i); j++) {
1857 const auto& label = GetMCLabel(i, j);
1858 if (GetMCLabelID(label) >= 0) {
1859 totalWeight += GetMCLabelWeight(label);
1860 if (GetMCTrackObj(mMCParam, label).pt >= 0.4) {
1861 weight400 += GetMCLabelWeight(label);
1862 }
1863 if (GetMCTrackObj(mMCParam, label).pt <= 0.04) {
1864 weight40 += GetMCLabelWeight(label);
1865 }
1866 }
1867 }
1868 if (totalWeight > 0 && 10.f * weight400 >= totalWeight) {
1869 if (!r.unattached && !r.protect && !r.physics) {
1870 mClusterCounts.nFakeRemove400++;
1871 int32_t totalFake = weight400 < 0.9f * totalWeight;
1872 if (totalFake) {
1873 mClusterCounts.nFullFakeRemove400++;
1874 }
1875 /*printf("Fake removal (%d): Hit %7d, attached %d lowPt %d looper %d tube200 %d highIncl %d tube %d bad %d recPt %7.2f recLabel %6d", totalFake, i, (int32_t) (mClusterParam[i].attached || mClusterParam[i].fakeAttached),
1876 (int32_t) lowPt, (int32_t) ((attach & gputpcgmmergertypes::attachGoodLeg) == 0), (int32_t) ((attach & gputpcgmmergertypes::attachTube) && mev200),
1877 (int32_t) ((attach & gputpcgmmergertypes::attachHighIncl) != 0), (int32_t) ((attach & gputpcgmmergertypes::attachTube) != 0), (int32_t) ((attach & gputpcgmmergertypes::attachGood) == 0),
1878 fabsf(qpt) > 0 ? 1.f / qpt : 0.f, id);
1879 for (int32_t j = 0;j < GetMCLabelNID(i);j++)
1880 {
1881 //if (GetMCLabelID(i, j) < 0) break;
1882 printf(" - label%d %6d weight %5d", j, GetMCLabelID(i, j), (int32_t) GetMCLabelWeight(i, j));
1883 if (GetMCLabelID(i, j) >= 0) printf(" - pt %7.2f", mMCParam[GetMCLabelID(i, j)].pt);
1884 else printf(" ");
1885 }
1886 printf("\n");*/
1887 }
1888 mClusterCounts.nAbove400++;
1889 }
1890 if (totalWeight > 0 && weight40 >= 0.9 * totalWeight) {
1891 mClusterCounts.nBelow40++;
1892 if (r.protect || r.physics) {
1893 mClusterCounts.nFakeProtect40++;
1894 }
1895 }
1896 }
1897
1898 if (r.physics) {
1899 mClusterCounts.nPhysics++;
1900 }
1901 if (r.protect) {
1902 mClusterCounts.nProt++;
1903 }
1904 if (r.unattached) {
1905 mClusterCounts.nUnattached++;
1906 }
1907 }
1908 if (mQATasks & taskClusterRejection) {
1909 if (mTracking && clNative) {
1910 const auto& cl = clNative->clustersLinear[i];
1911 mClRej[0]->Fill(cl.getPad() - GPUTPCGeometry::NPads(iRow) / 2 + 0.5, iRow, 1.f);
1912 if (!r.unattached && !r.protect) {
1913 mClRej[1]->Fill(cl.getPad() - GPUTPCGeometry::NPads(iRow) / 2 + 0.5, iRow, 1.f);
1914 }
1915 }
1916 }
1917 }
1918 }
1919 }
1920 }
1921
1922 // Process cluster count statistics
1923 if ((mQATasks & taskClusterCounts) && mConfig.clusterRejectionHistograms) {
1924 DoClusterCounts(nullptr);
1925 mClusterCounts = counts_t();
1926 }
1927
1928 if (timer.IsRunning()) {
1929 GPUInfo("QA Time: Cluster Counts:\t%6.0f us", timer.GetCurrentElapsedTime(true) * 1e6);
1930 }
1931
1932 if (mConfig.dumpToROOTLevel >= 1 && !tracksExternal) {
1933 if (!clNative || !mTracking || !mTracking->mIOPtrs.mergedTrackHitAttachment || !mTracking->mIOPtrs.mergedTracks) {
1934 throw std::runtime_error("Cannot dump non o2::tpc::clusterNative clusters, need also hit attachmend and GPU tracks");
1935 }
1936 uint32_t clid = 0;
1937 for (uint32_t i = 0; i < GPUChainTracking::NSECTORS; i++) {
1938 for (uint32_t j = 0; j < GPUTPCGeometry::NROWS; j++) {
1939 for (uint32_t k = 0; k < mClNative->nClusters[i][j]; k++) {
1940 const auto& cl = mClNative->clusters[i][j][k];
1941 uint32_t attach = mTracking->mIOPtrs.mergedTrackHitAttachment[clid];
1942 float x = 0, y = 0, z = 0;
1943 if (attach & gputpcgmmergertypes::attachFlagMask) {
1944 uint32_t track = attach & gputpcgmmergertypes::attachTrackMask;
1945 const auto& trk = mTracking->mIOPtrs.mergedTracks[track];
1946 mTracking->GetTPCTransform()->Transform(i, j, cl.getPad(), cl.getTime(), x, y, z, trk.GetParam().GetTOffset());
1947 mTracking->GetParam().Sector2Global(i, x, y, z, &x, &y, &z);
1948 }
1949 uint32_t extState = mTracking->mIOPtrs.mergedTrackHitStates ? mTracking->mIOPtrs.mergedTrackHitStates[clid] : 0;
1950
1951 if (mConfig.dumpToROOTLevel >= 2) {
1952 GPUTPCGMMergedTrack trk;
1953 GPUTPCGMMergedTrackHit trkHit;
1954 memset((void*)&trk, 0, sizeof(trk));
1955 memset((void*)&trkHit, 0, sizeof(trkHit));
1956 if (attach & gputpcgmmergertypes::attachFlagMask) {
1957 uint32_t track = attach & gputpcgmmergertypes::attachTrackMask;
1958 trk = mTracking->mIOPtrs.mergedTracks[track];
1959 for (uint32_t l = 0; l < trk.NClusters(); l++) {
1960 const auto& tmp = mTracking->mIOPtrs.mergedTrackHits[trk.FirstClusterRef() + l];
1961 if (tmp.num == clid) {
1962 trkHit = tmp;
1963 break;
1964 }
1965 }
1966 }
1967 static auto cldump = GPUROOTDump<o2::tpc::ClusterNative, GPUTPCGMMergedTrack, GPUTPCGMMergedTrackHit, uint32_t, uint32_t, float, float, float, uint32_t, uint32_t, uint32_t>::getNew("cluster", "track", "trackHit", "attach", "extState", "x", "y", "z", "sector", "row", "nEv", "clusterTree");
1968 cldump.Fill(cl, trk, trkHit, attach, extState, x, y, z, i, j, mNEvents - 1);
1969 } else {
1970 static auto cldump = GPUROOTDump<o2::tpc::ClusterNative, uint32_t, uint32_t, float, float, float, uint32_t, uint32_t, uint32_t>::getNew("cluster", "attach", "extState", "x", "y", "z", "sector", "row", "nEv", "clusterTree");
1971 cldump.Fill(cl, attach, extState, x, y, z, i, j, mNEvents - 1);
1972 }
1973 clid++;
1974 }
1975 }
1976 }
1977
1978 static auto trkdump = GPUROOTDump<uint32_t, GPUTPCGMMergedTrack>::getNew("nEv", "track", "tracksTree");
1979 for (uint32_t i = 0; i < mTracking->mIOPtrs.nMergedTracks; i++) {
1980 if (mTracking->mIOPtrs.mergedTracks[i].OK()) {
1981 trkdump.Fill(mNEvents - 1, mTracking->mIOPtrs.mergedTracks[i]);
1982 }
1983 }
1984
1985 if (mTracking && mTracking->GetProcessingSettings().createO2Output) {
1986 static auto o2trkdump = GPUROOTDump<uint32_t, o2::tpc::TrackTPC>::getNew("nEv", "track", "tracksO2Tree");
1987 for (uint32_t i = 0; i < mTracking->mIOPtrs.nOutputTracksTPCO2; i++) {
1988 o2trkdump.Fill(mNEvents - 1, mTracking->mIOPtrs.outputTracksTPCO2[i]);
1989 }
1990 }
1991 }
1992
1993 if (mConfig.compareTrackStatus) {
1994#ifdef GPUCA_DETERMINISTIC_MODE
1995 if (!mTracking || !mTracking->GetProcessingSettings().deterministicGPUReconstruction)
1996#endif
1997 {
1998 throw std::runtime_error("Need deterministic processing to compare track status");
1999 }
2000 std::vector<uint8_t> status(mTracking->mIOPtrs.nMergedTracks);
2001 for (uint32_t i = 0; i < mTracking->mIOPtrs.nMergedTracks; i++) {
2002 const auto& trk = mTracking->mIOPtrs.mergedTracks[i];
2003 status[i] = trk.OK() && trk.NClusters() && trk.GetParam().GetNDF() > 0 && (mConfig.noMC || (mTrackMCLabels[i].isValid() && !mTrackMCLabels[i].isFake()));
2004 }
2005 if (mConfig.compareTrackStatus == 1) {
2006 std::ofstream("track.status", std::ios::binary).write((char*)status.data(), status.size() * sizeof(status[0]));
2007 } else if (mConfig.compareTrackStatus == 2) {
2008 std::ifstream f("track.status", std::ios::binary | std::ios::ate);
2009 std::vector<uint8_t> comp(f.tellg());
2010 f.seekg(0);
2011 f.read((char*)comp.data(), comp.size());
2012
2013 if (comp.size() != status.size()) {
2014 throw std::runtime_error("Number of tracks candidates in track fit in track.status and in current reconstruction differ");
2015 }
2016 std::vector<uint32_t> missing, missingComp;
2017 for (uint32_t i = 0; i < status.size(); i++) {
2018 if (status[i] && !comp[i]) {
2019 missingComp.emplace_back(i);
2020 }
2021 if (comp[i] && !status[i]) {
2022 missing.emplace_back(i);
2023 }
2024 }
2025 auto printer = [](std::vector<uint32_t> m, const char* name) {
