#include <DuplicateChecker.h>
|
| virtual StatusCode | setupJob (Job &job) override |
| | effects: give the algorithm a chance to intialize the job with anything this algorithm needs. More...
|
| |
| virtual StatusCode | changeInput (bool firstFile) override |
| | effects: do all changes to work with a new input file, e.g. More...
|
| |
| virtual StatusCode | initialize () override |
| | effects: do everything that needs to be done before running the algorithm, e.g. More...
|
| |
| virtual StatusCode | execute () override |
| | effects: process the next event guarantee: basic failures: algorithm dependent More...
|
| |
| void | read_run_event_number () |
| | get the run and event number for the current event More...
|
| |
| | ClassDef (DuplicateChecker, 1) |
| |
| virtual StatusCode | fileExecute () |
| | effects: do all the processing that needs to be done once per file More...
|
| |
| virtual StatusCode | endOfFile () |
| | effects: do the post-processing for each input file guarantee: basic failures: algorithm dependent rationale: this is mainly used for specialized services that need to save partial results for each input file More...
|
| |
| virtual StatusCode | histInitialize () |
| | effects: this is a pre-initialization routine that is called before changeInput is called. More...
|
| |
| virtual StatusCode | postExecute () |
| | effects: do the post-processing for the event guarantee: basic failures: algorithm dependent rationale: this is mainly used for specialized services that need to get input from subsequent algorithms before filling their event data More...
|
| |
| virtual StatusCode | finalize () |
| | effects: do everything that needs to be done after completing work on this worker guarantee: basic failures: algorithm dependent rationale: currently there is no use foreseen, but this routine is provided regardless More...
|
| |
| virtual StatusCode | histFinalize () |
| | effects: this is a post-initialization routine that is called after finalize has been called. More...
|
| |
| virtual bool | hasName (const std::string &name) const |
| | returns: whether this algorithm has the given name guarantee: basic failures: algorithm dependent rationale: this is to allow an algorithm to be known by multiple names. More...
|
| |
| void | sysSetupJob (Job &job) |
| | effects: give the algorithm a chance to intialize the job with anything this algorithm needs. More...
|
| |
|
| std::string | m_eventInfoName |
| | the value returned by eventInfoName More...
|
| |
| std::string | m_outputTreeName |
| | the value returned by outputTreeName More...
|
| |
| std::map< std::pair< std::string, std::string >, std::map< Long64_t, std::pair< number_type, number_type > > > | m_duplicates |
| | the list of known duplicates to skip More...
|
| |
| std::map< Long64_t, std::pair< number_type, number_type > > * | m_currentDuplicates = nullptr |
| | the list of the duplicates in the current file to skip, or the null pointer if there are none More...
|
| |
| std::set< std::pair< number_type, number_type > > | m_processed |
| | the list of run-event numbers already encountered More...
|
| |
| xAOD::TEvent * | m_event = nullptr |
| | the event we are reading from More...
|
| |
| TTree * | m_outputTree = nullptr |
| | the output tree, if we are creating one More...
|
| |
| std::string | m_inputFileName |
| | the name of the input file (connected to m_outputTree, if present) More...
|
| |
| Long64_t | m_inputFileIndex |
| | the index in the input file (connected to m_outputTree, if present) More...
|
| |
| number_type | m_runNumber |
| | the run number of the current event (connected to m_outputTree, if present) More...
|
| |
| number_type | m_eventNumber |
| | the event number of the current event (connected to m_outputTree, if present) More...
|
| |
| Bool_t | m_processEvent |
| | whether the current event is/should be processed (connected to m_outputTree, if present) More...
|
| |
| IWorker * | m_wk |
| |
| asg::SgTEvent * | m_evtStorePtr = nullptr |
| | the value of evtStore More...
|
| |
| asg::SgTEvent | m_evtStore |
| | when configured, the object returned by evtStore More...
|
| |
| MsgStream * | m_msg = nullptr |
| | the message stream, if it has been instantiated More...
|
| |
| std::string | m_msgName |
| | the algorithm name for which the message stream has been instantiated More...
|
| |
| int | m_msgLevel = 3 |
| | the message level configured More...
|
| |
| std::string | m_nameCache |
| | the cache for name More...
|
| |
- Todo:
- add documentation
Definition at line 30 of file DuplicateChecker.h.
