laphecet [Sun, 30 Mar 2008 22:06:05 +0000 (22:06 +0000)]
Re-establishing creation of RecPoints in TreeR (Laurent)
Introducing tracking mode starting from trigger tracks (Laurent)
New options for tracking and some bugs fixed (Philippe P.)
A cluster server which is doing no more than serving clusters from an existing cluster store. Used to re-establish clustering in local reconstruction (see AliMUONReconstructor's changes)
- AliMUONTriggerTrackToTrackerClusters (Laurent)
A class to replace clusters in stations 4 and 5 (that are discarded (if they exist)) by "fake" ones, created from the extrapolation of trigger tracks to st45.
Created to allow reconstruction of data taken during Feb. 08 cosmic run where st45 were not ready yet.
Re-establishing, in the case of non-combined tracking, the filling of the TreeR with the clusters, i.e. clustering is done in local reconstruction in that case, and
not in tracking phase.
Changing ctor parameter from ref to pointer, as we're taking ownership of that
object, and adding a method to find out in which DE a global (x,y) position is.
AliMUONVClusterServer.h (Laurent)
Changing comment which did not follow interface change...
AliMUONTracker (Laurent)
If no cluster server given, use legacy one (see also changes in AliMUONReconstructor)
AliMUONRecoParam (Laurent & Philippe P.)
New options
AliMUONRefitter (Laurent)
Avoids code duplication and use new SimpleClusterServer interface.
AliMUON*Tracker* (Philippe P.)
- solved the problem of divergence in the tracking of low momentum muons
- added a reconstruction parameter defining the maximum non bending slope of the primary track candidates
- added the possibility to disable one or more chambers in the reconstruction (new option)
- added the possibility to reconstruct a track even if it has no cluster in one or more stations (new option)
- added the possibility to make a primary track candidate starting from one cluster on station 4 and one on station 5 (new option)
morsch [Sat, 29 Mar 2008 20:30:15 +0000 (20:30 +0000)]
Possibility to add friend trees.
Here only one additional tree can be added. This is for testing.
In the general case it will be possible to add any number of friends.
richterm [Thu, 27 Mar 2008 19:24:34 +0000 (19:24 +0000)]
bugfixes by Kenneth:
TPCClusterFinder: consider row offset correctly for stored clusters
(rows had been numbered starting from 0 on all the patches)
PCMapping: correct initialization of array bounds in cause of multiple instances.
richterm [Wed, 26 Mar 2008 22:06:10 +0000 (22:06 +0000)]
- handling of global AliHLTSystem singleton moved to BASE
- AliHLTSimulation uses AliHLTSystem singleton
- HLTOUT raw files written according to AliSimulation::SetWriteRawData
cvetan [Wed, 26 Mar 2008 14:11:44 +0000 (14:11 +0000)]
Major modifications to AliReconstruction that are needed in order to adapt it for online usage. The main idea was to extract the event loop from the run method and allow the user to add events (in memory via AliRawReaderDate) on the fly. More details will be given in a talk during the forthcoming offline week (Apr 2008)
cvetan [Wed, 26 Mar 2008 12:41:54 +0000 (12:41 +0000)]
Changes needed by the online event reconstruction: AliRawReaderDate's NextEvent() and RewindEvents() methods are now fully compatible regardless how the raw-reader has been initialized - via File or event in the memory.
hdalsgaa [Wed, 26 Mar 2008 11:49:11 +0000 (11:49 +0000)]
This is the base class for the FMD detector algorithms. This class does the works on the data and the derived classes provide the tool. The detector algorithms are part of libFMDutil .