2026 if (m.size()) {
2027 printf("Missing in %s reconstruction: (%zu)\n", name, m.size());
2028 for (uint32_t i = 0; i < m.size(); i++) {
2029 if (i) {
2030 printf(", ");
2031 }
2032 printf("%d", m[i]);
2033 }
2034 printf("\n");
2035 }
2036 };
2037 printer(missing, "current");
2038 printer(missingComp, "comparison");
2039 }
2040 }
2041
2042 mTrackingScratchBuffer.clear();
2043 mTrackingScratchBuffer.shrink_to_fit();
2044}
2045
2046void GPUQA::GetName(char* fname, int32_t k, bool noDash)
2047{
2048 const int32_t nNewInput = mConfig.inputHistogramsOnly ? 0 : 1;
2049 if (k || mConfig.inputHistogramsOnly || mConfig.name.size()) {
2050 if (!(mConfig.inputHistogramsOnly || k)) {
2051 snprintf(fname, 1024, "%s%s", mConfig.name.c_str(), noDash ? "" : " - ");
2052 } else if (mConfig.compareInputNames.size() > (unsigned)(k - nNewInput)) {
2053 snprintf(fname, 1024, "%s%s", mConfig.compareInputNames[k - nNewInput].c_str(), noDash ? "" : " - ");
2054 } else {
2055 strcpy(fname, mConfig.compareInputs[k - nNewInput].c_str());
2056 if (strlen(fname) > 5 && strcmp(fname + strlen(fname) - 5, ".root") == 0) {
2057 fname[strlen(fname) - 5] = 0;
2058 }
2059 if (!noDash) {
2060 strcat(fname, " - ");
2061 }
2062 }
2063 } else {
2064 fname[0] = 0;
2065 }
2066}
2067
2068template <class T>
2069T* GPUQA::GetHist(T*& ee, std::vector<std::unique_ptr<TFile>>& tin, int32_t k, int32_t nNewInput)
2070{
2071 T* e = ee;
2072 if ((mConfig.inputHistogramsOnly || k) && (e = dynamic_cast<T*>(tin[k - nNewInput]->Get(e->GetName()))) == nullptr) {
2073 GPUWarning("Missing histogram in input %s: %s", mConfig.compareInputs[k - nNewInput].c_str(), ee->GetName());
2074 return (nullptr);
2075 }
2076 ee = e;
2077 return (e);
2078}
2079
2080void GPUQA::DrawQAHistogramsCleanup()
2081{
2082 clearGarbagageCollector();
2083}
2084
2085void GPUQA::resetHists()
2086{
2087 if (!mQAInitialized) {
2088 throw std::runtime_error("QA not initialized");
2089 }
2090 if (mHaveExternalHists) {
2091 throw std::runtime_error("Cannot reset external hists");
2092 }
2093 for (auto& h : *mHist1D) {
2094 h.Reset();
2095 }
2096 for (auto& h : *mHist2D) {
2097 h.Reset();
2098 }
2099 for (auto& h : *mHist1Dd) {
2100 h.Reset();
2101 }
2102 for (auto& h : *mHistGraph) {
2103 h = TGraphAsymmErrors();
2104 }
2105 mClusterCounts = counts_t();
2106}
2107
2108int32_t GPUQA::DrawQAHistograms(TObjArray* qcout)
2109{
2110 const auto oldRootIgnoreLevel = gErrorIgnoreLevel;
2111 gErrorIgnoreLevel = kWarning;
2112 if (!mQAInitialized) {
2113 throw std::runtime_error("QA not initialized");
2114 }
2115
2116 if (mTracking && mTracking->GetProcessingSettings().debugLevel >= 2) {
2117 printf("Creating QA Histograms\n");
2118 }
2119
2120 std::vector<Color_t> colorNums(COLORCOUNT);
2121 if (!(qcout || mConfig.writeFileExt == "root" || mConfig.writeFileExt == "C")) {
2122 [[maybe_unused]] static int32_t initColorsInitialized = initColors();
2123 }
2124 for (int32_t i = 0; i < COLORCOUNT; i++) {
2125 colorNums[i] = (qcout || mConfig.writeFileExt == "root" || mConfig.writeFileExt == "C") ? defaultColorNums[i] : mColors[i]->GetNumber();
2126 }
2127
2128 bool mcAvail = mcPresent();
2129 char name[2048], fname[1024];
2130
2131 const int32_t nNewInput = mConfig.inputHistogramsOnly ? 0 : 1;
2132 const int32_t ConfigNumInputs = nNewInput + mConfig.compareInputs.size();
2133
2134 std::vector<std::unique_ptr<TFile>> tin;
2135 for (uint32_t i = 0; i < mConfig.compareInputs.size(); i++) {
2136 tin.emplace_back(std::make_unique<TFile>(mConfig.compareInputs[i].c_str()));
2137 }
2138 std::unique_ptr<TFile> tout = nullptr;
2139 if (mConfig.output.size()) {
2140 tout = std::make_unique<TFile>(mConfig.output.c_str(), "RECREATE");
2141 }
2142
2143 if (mConfig.enableLocalOutput || mConfig.shipToQCAsCanvas) {
2144 float legendSpacingString = 0.025;
2145 for (int32_t i = 0; i < ConfigNumInputs; i++) {
2146 GetName(fname, i);
2147 if (strlen(fname) * 0.006 > legendSpacingString) {
2148 legendSpacingString = strlen(fname) * 0.006;
2149 }
2150 }
2151
2152 // Create Canvas / Pads for Efficiency Histograms
2153 if (mQATasks & taskTrackingEff) {
2154 for (int32_t ii = 0; ii < 6; ii++) {
2155 snprintf(name, 1024, "eff_vs_%s_layout", VSPARAMETER_NAMES[ii]);
2156 mCEff[ii] = createGarbageCollected<TCanvas>(name, name, 0, 0, 700, 700. * 2. / 3.);
2157 mCEff[ii]->cd();
2158 float dy = 1. / 2.;
2159 mPEff[ii][0] = createGarbageCollected<TPad>("p0", "", 0.0, dy * 0, 0.5, dy * 1);
2160 mPEff[ii][0]->Draw();
2161 mPEff[ii][0]->SetRightMargin(0.04);
2162 mPEff[ii][1] = createGarbageCollected<TPad>("p1", "", 0.5, dy * 0, 1.0, dy * 1);
2163 mPEff[ii][1]->Draw();
2164 mPEff[ii][1]->SetRightMargin(0.04);
2165 mPEff[ii][2] = createGarbageCollected<TPad>("p2", "", 0.0, dy * 1, 0.5, dy * 2 - .001);
2166 mPEff[ii][2]->Draw();
2167 mPEff[ii][2]->SetRightMargin(0.04);
2168 mPEff[ii][3] = createGarbageCollected<TPad>("p3", "", 0.5, dy * 1, 1.0, dy * 2 - .001);
2169 mPEff[ii][3]->Draw();
2170 mPEff[ii][3]->SetRightMargin(0.04);
2171 mLEff[ii] = createGarbageCollected<TLegend>(0.92 - legendSpacingString * 1.45, 0.83 - (0.93 - 0.82) / 2. * (float)ConfigNumInputs, 0.98, 0.849);
2172 SetLegend(mLEff[ii]);
2173 }
2174 }
2175
2176 // Create Canvas / Pads for Resolution Histograms
2177 if (mQATasks & taskTrackingRes) {
2178 for (int32_t ii = 0; ii < 7; ii++) {
2179 if (ii == 6) {
2180 snprintf(name, 1024, "res_integral_layout");
2181 } else {
2182 snprintf(name, 1024, "res_vs_%s_layout", VSPARAMETER_NAMES[ii]);
2183 }
2184 mCRes[ii] = createGarbageCollected<TCanvas>(name, name, 0, 0, 700, 700. * 2. / 3.);
2185 mCRes[ii]->cd();
2186 gStyle->SetOptFit(1);
2187
2188 float dy = 1. / 2.;
2189 mPRes[ii][3] = createGarbageCollected<TPad>("p0", "", 0.0, dy * 0, 0.5, dy * 1);
2190 mPRes[ii][3]->Draw();
2191 mPRes[ii][3]->SetRightMargin(0.04);
2192 mPRes[ii][4] = createGarbageCollected<TPad>("p1", "", 0.5, dy * 0, 1.0, dy * 1);
2193 mPRes[ii][4]->Draw();
2194 mPRes[ii][4]->SetRightMargin(0.04);
2195 mPRes[ii][0] = createGarbageCollected<TPad>("p2", "", 0.0, dy * 1, 1. / 3., dy * 2 - .001);
2196 mPRes[ii][0]->Draw();
2197 mPRes[ii][0]->SetRightMargin(0.04);
2198 mPRes[ii][0]->SetLeftMargin(0.15);
2199 mPRes[ii][1] = createGarbageCollected<TPad>("p3", "", 1. / 3., dy * 1, 2. / 3., dy * 2 - .001);
2200 mPRes[ii][1]->Draw();
2201 mPRes[ii][1]->SetRightMargin(0.04);
2202 mPRes[ii][1]->SetLeftMargin(0.135);
2203 mPRes[ii][2] = createGarbageCollected<TPad>("p4", "", 2. / 3., dy * 1, 1.0, dy * 2 - .001);
2204 mPRes[ii][2]->Draw();
2205 mPRes[ii][2]->SetRightMargin(0.06);
2206 mPRes[ii][2]->SetLeftMargin(0.135);
2207 if (ii < 6) {
2208 mLRes[ii] = createGarbageCollected<TLegend>(0.9 - legendSpacingString * 1.45, 0.93 - (0.93 - 0.86) / 2. * (float)ConfigNumInputs, 0.98, 0.949);
2209 SetLegend(mLRes[ii]);
2210 }
2211 }
2212 }
2213
2214 // Create Canvas / Pads for Pull Histograms
2215 if (mQATasks & taskTrackingResPull) {
2216 for (int32_t ii = 0; ii < 7; ii++) {
2217 if (ii == 6) {
2218 snprintf(name, 1024, "pull_integral_layout");
2219 } else {
2220 snprintf(name, 1024, "pull_vs_%s_layout", VSPARAMETER_NAMES[ii]);
2221 }
2222 mCPull[ii] = createGarbageCollected<TCanvas>(name, name, 0, 0, 700, 700. * 2. / 3.);
2223 mCPull[ii]->cd();
2224 gStyle->SetOptFit(1);
2225
2226 float dy = 1. / 2.;
2227 mPPull[ii][3] = createGarbageCollected<TPad>("p0", "", 0.0, dy * 0, 0.5, dy * 1);
2228 mPPull[ii][3]->Draw();
2229 mPPull[ii][3]->SetRightMargin(0.04);
2230 mPPull[ii][4] = createGarbageCollected<TPad>("p1", "", 0.5, dy * 0, 1.0, dy * 1);
