TRIGGER-AWARE ANALYSIS TUTORIAL ARTEMIS Workshop – Pisa – 18 th to 19 th June 2009 Alessandro Cerri (CERN), Ricardo Gonçalo – Royal Holloway

ARTEMIS - Pisa, June Trigger Tutorial EMROI L2 calorim. L2 tracking OK? E.F.calorim. match? E.F.tracking track? e  OK? e/  reconst. L2 calorim. OK? E.F.calorim.  OK? e/  reconst. TrigEMCluster TrigInDetTracks CaloCluster egamma T.E. FEature eXtraction algos produce Features on which selection in HYPOthesis algos is based FEX HYPO features ESDAODTAGDPD features T.E. passed e/γ trigger decision

 Trigger configuation:  L1 Items: ▪ Name, version, CTP-Id, prescale  HLT Chains: ▪ Name, version, level, counter, prescale  Streams: ▪ Chains feeding into each stream  Chain-groups: ▪ Chains belonging to each group  Bunch-groups: ▪ Name of each of the 8 BG  Trigger objects (“features”)  Data objects: ▪ L2 has it’s own EDM: ▪ TrigTau, TrigInDetTrack, TrigEMCluster ▪ Event Filter uses mostly offline EDM ▪ CaloCluster, egamma, TrackParticle  Navigation links: TriggerElements  Trigger configuation:  L1 Items: ▪ Name, version, CTP-Id, prescale  HLT Chains: ▪ Name, version, level, counter, prescale  Streams: ▪ Chains feeding into each stream  Chain-groups: ▪ Chains belonging to each group  Bunch-groups: ▪ Name of each of the 8 BG  Trigger objects (“features”)  Data objects: ▪ L2 has it’s own EDM: ▪ TrigTau, TrigInDetTrack, TrigEMCluster ▪ Event Filter uses mostly offline EDM ▪ CaloCluster, egamma, TrackParticle  Navigation links: TriggerElements ARTEMIS - Pisa, June Trigger Tutorial TrigEMCluster TrigInDetTrack CaloCluster egamma T.E.

ARTEMIS - Pisa, June Trigger Tutorial

TrigDecisionTool Decision + Navigation EDMConfiguration ESDAODDPD contain persistent ConfigSvc IOVDbSvc IOV based P/T conv. EDM P/T convertion Unpacking Dec. + Nav. TrigDecisionFeature Nav.Configuration Upon Request Transient:

1. ToolHandle to a TrigDecisionTool in your algorithm header 2. Initialize ToolHandle in constructor as public tool 3. Retrieve tool in initialization 4. Use tool in execute() 1. ToolHandle to a TrigDecisionTool in your algorithm header 2. Initialize ToolHandle in constructor as public tool 3. Retrieve tool in initialization 4. Use tool in execute() ARTEMIS - Pisa, June Trigger Tutorial private: ToolHandle m_trigDec; private: ToolHandle m_trigDec; MyAlgo::MyAlgo(const std::string &name, …) : m_trigDec("Trig::TrigDecisionTool/TrigDecisionTool”) { declareProperty("TrigDecisionTool", m_trigDec, \ ”The tool to access TrigDecision”); } MyAlgo::MyAlgo(const std::string &name, …) : m_trigDec("Trig::TrigDecisionTool/TrigDecisionTool”) { declareProperty("TrigDecisionTool", m_trigDec, \ ”The tool to access TrigDecision”); } StatusCode sc = m_trigDec.retrieve(); (*m_log) << MSG::INFO << "L2_mu.* passed/failed: " isPassed("L2_mu.*") << endreq; (*m_log) << MSG::INFO << "L2_mu.* passed/failed: " isPassed("L2_mu.*") << endreq;

 Important concept in the new tool: ChainGroup  They work as an OR of triggers  Created from a vector of trigger names, comma- separated list or regular expressions:  const ChainGroup* getChainGroup(const string& names)  Can be used e.g. to get pass/fail information:  Can be created on the fly:   Important concept in the new tool: ChainGroup  They work as an OR of triggers  Created from a vector of trigger names, comma- separated list or regular expressions:  const ChainGroup* getChainGroup(const string& names)  Can be used e.g. to get pass/fail information:  Can be created on the fly:  ARTEMIS - Pisa, June Trigger Tutorial

 Each of the methods cbelow an also be accessed through the TrigDecisionTool instance: bool isPassed (condition=Physics) const ; const FeatureContainer features (condition=Physics) const float getPrescale (condition=Physics) const ; bool isPassedBits (condition=Physics) const ; vector getListOfTriggers() const vector getListOfStreams() const vector getListOfGroups() const vector > getListOfTriggerElements() const  A new ChainGroup will be created if necessary. For example: bool TrigDecisionTool::isPassed(const ChainGroup& chaingroup,condition=Physics) const bool TrigDecisionTool::isPassed(const string& names,condition=Physics) const  The conditions mask handles logical trigger conditions.  Each of the methods cbelow an also be accessed through the TrigDecisionTool instance: bool isPassed (condition=Physics) const ; const FeatureContainer features (condition=Physics) const float getPrescale (condition=Physics) const ; bool isPassedBits (condition=Physics) const ; vector getListOfTriggers() const vector getListOfStreams() const vector getListOfGroups() const vector > getListOfTriggerElements() const  A new ChainGroup will be created if necessary. For example: bool TrigDecisionTool::isPassed(const ChainGroup& chaingroup,condition=Physics) const bool TrigDecisionTool::isPassed(const string& names,condition=Physics) const  The conditions mask handles logical trigger conditions. ARTEMIS - Pisa, June Trigger Tutorial