◆ number_type
◆ DuplicateChecker()
| EL::DuplicateChecker::DuplicateChecker |
( |
| ) |
|
standard constructor
- Guarantee
- strong
- Failures
- out of memory I
◆ addKnownDuplicate()
| void EL::DuplicateChecker::addKnownDuplicate |
( |
const std::string & |
sampleName, |
|
|
const std::string & |
fileName, |
|
|
Long64_t |
entry, |
|
|
number_type |
runNumber, |
|
|
number_type |
eventNumber |
|
) |
| |
add a known duplicate event
- Guarantee
- strong
- Failures
- out of memory II
◆ addKnownDuplicatesFile()
| void EL::DuplicateChecker::addKnownDuplicatesFile |
( |
const std::string & |
duplicatesFile | ) |
|
add a file with known duplicates
- Guarantee
- strong
- Failures
- i/o errors
out of memory III
◆ book()
| void EL::Algorithm::book |
( |
const TH1 & |
hist | ) |
|
|
inherited |
book the given histogram
- Guarantee
- strong
- Failures
- histogram booking error
◆ changeInput()
| virtual StatusCode EL::DuplicateChecker::changeInput |
( |
bool |
firstFile | ) |
|
|
overrideprivatevirtual |
effects: do all changes to work with a new input file, e.g.
set new branch addresses. if firstFile is set, this method is called just before init() is called
Warning: If a file is split across multiple jobs this will be called more than once. This only happens for specific batch drivers and/or if it is explicitly configured by the user. With PROOF it could even happen multiple times within the same job, and while PROOF is no longer supported that behavior may come back if support for a similar framework is added in the future. As such, this method should not be used for accounting that relies to be called exactly once per file, take a look at fileExecute() if you want something that is guaranteed to be executed exactly once per input file.
Warning: The execution order of changeInput and fileExecute is currently unspecified.
guarantee: basic failures: algorithm dependent
Reimplemented from EL::Algorithm.
◆ ClassDef()
◆ endOfFile()
effects: do the post-processing for each input file guarantee: basic failures: algorithm dependent rationale: this is mainly used for specialized services that need to save partial results for each input file
Reimplemented in EL::MetricsSvc.
◆ eventInfoName()
| const std::string& EL::DuplicateChecker::eventInfoName |
( |
| ) |
const |
the name of the EventInfo structure to use
This is mostly meant, so that in my unit test code I can point it to my own specially prepared EventInfo
- Guarantee
- no-fail
◆ evtStore()
get the (main) event store for this algorithm
This is mostly to mirror the method of the same name in AthAlgorithm, allowing to make the tutorial instructions more dual-use.
- Guarantee
- strong
- Failures
- out of memory I
job not configured for xAODs
◆ execute()
| virtual StatusCode EL::DuplicateChecker::execute |
( |
| ) |
|
|
overrideprivatevirtual |
effects: process the next event guarantee: basic failures: algorithm dependent
Reimplemented from EL::Algorithm.
◆ fileExecute()
effects: do all the processing that needs to be done once per file
Warning: The user should not expect this to be called at any particular point in execution. If a file is split between multiple jobs this will be called in only one of these jobs, and not the others. It usually gets called before the first event in a file, but that is not guaranteed and relying on this is a bug. Take a look at changeInput if you want something that is guaranteed to be executed at the beginning of each input file.
Warning: The execution order of changeInput and fileExecute is currently unspecified.
guarantee: basic failures: algorithm dependent rationale: this is to read per-file accounting data, e.g. the list of lumi-blocks processed
Reimplemented in EL::UnitTestAlg1, EL::UnitTestAlg, EL::UnitTestAlgXAOD, and EL::MetricsSvc.