2231 mPPull[ii][4]->Draw();
2232 mPPull[ii][4]->SetRightMargin(0.04);
2233 mPPull[ii][0] = createGarbageCollected<TPad>("p2", "", 0.0, dy * 1, 1. / 3., dy * 2 - .001);
2234 mPPull[ii][0]->Draw();
2235 mPPull[ii][0]->SetRightMargin(0.04);
2236 mPPull[ii][0]->SetLeftMargin(0.15);
2237 mPPull[ii][1] = createGarbageCollected<TPad>("p3", "", 1. / 3., dy * 1, 2. / 3., dy * 2 - .001);
2238 mPPull[ii][1]->Draw();
2239 mPPull[ii][1]->SetRightMargin(0.04);
2240 mPPull[ii][1]->SetLeftMargin(0.135);
2241 mPPull[ii][2] = createGarbageCollected<TPad>("p4", "", 2. / 3., dy * 1, 1.0, dy * 2 - .001);
2242 mPPull[ii][2]->Draw();
2243 mPPull[ii][2]->SetRightMargin(0.06);
2244 mPPull[ii][2]->SetLeftMargin(0.135);
2245 if (ii < 6) {
2246 mLPull[ii] = createGarbageCollected<TLegend>(0.9 - legendSpacingString * 1.45, 0.93 - (0.93 - 0.86) / 2. * (float)ConfigNumInputs, 0.98, 0.949);
2247 SetLegend(mLPull[ii]);
2248 }
2249 }
2250 }
2251
2252 // Create Canvas for Cluster Histos
2253 if (mQATasks & taskClusterAttach) {
2254 for (int32_t i = 0; i < 3; i++) {
2255 snprintf(name, 1024, "clusters_%s_layout", CLUSTER_TYPES[i]);
2256 mCClust[i] = createGarbageCollected<TCanvas>(name, name, 0, 0, 700, 700. * 2. / 3.);
2257 mCClust[i]->cd();
2258 mPClust[i] = createGarbageCollected<TPad>("p0", "", 0.0, 0.0, 1.0, 1.0);
2259 mPClust[i]->Draw();
2260 float y1 = i != 1 ? 0.77 : 0.27, y2 = i != 1 ? 0.9 : 0.42;
2261 mLClust[i] = createGarbageCollected<TLegend>(i == 2 ? 0.1 : (0.65 - legendSpacingString * 1.45), y2 - (y2 - y1) * (ConfigNumInputs + (i != 1) / 2.) + 0.005, i == 2 ? (0.3 + legendSpacingString * 1.45) : 0.9, y2);
2262 SetLegend(mLClust[i]);
2263 }
2264 }
2265
2266 // Create Canvas for track statistic histos
2267 if (mQATasks & taskTrackStatistics) {
2268 for (int32_t i = 0; i < 3; i++) {
2269 snprintf(name, 2048, "ctracks%s", i ? (i == 2 ? "dedxmax" : "dedxtot") : "pt");
2270 mCTracks[i] = createGarbageCollected<TCanvas>(name, name, 0, 0, 700, 700. * 2. / 3.);
2271 mCTracks[i]->cd();
2272 mPTracks[i] = createGarbageCollected<TPad>("p0", "", 0.0, 0.0, 1.0, 1.0);
2273 mPTracks[i]->Draw();
2274 mLTracks[i] = createGarbageCollected<TLegend>(0.9 - legendSpacingString * 1.5, 0.93 - (0.93 - 0.86) / 2. * (float)ConfigNumInputs, 0.98, 0.949);
2275 SetLegend(mLTracks[i], true);
2276 }
2277
2278 for (int32_t i = 0; i < 2; i++) {
2279 snprintf(name, 2048, "ctrackst0%d", i);
2280 mCT0[i] = createGarbageCollected<TCanvas>(name, name, 0, 0, 700, 700. * 2. / 3.);
2281 mCT0[i]->cd();
2282 mPT0[i] = createGarbageCollected<TPad>("p0", "", 0.0, 0.0, 1.0, 1.0);
2283 mPT0[i]->Draw();
2284 mLT0[i] = createGarbageCollected<TLegend>(0.9 - legendSpacingString * 1.45, 0.93 - (0.93 - 0.86) / 2. * (float)ConfigNumInputs, 0.98, 0.949);
2285 SetLegend(mLT0[i]);
2286
2287 snprintf(name, 2048, "cncl%d", i);
2288 mCNCl[i] = createGarbageCollected<TCanvas>(name, name, 0, 0, 700, 700. * 2. / 3.);
2289 mCNCl[i]->cd();
2290 mPNCl[i] = createGarbageCollected<TPad>("p0", "", 0.0, 0.0, 1.0, 1.0);
2291 mPNCl[i]->Draw();
2292 mLNCl[i] = createGarbageCollected<TLegend>(0.9 - legendSpacingString * 1.45, 0.93 - (0.93 - 0.86) / 2. * (float)ConfigNumInputs, 0.98, 0.949); // TODO: Fix sizing of legend, and also fix font size
2293 SetLegend(mLNCl[i], true);
2294 }
2295
2296 mCClXY = createGarbageCollected<TCanvas>("clxy", "clxy", 0, 0, 700, 700. * 2. / 3.);
2297 mCClXY->cd();
2298 mPClXY = createGarbageCollected<TPad>("p0", "", 0.0, 0.0, 1.0, 1.0);
2299 mPClXY->Draw();
2300 }
2301
2302 if (mQATasks & taskClusterRejection) {
2303 for (int32_t i = 0; i < 3; i++) {
2304 snprintf(name, 2048, "cnclrej%d", i);
2305 mCClRej[i] = createGarbageCollected<TCanvas>(name, name, 0, 0, 700, 700. * 2. / 3.);
2306 mCClRej[i]->cd();
2307 mPClRej[i] = createGarbageCollected<TPad>("p0", "", 0.0, 0.0, 1.0, 1.0);
2308 mPClRej[i]->Draw();
2309 }
2310 mCClRejP = createGarbageCollected<TCanvas>("cnclrejp", "cnclrejp", 0, 0, 700, 700. * 2. / 3.);
2311 mCClRejP->cd();
2312 mPClRejP = createGarbageCollected<TPad>("p0", "", 0.0, 0.0, 1.0, 1.0);
2313 mPClRejP->Draw();
2314 }
2315
2316 if (mQATasks & taskClusterAttach) {
2317 for (int32_t i = 0; i < 4; i++) {
2318 snprintf(name, 2048, "cpadrow%d", i);
2319 mCPadRow[i] = createGarbageCollected<TCanvas>(name, name, 0, 0, 700, 700. * 2. / 3.);
2320 mCPadRow[i]->cd();
2321 mPPadRow[i] = createGarbageCollected<TPad>("p0", "", 0.0, 0.0, 1.0, 1.0);
2322 mPPadRow[i]->Draw();
2323 }
2324 }
2325 }
2326
2327 if (mConfig.enableLocalOutput && !mConfig.inputHistogramsOnly && (mQATasks & taskTrackingEff) && mcPresent()) {
2328 GPUInfo("QA Stats: Eff: Tracks Prim %d (Eta %d, Pt %d) %f%% (%f%%) Sec %d (Eta %d, Pt %d) %f%% (%f%%) - Res: Tracks %d (Eta %d, Pt %d)", (int32_t)mEff[3][1][0][0]->GetEntries(), (int32_t)mEff[3][1][0][3]->GetEntries(), (int32_t)mEff[3][1][0][4]->GetEntries(),
2329 mEff[0][0][0][0]->GetSumOfWeights() / std::max(1., mEff[3][0][0][0]->GetSumOfWeights()), mEff[0][1][0][0]->GetSumOfWeights() / std::max(1., mEff[3][1][0][0]->GetSumOfWeights()), (int32_t)mEff[3][1][1][0]->GetEntries(), (int32_t)mEff[3][1][1][3]->GetEntries(),
2330 (int32_t)mEff[3][1][1][4]->GetEntries(), mEff[0][0][1][0]->GetSumOfWeights() / std::max(1., mEff[3][0][1][0]->GetSumOfWeights()), mEff[0][1][1][0]->GetSumOfWeights() / std::max(1., mEff[3][1][1][0]->GetSumOfWeights()), (int32_t)mRes2[0][0]->GetEntries(),
2331 (int32_t)mRes2[0][3]->GetEntries(), (int32_t)mRes2[0][4]->GetEntries());
2332 }
2333
2334 int32_t flagShowVsPtLog = (mConfig.enableLocalOutput || mConfig.shipToQCAsCanvas) ? 1 : 0;
2335
2336 if (mQATasks & taskTrackingEff) {
2337 // Process / Draw Efficiency Histograms
2338 for (int32_t ii = 0; ii < 5 + flagShowVsPtLog; ii++) {
2339 int32_t i = ii == 5 ? 4 : ii;
2340 for (int32_t k = 0; k < ConfigNumInputs; k++) {
2341 for (int32_t j = 0; j < 4; j++) {
2342 if (mConfig.enableLocalOutput || mConfig.shipToQCAsCanvas) {
2343 mPEff[ii][j]->cd();
2344 if (ii == 5) {
2345 mPEff[ii][j]->SetLogx();
2346 }
2347 }
2348 for (int32_t l = 0; l < 3; l++) {
2349 if (k == 0 && mConfig.inputHistogramsOnly == 0 && ii != 5) {
2350 if (l == 0) {
2351 // Divide eff, compute all for fake/clone
2352 auto oldLevel = gErrorIgnoreLevel;
2353 gErrorIgnoreLevel = kError;
2354 mEffResult[0][j / 2][j % 2][i]->Divide(mEff[l][j / 2][j % 2][i], mEff[5][j / 2][j % 2][i], "cl=0.683 b(1,1) mode");
2355 gErrorIgnoreLevel = oldLevel;
2356 mEff[3][j / 2][j % 2][i]->Reset(); // Sum up rec + clone + fake for fake rate
2357 mEff[3][j / 2][j % 2][i]->Add(mEff[0][j / 2][j % 2][i]);
2358 mEff[3][j / 2][j % 2][i]->Add(mEff[1][j / 2][j % 2][i]);
2359 mEff[3][j / 2][j % 2][i]->Add(mEff[2][j / 2][j % 2][i]);
2360 mEff[4][j / 2][j % 2][i]->Reset(); // Sum up rec + clone for clone rate
2361 mEff[4][j / 2][j % 2][i]->Add(mEff[0][j / 2][j % 2][i]);
2362 mEff[4][j / 2][j % 2][i]->Add(mEff[1][j / 2][j % 2][i]);
2363 } else {
2364 // Divide fake/clone
2365 auto oldLevel = gErrorIgnoreLevel;
2366 gErrorIgnoreLevel = kError;
2367 mEffResult[l][j / 2][j % 2][i]->Divide(mEff[l][j / 2][j % 2][i], mEff[l == 1 ? 4 : 3][j / 2][j % 2][i], "cl=0.683 b(1,1) mode");
2368 gErrorIgnoreLevel = oldLevel;
2369 }
2370 }
2371
2372 TGraphAsymmErrors* e = mEffResult[l][j / 2][j % 2][i];
2373
2374 if (!mConfig.inputHistogramsOnly && k == 0) {
2375 if (tout) {
2376 mEff[l][j / 2][j % 2][i]->Write();
2377 e->Write();
2378 if (l == 2) {
2379 mEff[3][j / 2][j % 2][i]->Write(); // Store also all histogram!
2380 mEff[4][j / 2][j % 2][i]->Write(); // Store also all histogram!
