|
|
Getting Started
- To obtain a recorded luminosity measurement, you will need:
- A choice of trigger(s). You cannot default to "any" trigger.
- A list of Bad Runs you plan to exclude from your analysis. This list may be the sum of detector or object quality
bad runs. I have used the Run Quality database to
create a bad run list based on CAL, CFT and SMT detector quality flags. I am also excluding early runs when the L1 CAL
trigger did not have full central calorimeter coverage. See my Bad Runs page.
- A list of Bad LBNs you plan to exclude from your analysis.
- A run range (or default to 1-999999).
- A list of parentage files.
This section was last modified on Sat Feb 5 17:00:24 CST 2005.
|
Sidenote: A luminosity block number is associated to a quantity of recorded data. The LBN increments
every 60 seconds, or when an sclinit is issued, or when a run is started or resumed. Luminosity normalization
is done by trigger per LBN. The online luminosity system performs checks to determine the integrity of each
LBN. If an LBN is marked bad in the luminosity database, you must exclude that LBN if your analysis is dependent
on a luminosity normalization.
Online Cross Checks performed
by Luminosity system during validation of LBNs.
D0 Luminosity Determination including
measurement uncertainty.
|
Pass2
- The following method developed by Marco Verzocchi allows you to 'OR' a combination of unprescaled triggers. I have adapted it
for unprescaled single EM triggers.
- This little script calclumi performs all the
steps outlined below.
- Create a list of all the runs which are included in your favorite common sample group skim.
The following will create a subdirectory called EM1TRK in your current working directory.
- setup D0RunII p16.06.00
- setup sam
- setup python_dcoracle
- /home/mverzocc/scripts/genLBNtables -pass 2 -type RUNS -skim EM1TRK
- Make a copy of the luminosity parentage information into the EM1TRK directory.
- Dump the run quality database. The following will create a file called runQualityDB.dump.
There is a missing line in this dump for run 152237, so add it by hand:
152237 SPECIAL GOOD GOOD REASONABLE UNKNWON UNKNWON UNKNOWN #
- Create a file to be used for parsing the trigger database and deciding which trigger to use
for a given run. For the Z->ee analysis, we use a set of single EM triggers listed in the file
EM1TRK.trigrules. The program will try first to search for runs
in which the triggers E1_SHT20 and E2_SHT20 and E3_SHT20 and E1_SH30 are all unprescaled. If that
fails, the program then goes to the next line in the trigger rules file, so a heirarchy is followed.
- Create two files:
- EM1TRK.prescaleMap -
which contains for all the triggers in the EM1TRK.trigrules the list of prescales for all the runs
- EM1TRK.triggerMap -
which contains for each run the list of triggers which can be used in OR.
- /home/mverzocc/scripts/prescaleParser.py EM1TRK
You may inspect these files, eventually rerun after modifying the trigger or simply edit the
EM1TRK.triggerMap, which is the file used in the subsequent programs.
Note: In your analysis code, you must use the trigger map and select only those events which
pass a specific set of triggers on a run by run basis.
- Get the list of bad luminosity blocks from the
CalJetMet group.
I am using badLBNs.JET.pass2.cert6.0.
- Create a bad run list. See my
Bad Runs web page
where I provide a single text file containing the aggregrate of bad runs from the detector
groups of Calorimeter, CFT and SMT as well as the runs for which the L1 CAL trigger coverage
was limited to |eta|<0.8.
- Create a configuration file which gives the luminosity calcualation program the necessary
directives. I have called it Zee.lumi. Comment lines are embedded
to provide better instruction. There are two offline directories from which one can access
the Stage3 files. Also, one can choose 'recorded' or 'reconstructed' luminosity.
- Run the luminosity calculation program. This will take at least an hour.
- /work/cole-clued0/mverzocc/NEWLUMI/new/lumiCalc Zee.lumi >& Zee.output &
- Review the output from the luminosity calculation program. In addition to the
Zee.output file,
which contains a full dump of the program, of all the luminosities (by run,
by trigger combination and total), it also creates two additional files:
- Zee.run2skip:
a list of all the runs to be skipped in the analysis because they fail the
run quality requirements using the information contained in the run quality database
or because the are contained in the user provided list of bad runs
- Zee.lbn2skip:
a list of all the luminosity block numbers to be excluded from the analysis either
because there is no luminosity information available for those LBNs or because
they are contained in a user provided list of bad luminosity blocks.
Note: In your analysis code, you must use both lists and reject any event with a run
or luminosity block number which can be found in either list.
This section was last modified on Wed Oct 12 01:03:52 CDT 2005.
|
Luminosity calculation done by Gordon Watts for
the top group.
|
Pass1
- The same method as Pass2 is used, however, the list of parentage files is different, and the
bad LBN list was generated without using the t42 correction.
- You need the following files:
This section was last modified on Tue Nov 30 09:16:51 CST 2004.
|
|
|