◆ finalize()
effects: do everything that needs to be done after completing work on this worker guarantee: basic failures: algorithm dependent rationale: currently there is no use foreseen, but this routine is provided regardless
Reimplemented in EL::UnitTestAlg1, EL::UnitTestAlg, and EL::UnitTestAlgXAOD.
◆ hasName()
| virtual bool EL::Algorithm::hasName |
( |
const std::string & |
name | ) |
const |
|
privatevirtualinherited |
returns: whether this algorithm has the given name guarantee: basic failures: algorithm dependent rationale: this is to allow an algorithm to be known by multiple names.
this is needed for NTupleSvc, so that it can be located with and without the output tree name.
Reimplemented in EL::NTupleSvc.
◆ hist()
| TH1* EL::Algorithm::hist |
( |
const std::string & |
name | ) |
const |
|
inherited |
get the histogram with the given name
- Guarantee
- strong
- Failures
- histogram not found
◆ histFinalize()
| virtual StatusCode EL::Algorithm::histFinalize |
( |
| ) |
|
|
privatevirtualinherited |
◆ histInitialize()
| virtual StatusCode EL::Algorithm::histInitialize |
( |
| ) |
|
|
privatevirtualinherited |
◆ initialize()
| virtual StatusCode EL::DuplicateChecker::initialize |
( |
| ) |
|
|
overrideprivatevirtual |
effects: do everything that needs to be done before running the algorithm, e.g.
create output n-tuples and histograms. this method is called only once right after changeInput(true) is called guarantee: basic failures: algorithm dependent rationale: in principle all this work could be done on changeInput(true). However, providing this method should make it easier for the user to set up all his outputs and to do so only once.
Reimplemented from EL::Algorithm.
◆ msg() [1/2]
| MsgStream& EL::Algorithm::msg |
( |
| ) |
const |
|
inherited |
messaging interface
this is the interface to work with the standard messaging macros from AsgTools. Instead of enums I pass ints, so that I can avoid the include dependency (forward declarations are only allowed for enum classes AFAIK).
the standard message stream for this object
- Guarantee
- strong
- Failures
- code not compiled with AsgTools support
◆ msg() [2/2]
| MsgStream& EL::Algorithm::msg |
( |
int |
level | ) |
const |
|
inherited |
the message stream for this object, configured for the given level
- Guarantee
- strong
- Failures
- code not compiled with AsgTools support
◆ msgLvl()
| bool EL::Algorithm::msgLvl |
( |
int |
lvl | ) |
const |
|
inherited |
whether we are configured to print messages at the given level
- Guarantee
- no-fail
◆ name()
| virtual const std::string& EL::Algorithm::name |
( |
| ) |
const |
|
virtualinherited |
◆ outputTreeName()
| const std::string& EL::DuplicateChecker::outputTreeName |
( |
| ) |
const |
the name of the output tree to create, or the empty string if none is created
The output tree contains a list of run and event numbers for all events, and whether they were processed by this job. This can be used to check whether duplicate events were processed (or whether we somehow eliminated events as duplicates that we shouldn't have). It can also be used to create a list of duplicate events for future processing rounds.
◆ postExecute()
effects: do the post-processing for the event guarantee: basic failures: algorithm dependent rationale: this is mainly used for specialized services that need to get input from subsequent algorithms before filling their event data
Reimplemented in EL::NTupleSvc.
◆ processSummary() [1/2]
process the summary tree from the given submission
This will create a file "duplicates" inside the submission directory that contains the list of duplicates that can be fed into future submissions to filter them out.
- Returns
- whether the job was successfully, i.e. each input event was read exactly once and all duplicates were skipped
- Guarantee
- basic
- Failures
- i/o errors
This version of the method provides a lower level interface, in which the list of inputs is given via a sample handler (with the tree name properly set), and the output file name freely choosable.
◆ processSummary() [2/2]
| static bool EL::DuplicateChecker::processSummary |
( |
const std::string & |
submitdir, |
|
|
const std::string & |
treeName |
|
) |
| |
|
static |
process the summary tree from the given submission
This will create a file "duplicates" inside the submission directory that contains the list of duplicates that can be fed into future submissions to filter them out.