2381 }
2382 }
2383 } else if (GetHist(e, tin, k, nNewInput) == nullptr) {
2384 continue;
2385 }
2386 e->SetTitle(EFFICIENCY_TITLES[j]);
2387 e->GetYaxis()->SetTitle("(Efficiency)");
2388 e->GetXaxis()->SetTitle(XAXIS_TITLES[i]);
2389
2390 e->SetLineWidth(1);
2391 e->SetLineStyle(CONFIG_DASHED_MARKERS ? k + 1 : 1);
2392 SetAxisSize(e);
2393 if (qcout && !mConfig.shipToQCAsCanvas) {
2394 qcout->Add(e);
2395 }
2396 if (!mConfig.enableLocalOutput && !mConfig.shipToQCAsCanvas) {
2397 continue;
2398 }
2399 e->SetMarkerColor(kBlack);
2400 e->SetLineColor(colorNums[(k < 3 ? (l * 3 + k) : (k * 3 + l)) % COLORCOUNT]);
2401 e->GetHistogram()->GetYaxis()->SetRangeUser(-0.02, 1.02);
2402 e->Draw(k || l ? "same P" : "AP");
2403 if (j == 0) {
2404 GetName(fname, k);
2405 mLEff[ii]->AddEntry(e, Form("%s%s", fname, EFF_NAMES[l]), "l");
2406 }
2407 }
2408 if (!mConfig.enableLocalOutput && !mConfig.shipToQCAsCanvas) {
2409 continue;
2410 }
2411 mCEff[ii]->cd();
2412 ChangePadTitleSize(mPEff[ii][j], 0.056);
2413 }
2414 }
2415 if (!mConfig.enableLocalOutput && !mConfig.shipToQCAsCanvas) {
2416 continue;
2417 }
2418
2419 mLEff[ii]->Draw();
2420
2421 if (qcout) {
2422 qcout->Add(mCEff[ii]);
2423 }
2424 if (!mConfig.enableLocalOutput) {
2425 continue;
2426 }
2427 doPerfFigure(0.2, 0.295, 0.025);
2428 mCEff[ii]->Print(Form("%s/eff_vs_%s.pdf", mConfig.plotsDir.c_str(), VSPARAMETER_NAMES[ii]));
2429 if (mConfig.writeFileExt != "") {
2430 mCEff[ii]->Print(Form("%s/eff_vs_%s.%s", mConfig.plotsDir.c_str(), VSPARAMETER_NAMES[ii], mConfig.writeFileExt.c_str()));
2431 }
2432 }
2433 }
2434
2435 if (mQATasks & (taskTrackingRes | taskTrackingResPull)) {
2436 // Process / Draw Resolution Histograms
2437 TH1D *resIntegral[5] = {}, *pullIntegral[5] = {};
2438 TCanvas* cfit = nullptr;
2439 std::unique_ptr<TF1> customGaus = std::make_unique<TF1>("G", "[0]*exp(-(x-[1])*(x-[1])/(2.*[2]*[2]))");
2440 for (int32_t p = 0; p < 2; p++) {
2441 if ((p == 0 && (mQATasks & taskTrackingRes) == 0) || (p == 1 && (mQATasks & taskTrackingResPull) == 0)) {
2442 continue;
2443 }
2444 for (int32_t ii = 0; ii < 5 + flagShowVsPtLog; ii++) {
2445 TCanvas* can = p ? mCPull[ii] : mCRes[ii];
2446 TLegend* leg = p ? mLPull[ii] : mLRes[ii];
2447 int32_t i = ii == 5 ? 4 : ii;
2448 for (int32_t j = 0; j < 5; j++) {
2449 TH2F* src = p ? mPull2[j][i] : mRes2[j][i];
2450 TH1F** dst = p ? mPull[j][i] : mRes[j][i];
2451 TH1D*& dstIntegral = p ? pullIntegral[j] : resIntegral[j];
2452 TPad* pad = p ? mPPull[ii][j] : mPRes[ii][j];
2453
2454 if (!mConfig.inputHistogramsOnly && ii != 5) {
2455 if (cfit == nullptr) {
2456 cfit = createGarbageCollected<TCanvas>();
2457 }
2458 cfit->cd();
2459
2460 TAxis* axis = src->GetYaxis();
2461 int32_t nBins = axis->GetNbins();
2462 int32_t integ = 1;
2463 for (int32_t bin = 1; bin <= nBins; bin++) {
2464 int32_t bin0 = std::max(bin - integ, 0);
2465 int32_t bin1 = std::min(bin + integ, nBins);
2466 std::unique_ptr<TH1D> proj{src->ProjectionX("proj", bin0, bin1)};
2467 proj->ClearUnderflowAndOverflow();
2468 if (proj->GetEntries()) {
2469 uint32_t rebin = 1;
2470 while (proj->GetMaximum() < 50 && rebin < sizeof(RES_AXIS_BINS) / sizeof(RES_AXIS_BINS[0])) {
2471 proj->Rebin(RES_AXIS_BINS[rebin - 1] / RES_AXIS_BINS[rebin]);
2472 rebin++;
2473 }
2474
2475 if (proj->GetEntries() < 20 || proj->GetRMS() < 0.00001) {
2476 dst[0]->SetBinContent(bin, proj->GetRMS());
2477 dst[0]->SetBinError(bin, std::sqrt(proj->GetRMS()));
2478 dst[1]->SetBinContent(bin, proj->GetMean());
2479 dst[1]->SetBinError(bin, std::sqrt(proj->GetRMS()));
2480 } else {
2481 proj->GetXaxis()->SetRange(0, 0);
2482 proj->GetXaxis()->SetRangeUser(std::max(proj->GetXaxis()->GetXmin(), proj->GetMean() - 3. * proj->GetRMS()), std::min(proj->GetXaxis()->GetXmax(), proj->GetMean() + 3. * proj->GetRMS()));
2483 bool forceLogLike = proj->GetMaximum() < 20;
2484 for (int32_t k = forceLogLike ? 2 : 0; k < 3; k++) {
2485 proj->Fit("gaus", forceLogLike || k == 2 ? "sQl" : k ? "sQww" : "sQ");
2486 TF1* fitFunc = proj->GetFunction("gaus");
2487
2488 if (k && !forceLogLike) {
2489 customGaus->SetParameters(fitFunc->GetParameter(0), fitFunc->GetParameter(1), fitFunc->GetParameter(2));
2490 proj->Fit(customGaus.get(), "sQ");
2491 fitFunc = customGaus.get();
2492 }
2493
2494 const float sigma = fabs(fitFunc->GetParameter(2));
2495 dst[0]->SetBinContent(bin, sigma);
2496 dst[1]->SetBinContent(bin, fitFunc->GetParameter(1));
2497 dst[0]->SetBinError(bin, fitFunc->GetParError(2));
2498 dst[1]->SetBinError(bin, fitFunc->GetParError(1));
2499
2500 const bool fail1 = sigma <= 0.f;
2501 const bool fail2 = fabs(proj->GetMean() - dst[1]->GetBinContent(bin)) > std::min<float>(p ? PULL_AXIS : mConfig.nativeFitResolutions ? RES_AXES_NATIVE[j] : RES_AXES[j], 3.f * proj->GetRMS());
2502 const bool fail3 = dst[0]->GetBinContent(bin) > 3.f * proj->GetRMS() || dst[0]->GetBinError(bin) > 1 || dst[1]->GetBinError(bin) > 1;
2503 const bool fail4 = fitFunc->GetParameter(0) < proj->GetMaximum() / 5.;
2504 const bool fail = fail1 || fail2 || fail3 || fail4;
2505 // if (p == 0 && ii == 4 && j == 2) DrawHisto(proj, Form("Hist_bin_%d-%d_vs_%d____%d_%d___%f-%f___%f-%f___%d.pdf", p, j, ii, bin, k, dst[0]->GetBinContent(bin), proj->GetRMS(), dst[1]->GetBinContent(bin), proj->GetMean(), (int32_t) fail), "");
2506
2507 if (!fail) {
2508 break;
2509 } else if (k >= 2) {
2510 dst[0]->SetBinContent(bin, proj->GetRMS());
2511 dst[0]->SetBinError(bin, std::sqrt(proj->GetRMS()));
2512 dst[1]->SetBinContent(bin, proj->GetMean());
2513 dst[1]->SetBinError(bin, std::sqrt(proj->GetRMS()));
2514 }
2515 }
2516 }
2517 } else {
2518 dst[0]->SetBinContent(bin, 0.f);
2519 dst[0]->SetBinError(bin, 0.f);
2520 dst[1]->SetBinContent(bin, 0.f);
2521 dst[1]->SetBinError(bin, 0.f);
2522 }
2523 }
2524 if (ii == 0) {
2525 dstIntegral = src->ProjectionX(mConfig.nativeFitResolutions ? PARAMETER_NAMES_NATIVE[j] : PARAMETER_NAMES[j], 0, nBins + 1);
2526 uint32_t rebin = 1;
2527 while (dstIntegral->GetMaximum() < 50 && rebin < sizeof(RES_AXIS_BINS) / sizeof(RES_AXIS_BINS[0])) {
2528 dstIntegral->Rebin(RES_AXIS_BINS[rebin - 1] / RES_AXIS_BINS[rebin]);
2529 rebin++;
2530 }
2531 }
2532 }
2533 if (ii == 0) {
2534 if (mConfig.inputHistogramsOnly) {
2535 dstIntegral = createGarbageCollected<TH1D>();
2536 }
2537 dstIntegral->SetName(Form(p ? "IntPull%s" : "IntRes%s", VSPARAMETER_NAMES[j]));
2538 dstIntegral->SetTitle(Form(p ? "%s Pull" : "%s Resolution", p || mConfig.nativeFitResolutions ? PARAMETER_NAMES_NATIVE[j] : PARAMETER_NAMES[j]));
2539 }
2540 if (mConfig.enableLocalOutput || mConfig.shipToQCAsCanvas) {
2541 pad->cd();
2542 }
2543 int32_t numColor = 0;
2544 float tmpMax = -1000.;
2545 float tmpMin = 1000.;
2546
2547 for (int32_t l = 0; l < 2; l++) {
2548 for (int32_t k = 0; k < ConfigNumInputs; k++) {
2549 TH1F* e = dst[l];
2550 if (GetHist(e, tin, k, nNewInput) == nullptr) {
2551 continue;
2552 }
2553 if (nNewInput && k == 0 && ii != 5) {
2554 if (p == 0) {
2555 e->Scale(mConfig.nativeFitResolutions ? SCALE_NATIVE[j] : SCALE[j]);
2556 }
2557 }
2558 if (ii == 4) {
2559 e->GetXaxis()->SetRangeUser(0.2, PT_MAX);
2560 } else if (LOG_PT_MIN > 0 && ii == 5) {
2561 e->GetXaxis()->SetRangeUser(LOG_PT_MIN, PT_MAX);
2562 } else if (ii == 5) {
2563 e->GetXaxis()->SetRange(1, 0);
2564 }
2565 e->SetMinimum(-1111);
2566 e->SetMaximum(-1111);
2567
2568 if (e->GetMaximum() > tmpMax) {
2569 tmpMax = e->GetMaximum();
2570 }
2571 if (e->GetMinimum() < tmpMin) {
