TriggerTisTos tool
Some parts of documentation below are not up to date.
Please see
presentation at Trigger & Stripping meeting on Nov 1, 2010
for the most recent report on TisTos software. If you are
using Stripping MicroDSTs please also read
StrippingMDSTsTisTos page.
TriggerTisTos tool classifies triggers (L0 or Hlt) as TIS (Trigger Independent of Signal) or TOS (Trigger On Signal)
with respect to user defined signal (e.g. B candidate).
It can also be used to access summaries of candidates selected by the Trigger algorithms and saved in Raw Data.
There is also
TESTisTos tool which can be used to classify selections saved in TES, stripping decision in particular.
If you are using TupleToolTISTOS for your Tis/Tos info, you may want to read
How Do I configure TupleToolTrigger and TupleToolTISTOS? page.
TupleToolTISTOS in
DecayTreeTuple is an additional layer of software which calls
TriggerTisTos tool described here.
If you are interested in Tis/Tossing L0 trigger, please refer to
L0TriggerTisTos twiki.
There in now
TisTosParticleTagger which can be used to set-up filtering of Particles with respect to TisTos criteria, or to filter Events.
Please refer to separate page for
documentation.
Documentation
- The user interface of the tool is defined in ITriggerTisTos.h which contains also some documentation. See Doxygen documentation of the DaVinci project
. Make sure to follow link also to ITriggerSelectionTisTos from which ITriggerTisTos inherits.
- More extensive documentation with code examples can be found in various presentations (use the first one listed for the most recent one, though even this version is a bit obsolete):
- Update on TisTos software, Nov 1, 2010, Trigger & Stripping meeting
- Hlt Raw Data & TISTOS tool (C++ & Python) June 16, 2009, LHCb SoftwareWeek
.
- Hlt Raw Data & TISTOS tool, Mar 19, 2009, LHCb Software Week
.
- Hlt Data and Usage, Dec. 11, 2008, LHCb Software Week
.
- Trigger Info and TisTosTobbing Tutorial, Sep. 16, 2008, LHCb Software Week
.
- Python version of TisTosTobbing Tutorial, June 18, 2008, LHCb Software Week
.
- "Trigger and TisTos Tutorial", Oct.2, 2007, LHCb Software Week
- "TisTosTobbing", Sept.13, 2007, Plenary III at LHCb Week at St Nectaire
- "Status of TisTosTobbing", Sep.5, 2007, Tagging, propertime and mixing WG meeting
- "TisTosTobbing without tampering", T-Rec, December 2005
Links to related pages
Comments to various DaVinci releases
New set-up is needed for Stripping 17 MDSTs - see
StrippingMDSTsTisTos page.
- DaVinci v28r2p2 and similar
New 2011
DiMuon trigger is not yet supported in these releases. "getpack Phys/TisTosTobbing r122889" to make it work.
No known problems.
Contains support for TisTossing schemes on MDSTs.
Slight change in L0 TisTos of Muon triggers (use Trig/L0/MuonBCSU instead of Trig/L0/MuonCtrl).
Fixes the bugs outlined below.
However, it does contain bug for L0 trigger - see
L0TriggerTisTos twiki.
Results for "Hlt1.*Decision" wrong for most of 2010 data because of Hlt1L0.*Muon.*Decision (applies to all previous versions too).
See slide 19 in
presentation at Trigger & Stripping meeting on Nov 1, 2010
.
If you re-run L0+HLT triggers (using 1Track Hlt1 scheme) in
DaVinci you must add:
from Configurables import L0Conf
L0Conf().FullL0MuonDecoding = True
L0Conf().EnableL0DecodingOnDemand = True
otherwise your results for muons will be inaccurate and for hadrons completely wrong.
Major rewrite of the code (Phys/TisTosTobbing v5r0; Phys/DaVinciInterfaces v2r5).
More information will be provided. matchedTOSxxx methods were retired from the interface. Muon hit TOS threshold lowered from 0.7 to 0.6.
A lots of new methods to provide new functionality (the old methods will still work as well). See also
TisTosParticleTagger.
No known problems. This version includes
L0TriggerTisTos tool.
No known problems.
However, this and all earlier versions do not let you Tis/Tos L0 trigger to good accuracy (particularly L0DiMuon is affected).
Please refer to
L0TriggerTisTos info.
No known problems.
To work on slc5 platform you need to getpack Phys/TisTosTobbing v3r6 or higher (get a head version if this tag is not available yet) otherwise jobs can sometimes crash on the first event processed by the tool. slc4 platform appears to be fine.
There was a change in the tool catalog of known trigger names. In the previous version all names registered with HltANNSvc were considered valid trigger names, which in practice meant that there were a lot of spurious triggers in Hlt1 part. Now the trigger names are restricted to those appearing in HltDecReports which ensures that only executed triggers are known to the tool. Intermediate trigger selection names are restricted to those saved in HltSelReports. A consequence of these changes is that the list of known trigger names can vary from MC event to MC event (e.g. if L0 failed and Hlt was not excuted, there will be no known trigger names). Real data event all passed Hlt, thus the trigger list will not vary for the same
TCK, however list of intermediate selection names can still vary from event to event depending which ones were saved in HltSelReports. The tool no longer complains if the user asked for a trigger selection which is not found (the warning can be reactivated via a property of the tool); the tool will silently return default values (decision=false; tis=false; tos=false).
Do not use these version. There are problems in Calo code which cause crashes sometimes in Tis/Tos software.
IMPORTANT! From this version now on, wild character matching in the trigger specifications passed to the tool is handled by regex (see e.g.
Regular Expression Tutorial
). This provides for more functionality. Users must change their code to replace any occurrence of '*' by '.*' (for example 'Hlt1*Decision' must be replaced by 'Hlt1.*Decision').
- DaVinci v22r1, v22r2, v22r3, v23r0p, v23r1, v24r1
No known problems.
Options to run trigger simulations on DC06 data:
from Gaudi.Configuration import *
from Configurables import DaVinci
DaVinci().ReplaceL0BanksWithEmulated = True ## enable if you want to rerun L0
DaVinci().HltType = 'Hlt1+Hlt2' ## pick one of 'Hlt1', 'Hlt2', or 'Hlt1+Hlt2'
--
TomaszSkwarnicki - 11 Oct 2007, last update 1 Nov 2011