 Use chains or chain groups to access trigger features produced by each trigger/group  FeatureContainer gives access to Trigger Elements and trigger objects for simple or complex triggers  Combinations allows to retrieve e.g. combinations of objects satisfying combined triggers like EF_tau16_2j23  Use chains or chain groups to access trigger features produced by each trigger/group  FeatureContainer gives access to Trigger Elements and trigger objects for simple or complex triggers  Combinations allows to retrieve e.g. combinations of objects satisfying combined triggers like EF_tau16_2j23 ARTEMIS - Pisa, June Trigger Tutorial FeatureContainer f = m_trigDec->features(“L2_e15”); FeatureContainer::combination_const_iterator it; for (it = f.getCombinations().begin(); it != f.getCombinations().end(); ++it) { std::vector > initRois = cIt->get ("initialRoI"); }

ARTEMIS - Pisa, June Trigger Tutorial

 Run-summary page: trigger chains, prescales, rates – web based  Query single run  AtlCoolTrigger.py: same, but command line  Views many runs  TriggerTool: Java based GUI to browse the TriggerDB (replica) for all information  as web front-end to TriggerTool and AtlCoolTrigger.py  Run-summary page: trigger chains, prescales, rates – web based  Query single run  AtlCoolTrigger.py: same, but command line  Views many runs  TriggerTool: Java based GUI to browse the TriggerDB (replica) for all information  as web front-end to TriggerTool and AtlCoolTrigger.py ARTEMIS - Pisa, June Trigger Tutorial

 Java based front end to TriggerDB, launch from the web (Java web-start):  Overview of all trigger configurations  Detailed and convenient investigation of trigger menus ▪ Trigger definition L1->L2->EF: prescales, threshold algorithms, selection criteria, streaming information, etc.  Possibility to compare different trigger configurations  Java based front end to TriggerDB, launch from the web (Java web-start):  Overview of all trigger configurations  Detailed and convenient investigation of trigger menus ▪ Trigger definition L1->L2->EF: prescales, threshold algorithms, selection criteria, streaming information, etc.  Possibility to compare different trigger configurations ARTEMIS - Pisa, June Trigger Tutorial

 Web interface  Runs TriggerTool on the server, result presented as dynamic html pages  Web interface  Runs TriggerTool on the server, result presented as dynamic html pages ARTEMIS - Pisa, June Trigger Tutorial 1. Search run-range 2. Run list 3. Trigger configuration (browsable) (definition, algorithms, selection cuts) Also with simple comparison functionality

 checkTrigger.py AOD.pool.root  Runs over ESD/AOD/DPD and presents detailed (chain‐wise) counts of the trigger decision  checkTriggerConfig.py ‐d AOD.pool.root  Runs over ESD/AOD/DPD and presents detailed trigger configuration(s) in the file  Shows multiple configurations, in DPD possible/likely  checkTrigger.py AOD.pool.root  Runs over ESD/AOD/DPD and presents detailed (chain‐wise) counts of the trigger decision  checkTriggerConfig.py ‐d AOD.pool.root  Runs over ESD/AOD/DPD and presents detailed trigger configuration(s) in the file  Shows multiple configurations, in DPD possible/likely ARTEMIS - Pisa, June Trigger Tutorial

 Triger Menu and L1 rates stored in COOL, HLT rates coming. Quick access via  Run summary pages (WEB based)  db-runlist/query.html db-runlist/query.html ▪ Trigger names, rates  AtlCoolTrigger.py (command line tool) ▪ AtlCoolTrigger –r 91000‐99000 (many run summary) ▪ AtlCoolTrigger –v –m –r (single run menu) ▪ Prints keys, trigger menus, streams, allows diff‐ing of menus in different runs  Triger Menu and L1 rates stored in COOL, HLT rates coming. Quick access via  Run summary pages (WEB based)  db-runlist/query.html db-runlist/query.html ▪ Trigger names, rates  AtlCoolTrigger.py (command line tool) ▪ AtlCoolTrigger –r 91000‐99000 (many run summary) ▪ AtlCoolTrigger –v –m –r (single run menu) ▪ Prints keys, trigger menus, streams, allows diff‐ing of menus in different runs ARTEMIS - Pisa, June Trigger Tutorial

General Trigger info: Tutorials: Trigger Event Data Model (EDM) : TrigDecisionTool – all releases/versions: Twiki: Athena examples: TrigAnalysisExamples package: TrigDecisionChecker (pre-15.X.0 for the moment): atlas.cgi/offline/Trigger/TrigValidation/TrigValAlgs/TrigValAlgs/TrigDecisionChecker.h?revision=1.2&view=markup atlas.cgi/offline/Trigger/TrigValidation/TrigValAlgs/TrigValAlgs/TrigDecisionChecker.h?revision=1.2&view=markup ARA example: ARA: Trigger Configuration: Run summary pages: (include link to start up the TriggerTool) Trigger information slimming: TriggerMenu working group: Help! Hypernews forum

ARTEMIS - Pisa, June Trigger Tutorial

 Analysis based on single trigger chain or an ‘OR’ of a few chains  Chain definition – algorithms, cuts, multiplicities – do not change during a run, but can change between runs  Important for analysis on DPD, where multiple runs are merged  Prescales at LVL1 or at HLT can change between luminosity blocks  A negative prescale means that this trigger is off. This is important for calculating the integrated luminosity  Analysis based on single trigger chain or an ‘OR’ of a few chains  Chain definition – algorithms, cuts, multiplicities – do not change during a run, but can change between runs  Important for analysis on DPD, where multiple runs are merged  Prescales at LVL1 or at HLT can change between luminosity blocks  A negative prescale means that this trigger is off. This is important for calculating the integrated luminosity ARTEMIS - Pisa, June Trigger Tutorial

ESD AOD TAG Configures Stores decoded Trigger Menu DPD With decreasing amount of detail