2572 tmpMin = e->GetMinimum();
2573 }
2574 }
2575 }
2576
2577 float tmpSpan;
2578 tmpSpan = tmpMax - tmpMin;
2579 tmpMax += tmpSpan * .02;
2580 tmpMin -= tmpSpan * .02;
2581 if (j == 2 && i < 3) {
2582 tmpMax += tmpSpan * 0.13 * ConfigNumInputs;
2583 }
2584
2585 for (int32_t k = 0; k < ConfigNumInputs; k++) {
2586 for (int32_t l = 0; l < 2; l++) {
2587 TH1F* e = dst[l];
2588 if (!mConfig.inputHistogramsOnly && k == 0) {
2589 e->SetTitle(Form(p ? "%s Pull" : "%s Resolution", p || mConfig.nativeFitResolutions ? PARAMETER_NAMES_NATIVE[j] : PARAMETER_NAMES[j]));
2590 e->SetStats(kFALSE);
2591 if (tout) {
2592 if (l == 0) {
2593 mRes2[j][i]->SetOption("colz");
2594 mRes2[j][i]->Write();
2595 }
2596 e->Write();
2597 }
2598 } else if (GetHist(e, tin, k, nNewInput) == nullptr) {
2599 continue;
2600 }
2601 e->SetMaximum(tmpMax);
2602 e->SetMinimum(tmpMin);
2603 e->SetLineWidth(1);
2604 e->SetLineStyle(CONFIG_DASHED_MARKERS ? k + 1 : 1);
2605 SetAxisSize(e);
2606 e->GetYaxis()->SetTitle(p ? AXIS_TITLES_PULL[j] : mConfig.nativeFitResolutions ? AXIS_TITLES_NATIVE[j] : AXIS_TITLES[j]);
2607 e->GetXaxis()->SetTitle(XAXIS_TITLES[i]);
2608 if (LOG_PT_MIN > 0 && ii == 5) {
2609 e->GetXaxis()->SetRangeUser(LOG_PT_MIN, PT_MAX);
2610 }
2611
2612 if (j == 0) {
2613 e->GetYaxis()->SetTitleOffset(1.5);
2614 } else if (j < 3) {
2615 e->GetYaxis()->SetTitleOffset(1.4);
2616 }
2617 if (qcout && !mConfig.shipToQCAsCanvas) {
2618 qcout->Add(e);
2619 }
2620 if (!mConfig.enableLocalOutput && !mConfig.shipToQCAsCanvas) {
2621 continue;
2622 }
2623
2624 e->SetMarkerColor(kBlack);
2625 e->SetLineColor(colorNums[numColor++ % COLORCOUNT]);
2626 e->Draw(k || l ? "same" : "");
2627 if (j == 0) {
2628 GetName(fname, k);
2629 leg->AddEntry(e, Form("%s%s", fname, l ? "Mean" : (p ? "Pull" : "Resolution")), "l");
2630 }
2631 }
2632 }
2633 if (!mConfig.enableLocalOutput && !mConfig.shipToQCAsCanvas) {
2634 continue;
2635 }
2636
2637 if (ii == 5) {
2638 pad->SetLogx();
2639 }
2640 can->cd();
2641 if (j == 4) {
2642 ChangePadTitleSize(pad, 0.056);
2643 }
2644 }
2645 if (!mConfig.enableLocalOutput && !mConfig.shipToQCAsCanvas) {
2646 continue;
2647 }
2648
2649 leg->Draw();
2650
2651 if (qcout) {
2652 qcout->Add(can);
2653 }
2654 if (!mConfig.enableLocalOutput) {
2655 continue;
2656 }
2657 doPerfFigure(0.2, 0.295, 0.025);
2658 can->Print(Form(p ? "%s/pull_vs_%s.pdf" : "%s/res_vs_%s.pdf", mConfig.plotsDir.c_str(), VSPARAMETER_NAMES[ii]));
2659 if (mConfig.writeFileExt != "") {
2660 can->Print(Form(p ? "%s/pull_vs_%s.%s" : "%s/res_vs_%s.%s", mConfig.plotsDir.c_str(), VSPARAMETER_NAMES[ii], mConfig.writeFileExt.c_str()));
2661 }
2662 }
2663 }
2664
2665 // Process Integral Resolution Histogreams
2666 for (int32_t p = 0; p < 2; p++) {
2667 if ((p == 0 && (mQATasks & taskTrackingRes) == 0) || (p == 1 && (mQATasks & taskTrackingResPull) == 0)) {
2668 continue;
2669 }
2670 TCanvas* can = p ? mCPull[6] : mCRes[6];
2671 for (int32_t i = 0; i < 5; i++) {
2672 TPad* pad = p ? mPPull[6][i] : mPRes[6][i];
2673 TH1D* hist = p ? pullIntegral[i] : resIntegral[i];
2674 int32_t numColor = 0;
2675 if (mConfig.enableLocalOutput || mConfig.shipToQCAsCanvas) {
2676 pad->cd();
2677 }
2678 if (!mConfig.inputHistogramsOnly && mcAvail) {
2679 TH1D* e = hist;
2680 if (e && e->GetEntries()) {
2681 e->Fit("gaus", "sQ");
2682 }
2683 }
2684
2685 float tmpMax = 0;
2686 for (int32_t k = 0; k < ConfigNumInputs; k++) {
2687 TH1D* e = hist;
2688 if (GetHist(e, tin, k, nNewInput) == nullptr) {
2689 continue;
2690 }
2691 e->SetMaximum(-1111);
2692 if (e->GetMaximum() > tmpMax) {
2693 tmpMax = e->GetMaximum();
2694 }
2695 }
2696
2697 for (int32_t k = 0; k < ConfigNumInputs; k++) {
2698 TH1D* e = hist;
2699 if (GetHist(e, tin, k, nNewInput) == nullptr) {
2700 continue;
2701 }
2702 e->SetMaximum(tmpMax * 1.02);
2703 e->SetMinimum(tmpMax * -0.02);
2704 if (tout && !mConfig.inputHistogramsOnly && k == 0) {
2705 e->Write();
2706 }
2707 if (qcout && !mConfig.shipToQCAsCanvas) {
2708 qcout->Add(e);
2709 }
2710 if (!mConfig.enableLocalOutput && !mConfig.shipToQCAsCanvas) {
2711 continue;
2712 }
2713
2714 e->SetLineColor(colorNums[numColor++ % COLORCOUNT]);
2715 e->Draw(k == 0 ? "" : "same");
2716 }
2717 if (!mConfig.enableLocalOutput && !mConfig.shipToQCAsCanvas) {
2718 continue;
2719 }
2720 can->cd();
2721 }
2722 if (qcout) {
2723 qcout->Add(can);
2724 }
2725 if (!mConfig.enableLocalOutput) {
2726 continue;
2727 }
2728
2729 can->Print(Form(p ? "%s/pull_integral.pdf" : "%s/res_integral.pdf", mConfig.plotsDir.c_str()));
2730 if (mConfig.writeFileExt != "") {
2731 can->Print(Form(p ? "%s/pull_integral.%s" : "%s/res_integral.%s", mConfig.plotsDir.c_str(), mConfig.writeFileExt.c_str()));
2732 }
2733 }
2734 }
2735
2736 uint64_t attachClusterCounts[N_CLS_HIST];
2737 if (mQATasks & taskClusterAttach) {
2738 // Process Cluster Attachment Histograms
2739 if (mConfig.inputHistogramsOnly == 0) {
2740 for (int32_t i = N_CLS_HIST; i < N_CLS_TYPE * N_CLS_HIST - 1; i++) {
2741 mClusters[i]->Sumw2(true);
2742 }
2743 double totalVal = 0;
2744 if (!CLUST_HIST_INT_SUM) {
2745 for (int32_t j = 0; j < mClusters[N_CLS_HIST - 1]->GetXaxis()->GetNbins() + 2; j++) {
2746 totalVal += mClusters[N_CLS_HIST - 1]->GetBinContent(j);
2747 }
2748 }
2749 if (totalVal == 0.) {
2750 totalVal = 1.;
2751 }
2752 for (int32_t i = 0; i < N_CLS_HIST; i++) {
2753 double val = 0;
2754 for (int32_t j = 0; j < mClusters[i]->GetXaxis()->GetNbins() + 2; j++) {
2755 val += mClusters[i]->GetBinContent(j);
2756 mClusters[2 * N_CLS_HIST - 1 + i]->SetBinContent(j, val / totalVal);
2757 }
2758 attachClusterCounts[i] = val;
2759 }
2760
2761 if (!CLUST_HIST_INT_SUM) {
2762 for (int32_t i = 0; i < N_CLS_HIST; i++) {
2763 mClusters[2 * N_CLS_HIST - 1 + i]->SetMaximum(1.02);
2764 mClusters[2 * N_CLS_HIST - 1 + i]->SetMinimum(-0.02);
2765 }
2766 }
2767
2768 for (int32_t i = 0; i < N_CLS_HIST - 1; i++) {
2769 auto oldLevel = gErrorIgnoreLevel;
2770 gErrorIgnoreLevel = kError;
2771 mClusters[N_CLS_HIST + i]->Divide(mClusters[i], mClusters[N_CLS_HIST - 1], 1, 1, "B");
2772 gErrorIgnoreLevel = oldLevel;
2773 mClusters[N_CLS_HIST + i]->SetMinimum(-0.02);
2774 mClusters[N_CLS_HIST + i]->SetMaximum(1.02);
2775 }
2776 }
2777
2778 float tmpMax[2] = {0, 0}, tmpMin[2] = {0, 0};
2779 for (int32_t l = 0; l <= CLUST_HIST_INT_SUM; l++) {
2780 for (int32_t k = 0; k < ConfigNumInputs; k++) {
2781 TH1* e = mClusters[l ? (N_CLS_TYPE * N_CLS_HIST - 2) : (N_CLS_HIST - 1)];
2782 if (GetHist(e, tin, k, nNewInput) == nullptr) {
2783 continue;
2784 }
2785 e->SetMinimum(-1111);
2786 e->SetMaximum(-1111);
2787 if (l == 0) {
2788 e->GetXaxis()->SetRange(2, AXIS_BINS[4]);
2789 }
2790 if (e->GetMaximum() > tmpMax[l]) {
2791 tmpMax[l] = e->GetMaximum();
2792 }
2793 if (e->GetMinimum() < tmpMin[l]) {
2794 tmpMin[l] = e->GetMinimum();
2795 }
2796 }
2797 for (int32_t k = 0; k < ConfigNumInputs; k++) {
2798 for (int32_t i = 0; i < N_CLS_HIST; i++) {
2799 TH1* e = mClusters[l ? (2 * N_CLS_HIST - 1 + i) : i];
2800 if (GetHist(e, tin, k, nNewInput) == nullptr) {
2801 continue;
2802 }
2803 e->SetMaximum(tmpMax[l] * 1.02);
2804 e->SetMinimum(tmpMax[l] * -0.02);
2805 }
2806 }
2807 }
2808
2809 for (int32_t i = 0; i < N_CLS_TYPE; i++) {
2810 if (mConfig.enableLocalOutput || mConfig.shipToQCAsCanvas) {
2811 mPClust[i]->cd();
2812 mPClust[i]->SetLogx();
2813 }
2814 int32_t begin = i == 2 ? (2 * N_CLS_HIST - 1) : i == 1 ? N_CLS_HIST : 0;