- Returns
- whether the job was successfully, i.e. each input event was read exactly once and all duplicates were skipped
- Guarantee
- basic
- Failures
- i/o errors
◆ read_run_event_number()
| void EL::DuplicateChecker::read_run_event_number |
( |
| ) |
|
|
private |
get the run and event number for the current event
◆ setEventInfoName()
| void EL::DuplicateChecker::setEventInfoName |
( |
const std::string & |
val_eventInfoName | ) |
|
set the value of eventInfoName
- Guarantee
- strong
- Failures
- out of memory II
◆ setMsgLevel()
| void EL::Algorithm::setMsgLevel |
( |
int |
level | ) |
|
|
inherited |
set the message level for the message stream for this object
- Guarantee
- no-fail
◆ setOutputTreeName()
| void EL::DuplicateChecker::setOutputTreeName |
( |
const std::string & |
val_outputTreeName | ) |
|
set the value of outputTreeName
- Guarantee
- strong
- Failures
- out of memory II
◆ setupJob()
effects: give the algorithm a chance to intialize the job with anything this algorithm needs.
this method is automatically called before the algorithm is actually added to the job. guarantee: basic failures: algorithm dependent rationale: this is currently used to give algorithms a chance to register their output datasets, but can also be used for other stuff.
Reimplemented from EL::Algorithm.
◆ sysSetupJob()
| void EL::Algorithm::sysSetupJob |
( |
Job & |
job | ) |
|
|
privateinherited |
effects: give the algorithm a chance to intialize the job with anything this algorithm needs.
this method is automatically called before the algorithm is actually added to the job. guarantee: basic failures: algorithm dependent rationale: this is currently used to give algorithms a chance to register their output datasets, but can also be used for other stuff.
◆ testInvariant()
| void EL::DuplicateChecker::testInvariant |
( |
| ) |
const |
test the invariant of this object
- Guarantee
- no-fail
◆ wk()
| IWorker* EL::Algorithm::wk |
( |
| ) |
const |
|
inherited |
description: the worker that is controlling us guarantee: no-fail
◆ m_currentDuplicates
the list of the duplicates in the current file to skip, or the null pointer if there are none
Definition at line 183 of file DuplicateChecker.h.
◆ m_duplicates
| std::map<std::pair<std::string,std::string>,std::map<Long64_t,std::pair<number_type,number_type> > > EL::DuplicateChecker::m_duplicates |
|
private |
◆ m_event
◆ m_eventInfoName
| std::string EL::DuplicateChecker::m_eventInfoName |
|
private |
◆ m_eventNumber
◆ m_evtStore
◆ m_evtStorePtr
◆ m_inputFileIndex
| Long64_t EL::DuplicateChecker::m_inputFileIndex |
|
private |
◆ m_inputFileName
| std::string EL::DuplicateChecker::m_inputFileName |
|
private |
◆ m_msg
| MsgStream* EL::Algorithm::m_msg = nullptr |
|
mutableprivateinherited |
the message stream, if it has been instantiated
Definition at line 333 of file Algorithm.h.
◆ m_msgLevel
| int EL::Algorithm::m_msgLevel = 3 |
|
privateinherited |
the message level configured
Definition at line 342 of file Algorithm.h.
◆ m_msgName
| std::string EL::Algorithm::m_msgName |
|
mutableprivateinherited |
the algorithm name for which the message stream has been instantiated
Definition at line 338 of file Algorithm.h.
◆ m_nameCache
| std::string EL::Algorithm::m_nameCache |
|
mutableprivateinherited |
◆ m_outputTree
| TTree* EL::DuplicateChecker::m_outputTree = nullptr |
|
private |
◆ m_outputTreeName
| std::string EL::DuplicateChecker::m_outputTreeName |
|
private |
◆ m_processed
◆ m_processEvent
| Bool_t EL::DuplicateChecker::m_processEvent |
|
private |
◆ m_runNumber
◆ m_wk
The documentation for this class was generated from the following file:
1.8.18