- using AliVertexerTracks initialization provided by Andrea
- reading cuts from AliGRPRecoParam
- do not require kITSrefit flag
- using mode ITS
- publishing the AliESDVertex object
- updated the corresponding macro
- extended chain with GlobalVertexerHisto and connected ROOTFileWriter
- two parallel chains with GlobalVertexer and GlobalOfflineVertexer
- new running mode to analyse ESDs directly, either esdTree or HLTesdTree
- fill the number of cluster per track only for TPC or TPC+ITS tracks (0 entries rejected) (Hege Erdal)
- fill the eta histogram with a DCA cut (<5cm) for selecting primary tracks and a pt cut (>300MeV/c) to get rid of looping particles
- reject laser tracks in all histograms
- included histo with the status flag of every track
initialization of magnetic field, setting the kITSrefit flag for the input tracks as a first attempt
more detailed studies underway
the component now also compiles with the v4-18-Release where the ESD parameter to the AliVertexerTracks was not const
ivana [Mon, 19 Apr 2010 11:09:10 +0000 (11:09 +0000)]
In AliMUONReconstructor:
Putting back unloading mapping from ctor, but now only if the mapping objects
exist. (This removes the warning from AliCDBManager with new OCDB where the
old objects are removed and keeps the code working with old OCDBs where the old
objects are present.)
adding component 'GlobalOfflineVertexer' (class AliHLTGlobalOfflineVertexerComponent) to the libAliHLTGlobal
adding test macro to run the component on raw data, subscribing to the default HLT reconstruction chain
adding new interface function for the announcement of required OCDB objects and
implemented a check for the objects in the component initialization
/**
* Get a list of OCDB object description.
* The list of objects is provided in a TMap
* - key: complete OCDB path, e.g. GRP/GRP/Data
* - value: short description why the object is needed
* Key and value objects created inside this class go into ownership of
* target TMap.
* @param targetMap TMap instance receiving the list
* @return void
*/
virtual void GetOCDBObjectDescription( TMap* const targetArray);
- applying a workaround for reading the ITS TreeD which was supposed to be sent via TMessage. This seems however not possible. More investigation necessary.
- bugfix: Setting the directory of the target tree to NULL solves seg fault when reading it
1.) runFlowTask.C:
a) Included by default some other methods besides SP;
b) Added Bool_t DATA = kFALSE/kTRUE to the macro's argument when running in mode mGrid as well;
c) Added method CrossCheckUserSettings() which for instance prevents setting MCEP = DATA = kTRUE. Implementation of other warnings (e.g. for LYZ1SUM = LYZ2SUM = kTRUE shall be implemented within this method in order not to spoil to much the main body of the macro with trivial things);
2.) CreateAlienHandler.C:
a) When running in run mode "test" by default 10 files will be copied locally and analysis will be performed only on them, which can take more than 1h in reality. Added setter for number of files to be copied locally: plugin->SetNtestFiles(2);
b) Updated Root and AliRoot versions: plugin->SetROOTVersion("v5-26-00b-2"); plugin->SetAliROOTVersion("v4-19-10-AN");
c) There is a method to add more runs of a certain production in one go: plugin->SetRunRange(105000,106000);
d) Instead of storing outputs in directories 000,001, etc, the output files will be stored in directories named after the run number of a certain production: plugin->SetOutputToRunNo();
e) Provided example of plugin usage for one MC production and for real data on Grid.
3.) In *Task*.cxx:
a) Added in accordance with the 3rd point at https://aliceinfo.cern.ch/Offline/Activities/Analysis/AnalysisFramework/analysisFAQ.html the line PostData(1,fListHistos);
at the end of each method UserCreateOutputObjects() in the task's classes. Now even if no event passes the selection criteria the output files of each method still will be in "AnalysisResults.root" but with the empty histograms. This is extremely relevant for merging procedure and for instance for validating jobs in plugin's "test" mode.
re-organizing the configurations for globel HLT components:
- GlobalVertexer gets input from TPC merger and ITS tracking
- GlobalESDConverter gets input from TPC merger, TPC MC, Global Vertexer and ITS tracking
- Vertex histogram componenent gets input from Global Vertexer and SPD vertexer
all inputs are optional
Adding support for ITS reconstruction from digits. The DigitPublisher configuration is added, however not yet enabled in the ITS tracking due to a problem in reading the TTree input which needs to be solved first.
Reducing memory profile of HLT ITS clusterfinding by using only one CF per detector instead of one per DDL.
bugfix: removing calls to old AliGRPObject functions, check of libAliHLTPHOS temporarily disabled due to various problems in the library - cleanup ongoing
correcting unresolved symbols of libAliHLTTrigger:
classes AliHLTTriggerCaloClusterEnergy and AliHLTTriggerEmcalClusterEnergy added
to build system and LinkDef
- Init exchange containers independently needed for usage of the task
outside the analysis train
- comment out tracklet exchanged containers for the moment