2815 int32_t end = i == 2 ? (3 * N_CLS_HIST - 1) : i == 1 ? (2 * N_CLS_HIST - 1) : N_CLS_HIST;
2816 int32_t numColor = 0;
2817 for (int32_t k = 0; k < ConfigNumInputs; k++) {
2818 for (int32_t j = end - 1; j >= begin; j--) {
2819 TH1* e = mClusters[j];
2820 if (GetHist(e, tin, k, nNewInput) == nullptr) {
2821 continue;
2822 }
2823
2824 e->SetTitle(mConfig.plotsNoTitle ? "" : CLUSTER_TITLES[i]);
2825 e->GetYaxis()->SetTitle(i == 0 ? "Number of TPC clusters" : i == 1 ? "Fraction of TPC clusters" : CLUST_HIST_INT_SUM ? "Total TPC clusters (integrated)" : "Fraction of TPC clusters (integrated)");
2826 e->GetXaxis()->SetTitle("#it{p}_{Tmc} (GeV/#it{c})");
2827 e->GetXaxis()->SetTitleOffset(1.1);
2828 e->GetXaxis()->SetLabelOffset(-0.005);
2829 if (tout && !mConfig.inputHistogramsOnly && k == 0) {
2830 e->Write();
2831 }
2832 e->SetStats(kFALSE);
2833 e->SetLineWidth(1);
2834 e->SetLineStyle(CONFIG_DASHED_MARKERS ? j + 1 : 1);
2835 if (i == 0) {
2836 e->GetXaxis()->SetRange(2, AXIS_BINS[4]);
2837 }
2838 if (qcout && !mConfig.shipToQCAsCanvas) {
2839 qcout->Add(e);
2840 }
2841 if (!mConfig.enableLocalOutput && !mConfig.shipToQCAsCanvas) {
2842 continue;
2843 }
2844
2845 e->SetMarkerColor(kBlack);
2846 e->SetLineColor(colorNums[numColor++ % COLORCOUNT]);
2847 e->Draw(j == end - 1 && k == 0 ? "" : "same");
2848 GetName(fname, k);
2849 mLClust[i]->AddEntry(e, Form("%s%s", fname, CLUSTER_NAMES[j - begin]), "l");
2850 }
2851 }
2852 if (ConfigNumInputs == 1) {
2853 TH1* e = reinterpret_cast<TH1F*>(mClusters[begin + CL_att_adj]->Clone());
2854 e->Add(mClusters[begin + CL_prot], -1);
2855 if (qcout && !mConfig.shipToQCAsCanvas) {
2856 qcout->Add(e);
2857 }
2858 if (!mConfig.enableLocalOutput && !mConfig.shipToQCAsCanvas) {
2859 continue;
2860 }
2861
2862 e->SetLineColor(colorNums[numColor++ % COLORCOUNT]);
2863 e->Draw("same");
2864 mLClust[i]->AddEntry(e, "Removed (Strategy A)", "l");
2865 }
2866 if (!mConfig.enableLocalOutput && !mConfig.shipToQCAsCanvas) {
2867 continue;
2868 }
2869
2870 mLClust[i]->Draw();
2871
2872 if (qcout) {
2873 qcout->Add(mCClust[i]);
2874 }
2875 if (!mConfig.enableLocalOutput) {
2876 continue;
2877 }
2878 doPerfFigure(i == 0 ? 0.37 : (i == 1 ? 0.34 : 0.6), 0.295, 0.030);
2879 mCClust[i]->cd();
2880 mCClust[i]->Print(Form(i == 2 ? "%s/clusters_integral.pdf" : i == 1 ? "%s/clusters_relative.pdf" : "%s/clusters.pdf", mConfig.plotsDir.c_str()));
2881 if (mConfig.writeFileExt != "") {
2882 mCClust[i]->Print(Form(i == 2 ? "%s/clusters_integral.%s" : i == 1 ? "%s/clusters_relative.%s" : "%s/clusters.%s", mConfig.plotsDir.c_str(), mConfig.writeFileExt.c_str()));
2883 }
2884 }
2885
2886 for (int32_t i = 0; i < 4; i++) {
2887 auto* e = mPadRow[i];
2888 if (tout && !mConfig.inputHistogramsOnly) {
2889 e->Write();
2890 }
2891 mPPadRow[i]->cd();
2892 e->SetOption("colz");
2893 std::string title = "First Track Pad Row (p_{T} > 1GeV, N_{Cl} #geq " + std::to_string(PADROW_CHECK_MINCLS);
2894 if (i >= 2) {
2895 title += ", row_{trk} > row_{MC} + 3, row_{MC} < 10";
2896 }
2897 if (i >= 3) {
2898 title += ", #Phi_{Cl} < 0.15";
2899 }
2900 title += ")";
2901
2902 e->SetTitle(mConfig.plotsNoTitle ? "" : title.c_str());
2903 e->GetXaxis()->SetTitle(i == 3 ? "Local Occupancy" : (i ? "#Phi_{Cl} (sector)" : "First MC Pad Row"));
2904 e->GetYaxis()->SetTitle("First Pad Row");
2905 e->Draw();
2906 mCPadRow[i]->cd();
2907 static const constexpr char* PADROW_NAMES[4] = {"MC", "Phi", "Phi1", "Occ"};
2908 mCPadRow[i]->Print(Form("%s/padRow%s.pdf", mConfig.plotsDir.c_str(), PADROW_NAMES[i]));
2909 if (mConfig.writeFileExt != "") {
2910 mCPadRow[i]->Print(Form("%s/padRow%s.%s", mConfig.plotsDir.c_str(), PADROW_NAMES[i], mConfig.writeFileExt.c_str()));
2911 }
2912 }
2913 }
2914
2915 // Process cluster count statistics
2916 if ((mQATasks & taskClusterCounts) && !mHaveExternalHists && !mConfig.clusterRejectionHistograms && !mConfig.inputHistogramsOnly) {
2917 DoClusterCounts(attachClusterCounts);
2918 }
2919 if ((qcout || tout) && (mQATasks & taskClusterCounts) && mConfig.clusterRejectionHistograms) {
2920 for (uint32_t i = 0; i < mHistClusterCount.size(); i++) {
2921 if (tout) {
2922 mHistClusterCount[i]->Write();
2923 }
2924 if (qcout) {
2925 qcout->Add(mHistClusterCount[i]);
2926 }
2927 }
2928 }
2929
2930 if (mQATasks & taskTrackStatistics) {
2931 // Process track statistic histograms
2932
2933 float tmpMax = 0.;
2934 for (int32_t k = 0; k < ConfigNumInputs; k++) { // TODO: Simplify this drawing, avoid copy&paste
2935 TH1F* e = mTrackPt;
2936 if (GetHist(e, tin, k, nNewInput) == nullptr) {
2937 continue;
2938 }
2939 e->SetMaximum(-1111);
2940 if (e->GetMaximum() > tmpMax) {
2941 tmpMax = e->GetMaximum();
2942 }
2943 }
2944 mPTracks[0]->cd();
2945 mPTracks[0]->SetLogx();
2946 for (int32_t k = 0; k < ConfigNumInputs; k++) {
2947 TH1F* e = mTrackPt;
2948 if (GetHist(e, tin, k, nNewInput) == nullptr) {
2949 continue;
2950 }
2951 if (tout && !mConfig.inputHistogramsOnly && k == 0) {
2952 e->Write();
2953 }
2954 e->SetMaximum(tmpMax * 1.02);
2955 e->SetMinimum(tmpMax * -0.02);
2956 e->SetStats(kFALSE);
2957 e->SetLineWidth(1);
2958 e->SetTitle(mConfig.plotsNoTitle ? "" : "Number of Tracks vs #it{p}_{T}");
2959 e->GetYaxis()->SetTitle("Number of Tracks");
2960 e->GetXaxis()->SetTitle("#it{p}_{T} (GeV/#it{c})");
2961 e->GetXaxis()->SetTitleOffset(1.2);
2962 if (qcout) {
2963 qcout->Add(e);
2964 }
2965 e->SetMarkerColor(kBlack);
2966 e->SetLineColor(colorNums[k % COLORCOUNT]);
2967 e->Draw(k == 0 ? "" : "same");
2968 GetName(fname, k, mConfig.inputHistogramsOnly);
2969 mLTracks[0]->AddEntry(e, Form(mConfig.inputHistogramsOnly ? "%s" : "%sTrack #it{p}_{T}", fname), "l");
2970 }
2971 mLTracks[0]->Draw();
2972 doPerfFigure(0.63, 0.7, 0.030);
2973 mCTracks[0]->cd();
2974 mCTracks[0]->Print(Form("%s/tracks%s.pdf", mConfig.plotsDir.c_str(), "pt"));
2975 if (mConfig.writeFileExt != "") {
2976 mCTracks[0]->Print(Form("%s/tracks%s.%s", mConfig.plotsDir.c_str(), "pt", mConfig.writeFileExt.c_str()));
2977 }
2978
2979 for (int32_t i = 0; i < 2; i++) {
2980 mPTracks[1 + i]->cd();
2981 {
2982 TH2F* e = mTrackdEdx[i];
2983 if (tout && !mConfig.inputHistogramsOnly) {
2984 e->Write();
2985 }
2986 // e->SetStats(kFALSE);
2987 e->SetTitle(mConfig.plotsNoTitle ? "" : (i ? "Track dE/dx (Max)" : "Track dE/dx (Tot)"));
2988 e->GetYaxis()->SetTitle(i ? "dE/dx (max)" : "dE/dx (tot)");
2989 e->GetXaxis()->SetTitle("log(#it{p})");
2990 e->GetXaxis()->SetTitleOffset(1.2);
2991 if (qcout) {
2992 qcout->Add(e);
2993 }
2994 e->SetOption("colz");
2995 e->Draw();
2996 }
2997 mCTracks[1 + i]->cd();
2998 mCTracks[1 + i]->Print(Form("%s/tracks%s.pdf", mConfig.plotsDir.c_str(), i ? "dedx_max" : "dedx_tot"));
2999 if (mConfig.writeFileExt != "") {
3000 mCTracks[1 + i]->Print(Form("%s/tracks%s.%s", mConfig.plotsDir.c_str(), i ? "dedx_max" : "dedx_tot", mConfig.writeFileExt.c_str()));
3001 }
3002
3003 tmpMax = 0.;
3004 for (int32_t k = 0; k < ConfigNumInputs; k++) {
3005 TH1F* e = mT0[i];
3006 if (GetHist(e, tin, k, nNewInput) == nullptr) {
3007 continue;
3008 }
3009 e->SetMaximum(-1111);
3010 if (e->GetMaximum() > tmpMax) {
3011 tmpMax = e->GetMaximum();
3012 }
3013 }
3014 mPT0[i]->cd();
3015 for (int32_t k = 0; k < ConfigNumInputs; k++) {
3016 TH1F* e = mT0[i];
3017 if (GetHist(e, tin, k, nNewInput) == nullptr) {
3018 continue;
3019 }
3020 if (tout && !mConfig.inputHistogramsOnly && k == 0) {
3021 e->Write();
3022 }
3023 e->SetMaximum(tmpMax * 1.02);
3024 e->SetMinimum(tmpMax * -0.02);
3025 e->SetStats(kFALSE);
3026 e->SetLineWidth(1);
3027 e->SetTitle(mConfig.plotsNoTitle ? "" : (i ? "Track t_{0} resolution" : "Track t_{0} distribution"));
3028 e->GetYaxis()->SetTitle("a.u.");
3029 e->GetXaxis()->SetTitle(i ? "t_{0} - t_{0, mc}" : "t_{0}");
3030 if (qcout) {
3031 qcout->Add(e);
3032 }
3033 e->SetMarkerColor(kBlack);
3034 e->SetLineColor(colorNums[k % COLORCOUNT]);
3035 e->Draw(k == 0 ? "" : "same");
3036 GetName(fname, k, mConfig.inputHistogramsOnly);
3037 mLT0[i]->AddEntry(e, Form(mConfig.inputHistogramsOnly ? "%s (%s)" : "%sTrack t_{0} %s", fname, i ? "" : "resolution"), "l");
3038 }
3039 mLT0[i]->Draw();
3040 doPerfFigure(0.63, 0.7, 0.030);
3041 mCT0[i]->cd();
3042 mCT0[i]->Print(Form("%s/t0%s.pdf", mConfig.plotsDir.c_str(), i ? "_res" : ""));
3043 if (mConfig.writeFileExt != "") {
3044 mCT0[i]->Print(Form("%s/t0%s.%s", mConfig.plotsDir.c_str(), i ? "_res" : "", mConfig.writeFileExt.c_str()));
3045 }
3046
3047 tmpMax = 0.;
3048 for (int32_t k = 0; k < ConfigNumInputs; k++) {
3049 TH1F* e = mNCl[i];
3050 if (GetHist(e, tin, k, nNewInput) == nullptr) {
3051 continue;
3052 }
3053 e->SetMaximum(-1111);
3054 if (e->GetMaximum() > tmpMax) {
3055 tmpMax = e->GetMaximum();
3056 }
3057 }
3058 mPNCl[i]->cd();
3059 for (int32_t k = 0; k < ConfigNumInputs; k++) {
3060 TH1F* e = mNCl[i];
3061 if (GetHist(e, tin, k, nNewInput) == nullptr) {
3062 continue;
3063 }
3064 if (tout && !mConfig.inputHistogramsOnly && k == 0) {
3065 e->Write();
3066 }
3067 e->SetMaximum(tmpMax * 1.02);
3068 e->SetMinimum(tmpMax * -0.02);
3069 e->SetStats(kFALSE);
3070 e->SetLineWidth(1);
3071 e->SetTitle(mConfig.plotsNoTitle ? "" : (i ? "Number of Rows with attached Cluster" : "Number of Clusters"));
3072 e->GetYaxis()->SetTitle("a.u.");
3073 e->GetXaxis()->SetTitle(i ? "N_{Rows with Clusters}" : "N_{Clusters}");
3074 if (qcout) {
3075 qcout->Add(e);
3076 }
3077 e->SetMarkerColor(kBlack);
3078 e->SetLineColor(colorNums[k % COLORCOUNT]);
3079 e->Draw(k == 0 ? "" : "same");
3080 GetName(fname, k, mConfig.inputHistogramsOnly);
3081 mLNCl[i]->AddEntry(e, Form(mConfig.inputHistogramsOnly ? "%s" : (i ? "%sN_{Clusters}" : "%sN_{Rows with Clusters}"), fname), "l");
3082 }
3083 mLNCl[i]->Draw();
3084 doPerfFigure(0.6, 0.7, 0.030);
3085 mCNCl[i]->cd();
3086 mCNCl[i]->Print(Form("%s/nClusters%s.pdf", mConfig.plotsDir.c_str(), i ? "_corrected" : ""));
3087 if (mConfig.writeFileExt != "") {
3088 mCNCl[i]->Print(Form("%s/nClusters%s.%s", mConfig.plotsDir.c_str(), i ? "_corrected" : "", mConfig.writeFileExt.c_str()));
3089 }
3090 }
3091
3092 mPClXY->cd(); // TODO: This should become a separate task category
3093 mClXY->SetOption("colz");
3094 mClXY->Draw();
3095 mCClXY->cd();
3096 mCClXY->Print(Form("%s/clustersXY.pdf", mConfig.plotsDir.c_str()));
3097 if (mConfig.writeFileExt != "") {
3098 mCClXY->Print(Form("%s/clustersXY.%s", mConfig.plotsDir.c_str(), mConfig.writeFileExt.c_str()));
3099 }
3100 }
3101
3102 if (mQATasks & taskClusterRejection) {
3103 mClRej[2]->Divide(mClRej[1], mClRej[0]);
3104
3105 for (int32_t i = 0; i < 3; i++) {
3106 if (tout && !mConfig.inputHistogramsOnly) {
3107 mClRej[i]->Write();
3108 }
3109 mPClRej[i]->cd();
3110 mClRej[i]->SetTitle(mConfig.plotsNoTitle ? "" : REJECTED_NAMES[i]);
3111 mClRej[i]->SetOption("colz");
3112 mClRej[i]->Draw();
3113 mCClRej[i]->cd();
3114 mCClRej[i]->Print(Form("%s/clustersRej%d%s.pdf", mConfig.plotsDir.c_str(), i, REJECTED_NAMES[i]));
3115 if (mConfig.writeFileExt != "") {
3116 mCClRej[i]->Print(Form("%s/clustersRej%d%s.%s", mConfig.plotsDir.c_str(), i, REJECTED_NAMES[i], mConfig.writeFileExt.c_str()));
3117 }
3118 }
3119
3120 mPClRejP->cd();
3121 for (int32_t k = 0; k < ConfigNumInputs; k++) {
3122 auto* tmp = mClRej[0];
3123 if (GetHist(tmp, tin, k, nNewInput) == nullptr) {
3124 continue;
3125 }
3126 TH1D* proj1 = tmp->ProjectionY(Form("clrejptmp1%d", k)); // TODO: Clean up names
3127 proj1->SetDirectory(nullptr);
3128 tmp = mClRej[1];
3129 if (GetHist(tmp, tin, k, nNewInput) == nullptr) {
3130 continue;
3131 }
3132 TH1D* proj2 = tmp->ProjectionY(Form("clrejptmp2%d", k));
3133 proj2->SetDirectory(nullptr);
3134
3135 auto* e = mClRejP;
3136 if (GetHist(e, tin, k, nNewInput) == nullptr) {
3137 continue;
3138 }
3139 e->Divide(proj2, proj1);
3140 if (tout && !mConfig.inputHistogramsOnly && k == 0) {
3141 e->Write();
3142 }
3143 delete proj1;
3144 delete proj2;
3145 e->SetMinimum(-0.02);
3146 e->SetMaximum(0.22);
3147 e->SetTitle(mConfig.plotsNoTitle ? "" : "Rejected Clusters");
3148 e->GetXaxis()->SetTitle("Pad Row");
3149 e->GetYaxis()->SetTitle("Rejected Clusters (fraction)");
3150 e->Draw(k == 0 ? "" : "same");
3151 }
3152 mPClRejP->Print(Form("%s/clustersRejProjected.pdf", mConfig.plotsDir.c_str()));
3153 if (mConfig.writeFileExt != "") {
3154 mPClRejP->Print(Form("%s/clustersRejProjected.%s", mConfig.plotsDir.c_str(), mConfig.writeFileExt.c_str()));
3155 }
3156 }
3157
3158 if (tout && !mConfig.inputHistogramsOnly && mConfig.writeMCLabels) {
3159 gInterpreter->GenerateDictionary("vector<vector<int32_t>>", "");
3160 tout->WriteObject(&mcEffBuffer, "mcEffBuffer");
3161 tout->WriteObject(&mcLabelBuffer, "mcLabelBuffer");
3162 remove("AutoDict_vector_vector_int__.cxx");
3163 remove("AutoDict_vector_vector_int___cxx_ACLiC_dict_rdict.pcm");
3164 remove("AutoDict_vector_vector_int___cxx.d");
3165 remove("AutoDict_vector_vector_int___cxx.so");
3166 }
3167
3168 if (tout) {
3169 tout->Close();
3170 }
3171 for (uint32_t i = 0; i < mConfig.compareInputs.size(); i++) {
3172 tin[i]->Close();
3173 }
3174 if (!qcout) {
3175 clearGarbagageCollector();
3176 }
3177 GPUInfo("GPU TPC QA histograms have been written to pdf%s%s files", mConfig.writeFileExt == "" ? "" : " and ", mConfig.writeFileExt.c_str());
3178 gErrorIgnoreLevel = oldRootIgnoreLevel;
3179 return (0);
3180}
3181
3182void GPUQA::PrintClusterCount(int32_t mode, int32_t& num, const char* name, uint64_t n, uint64_t normalization)
3183{
3184 if (mode == 2) {
3185 // do nothing, just count num
3186 } else if (mode == 1) {
3187 char name2[128];
3188 snprintf(name2, 128, "clusterCount%d_", num);
3189 char* ptr = name2 + strlen(name2);
3190 for (uint32_t i = 0; i < strlen(name); i++) {
3191 if ((name[i] >= 'a' && name[i] <= 'z') || (name[i] >= 'A' && name[i] <= 'Z') || (name[i] >= '0' && name[i] <= '9')) {
3192 *(ptr++) = name[i];
3193 }
3194 }
3195 *ptr = 0;
3196 createHist(mHistClusterCount[num], name2, name, 1000, 0, mConfig.histMaxNClusters, 1000, 0, 100);
3197 } else if (mode == 0) {
3198 if (normalization && mConfig.enableLocalOutput) {
3199 for (uint32_t i = 0; i < 1 + (mTextDump != nullptr); i++) {
3200 fprintf(i ? mTextDump : stdout, "\t%40s: %'12" PRIu64 " (%6.2f%%)\n", name, n, 100.f * n / normalization);
3201 }
3202 }
3203 if (mConfig.clusterRejectionHistograms) {
3204 float ratio = 100.f * n / std::max<uint64_t>(normalization, 1);
3205 mHistClusterCount[num]->Fill(normalization, ratio, 1);
3206 }
3207 }
3208 num++;
3209}
3210
3211int32_t GPUQA::DoClusterCounts(uint64_t* attachClusterCounts, int32_t mode)
3212{
3213 if (mConfig.enableLocalOutput && !mConfig.inputHistogramsOnly && mConfig.plotsDir != "") {
3214 mTextDump = fopen((mConfig.plotsDir + "/clusterCounts.txt").c_str(), "w+");
3215 }
3216 int32_t num = 0;
3217 if (mcPresent() && (mQATasks & taskClusterAttach) && attachClusterCounts) {
3218 for (int32_t i = 0; i < N_CLS_HIST; i++) { // TODO: Check that these counts are still printed correctly!
3219 PrintClusterCount(mode, num, CLUSTER_NAMES[i], attachClusterCounts[i], mClusterCounts.nTotal);
3220 }
3221 PrintClusterCount(mode, num, "Unattached", attachClusterCounts[N_CLS_HIST - 1] - attachClusterCounts[CL_att_adj], mClusterCounts.nTotal);
3222 PrintClusterCount(mode, num, "Removed (Strategy A)", attachClusterCounts[CL_att_adj] - attachClusterCounts[CL_prot], mClusterCounts.nTotal); // Attached + Adjacent (also fake) - protected
3223 PrintClusterCount(mode, num, "Unaccessible", mClusterCounts.nUnaccessible, mClusterCounts.nTotal); // No contribution from track >= 10 MeV, unattached or fake-attached/adjacent
3224 } else {
3225 PrintClusterCount(mode, num, "All Clusters", mClusterCounts.nTotal, mClusterCounts.nTotal);
3226 PrintClusterCount(mode, num, "Used in Physics", mClusterCounts.nPhysics, mClusterCounts.nTotal);
3227 PrintClusterCount(mode, num, "Protected", mClusterCounts.nProt, mClusterCounts.nTotal);
3228 PrintClusterCount(mode, num, "Unattached", mClusterCounts.nUnattached, mClusterCounts.nTotal);
3229 PrintClusterCount(mode, num, "Removed (Strategy A)", mClusterCounts.nTotal - mClusterCounts.nUnattached - mClusterCounts.nProt, mClusterCounts.nTotal);
3230 PrintClusterCount(mode, num, "Removed (Strategy B)", mClusterCounts.nTotal - mClusterCounts.nProt, mClusterCounts.nTotal);
3231 }
3232
3233 PrintClusterCount(mode, num, "Merged Loopers (Track Merging)", mClusterCounts.nMergedLooperConnected, mClusterCounts.nTotal);
3234 PrintClusterCount(mode, num, "Merged Loopers (Afterburner)", mClusterCounts.nMergedLooperUnconnected, mClusterCounts.nTotal);
3235 PrintClusterCount(mode, num, "Looping Legs (other)", mClusterCounts.nLoopers, mClusterCounts.nTotal);
3236 PrintClusterCount(mode, num, "High Inclination Angle", mClusterCounts.nHighIncl, mClusterCounts.nTotal);
3237 PrintClusterCount(mode, num, "Rejected", mClusterCounts.nRejected, mClusterCounts.nTotal);
3238 PrintClusterCount(mode, num, "Tube (> 200 MeV)", mClusterCounts.nTube, mClusterCounts.nTotal);
3239 PrintClusterCount(mode, num, "Tube (< 200 MeV)", mClusterCounts.nTube200, mClusterCounts.nTotal);
3240 PrintClusterCount(mode, num, "Low Pt < 50 MeV", mClusterCounts.nLowPt, mClusterCounts.nTotal);
3241 PrintClusterCount(mode, num, "Low Pt < 200 MeV", mClusterCounts.n200MeV, mClusterCounts.nTotal);
3242
3243 if (mcPresent() && (mQATasks & taskClusterAttach)) {
3244 PrintClusterCount(mode, num, "Tracks > 400 MeV", mClusterCounts.nAbove400, mClusterCounts.nTotal);
3245 PrintClusterCount(mode, num, "Fake Removed (> 400 MeV)", mClusterCounts.nFakeRemove400, mClusterCounts.nAbove400);
3246 PrintClusterCount(mode, num, "Full Fake Removed (> 400 MeV)", mClusterCounts.nFullFakeRemove400, mClusterCounts.nAbove400);
3247 PrintClusterCount(mode, num, "Tracks < 40 MeV", mClusterCounts.nBelow40, mClusterCounts.nTotal);
3248 PrintClusterCount(mode, num, "Fake Protect (< 40 MeV)", mClusterCounts.nFakeProtect40, mClusterCounts.nBelow40);
3249 }
3250 if (mcPresent() && (mQATasks & taskTrackStatistics)) {
3251 PrintClusterCount(mode, num, "Correctly Attached all-trk normalized", mClusterCounts.nCorrectlyAttachedNormalized, mClusterCounts.nTotal);
3252 PrintClusterCount(mode, num, "Correctly Attached non-fake normalized", mClusterCounts.nCorrectlyAttachedNormalizedNonFake, mClusterCounts.nTotal);
3253 }
3254 if (mTextDump) {
3255 fclose(mTextDump);
3256 mTextDump = nullptr;
3257 }
3258 return num;
3259}
3260
3261void* GPUQA::AllocateScratchBuffer(size_t nBytes)
3262{
3263 mTrackingScratchBuffer.resize((nBytes + sizeof(mTrackingScratchBuffer[0]) - 1) / sizeof(mTrackingScratchBuffer[0]));
3264 return mTrackingScratchBuffer.data();
3265}
labels
std::vector< std::string > labels
Definition AODReaderHelpers.cxx:32
ConstMCTruthContainer.h
A const (ready only) version of MCTruthContainer.
charge
int16_t charge
Definition RawEventData.h:5
time
int16_t time
Definition RawEventData.h:4
i
int32_t i
Definition GPUCommonAlgorithm.h:451
TRACK_EXPECTED_REFERENCE_X_DEFAULT
#define TRACK_EXPECTED_REFERENCE_X_DEFAULT
Definition GPUQA.cxx:206
TRACK_EXPECTED_REFERENCE_X
#define TRACK_EXPECTED_REFERENCE_X
Definition GPUQA.cxx:231
QA_DEBUG
#define QA_DEBUG
Definition GPUQA.cxx:15
QA_TIMING
#define QA_TIMING
Definition GPUQA.cxx:16
Color_t
int16_t Color_t
Definition GPUQA.h:33
iSector
uint32_t iSector
Definition GPUReconstructionConvert.cxx:191
chain
GPUChain * chain
Definition GPUReconstruction.cxx:57
MCTrack.h
Definition of the MCTrack class.
NameConf.h
Definition of the Names Generator class.
pos
uint16_t pos
Definition RawData.h:3
j
uint32_t j
Definition RawData.h:0
side
uint32_t side
Definition RawData.h:0
pid
uint16_t pid
Definition RawData.h:2
c
uint32_t c
Definition RawData.h:2
TPCFastTransformPOD.h
POD correction map.
ptr
TBranch * ptr
Definition TableTreeHelpers.cxx:293
nClusters
int nClusters
Definition bench_Clusterizer.cxx:120
num
double num
Definition bench_Clusterizer.cxx:122
h
Class for time synchronization of RawReader instances.
o2::detectors::DetID::TPC
static constexpr ID TPC
Definition DetID.h:64
o2::gpu::GPUChain::NSECTORS
static constexpr int32_t NSECTORS
Definition GPUChain.h:58
o2::gpu::GPUQA::ReadO2MCData
int32_t ReadO2MCData(const char *filename)
Definition GPUQA.h:55
o2::gpu::GPUQA::clusterRemovable
bool clusterRemovable(int32_t attach, bool prot) const
Definition GPUQA.h:53
o2::gpu::GPUQA::~GPUQA
~GPUQA()=default
Definition GPUQA.cxx:338
o2::gpu::GPUQA::AllocateScratchBuffer
void * AllocateScratchBuffer(size_t nBytes)
Definition GPUQA.h:56
o2::gpu::GPUQA::SetMCTrackRange
void SetMCTrackRange(int32_t min, int32_t max)
Definition GPUQA.h:48
o2::gpu::GPUQA::GetMCTrackLabel
mcLabelI_t GetMCTrackLabel(uint32_t trackId) const
Definition GPUQA.h:52
o2::gpu::GPUQA::DrawQAHistograms
int32_t DrawQAHistograms()
Definition GPUQA.h:47
o2::gpu::GPUQA::InitQA
int32_t InitQA(int32_t tasks=0)
Definition GPUQA.h:45
o2::gpu::GPUQA::DumpO2MCData
void DumpO2MCData(const char *filename) const
Definition GPUQA.h:54
o2::gpu::GPUQA::mcLabelI_t
int32_t mcLabelI_t
Definition GPUQA.h:44
o2::gpu::GPUQA::RunQA
void RunQA(bool matchOnly=false)
Definition GPUQA.h:46
o2::gpu::GPUQA::GPUQA
GPUQA(void *chain)
Definition GPUQA.h:42
o2::gpu::GPUTPCGeometry::NROWS
static constexpr uint32_t NROWS
Definition GPUTPCGeometry.h:108
o2::gpu::GPUTPCGeometry::NSECTORS
static constexpr uint32_t NSECTORS
Definition GPUTPCGeometry.h:107
o2::steer::DigitizationContext::loadFromFile
static DigitizationContext * loadFromFile(std::string_view filename="")
Definition DigitizationContext.cxx:290
n
GLdouble n
Definition glcorearb.h:1982
alpha
GLfloat GLfloat GLfloat alpha
Definition glcorearb.h:279
x
GLint GLenum GLint x
Definition glcorearb.h:403
m
const GLfloat * m
Definition glcorearb.h:4066
mode
GLenum mode
Definition glcorearb.h:266
src
GLenum src
Definition glcorearb.h:1767
y1
GLuint GLfloat GLfloat GLfloat GLfloat y1
Definition glcorearb.h:5034
size
GLsizeiptr size
Definition glcorearb.h:659
end
GLuint GLuint end
Definition glcorearb.h:469
v
const GLdouble * v
Definition glcorearb.h:832
name
GLuint const GLchar * name
Definition glcorearb.h:781
f
GLdouble f
Definition glcorearb.h:310
weight
GLuint GLuint GLfloat weight
Definition glcorearb.h:5477
range
GLenum GLint * range
Definition glcorearb.h:1899
y
GLint y
Definition glcorearb.h:270
dst
GLenum GLenum dst
Definition glcorearb.h:1767
data
GLboolean * data
Definition glcorearb.h:298
label
GLuint GLsizei const GLchar * label
Definition glcorearb.h:2519
val
GLuint GLfloat * val
Definition glcorearb.h:1582
r
GLboolean r
Definition glcorearb.h:1233
param
GLenum GLfloat param
Definition glcorearb.h:271
id
GLuint id
Definition glcorearb.h:650
z
GLdouble GLdouble GLdouble z
Definition glcorearb.h:843
o2::its::math_utils::p
float p
Definition MathUtils.h:123
o2::its::math_utils::y2
float float float y2
Definition MathUtils.h:49
o2::its::track::nCl
o2::track::TrackParCov int int int float int nCl
Definition TrackHelpers.h:196
o2::its::track
const bool const int TrackITSInternal< NLayers > & track
Definition TrackFollower.h:62
o2::math_utils::detail::phi
T phi
Definition trigonometric.h:221
o2::math_utils::detail::s
T & s
Definition trigonometric.h:121
o2::math_utils::detail::T
T
Definition SMatrixGPU.h:749
o2::tpc::constants::LHCBCPERTIMEBIN
constexpr int LHCBCPERTIMEBIN
Definition Constants.h:38
o2::tpc::begin
Enum< T >::Iterator begin(Enum< T >)
Definition Defs.h:156
o2::track::f3
value_T f3
Definition TrackUtils.h:93
o2::track::f1
value_T f1
Definition TrackUtils.h:91
o2::track::f2
value_T f2
Definition TrackUtils.h:92
o2::upgrades_utils::tracks
struct o2::upgrades_utils::@469 tracks
structure to keep trigger-related info
std::to_string
std::string to_string(gsl::span< T, Size > span)
Definition common.h:52
filename
std::string filename()
Definition o2FairMQHeaderSizeTest.cxx:55
isValid
bool isValid(std::string alias)
Definition scan-hvlv-ccdb.cxx:74
o2::InteractionRecord::differenceInBC
int64_t differenceInBC(const InteractionRecord &other) const
Definition InteractionRecord.h:94
o2::gpu::internal::GPUQAGarbageCollection::v
std::tuple< std::vector< std::unique_ptr< TCanvas > >, std::vector< std::unique_ptr< TLegend > >, std::vector< std::unique_ptr< TPad > >, std::vector< std::unique_ptr< TLatex > >, std::vector< std::unique_ptr< TH1D > > > v
Definition GPUQA.cxx:316
o2::raw::HBFUtils::getFirstIRofTF
IR getFirstIRofTF(const IR &rec) const
get 1st IR of TF corresponding to the 1st sampled orbit (in MC)
Definition HBFUtils.h:71
o2::raw::HBFUtils::getFirstIR
IR getFirstIR() const
Definition HBFUtils.h:47
o2::tpc::ClusterNativeAccess::nClusters
unsigned int nClusters[constants::MAXSECTOR][constants::MAXGLOBALPADROW]
Definition ClusterNative.h:307
o2::tpc::ClusterNativeAccess::clusterOffset
unsigned int clusterOffset[constants::MAXSECTOR][constants::MAXGLOBALPADROW]
Definition ClusterNative.h:309
o2::tpc::ClusterNativeAccess::clustersLinear
const ClusterNative * clustersLinear
Definition ClusterNative.h:304
min
constexpr size_t min
Definition test_Algorithm.cxx:50
max
constexpr size_t max
Definition test_Algorithm.cxx:51
ir
o2::InteractionRecord ir(0, 0)
clear
vec clear()
ir0
o2::InteractionRecord ir0(3, 5)