How to use the Hall A Gen Library for Data Anlysis...

Bodo Reitz
Revision 0.0, September 2, 05


In the following you will find some hints of how to use 
the Hall A analyzer for analysis of data taken for the Hall A Gen
experiment. I will concentrate on the example of the test data 
taken end of January 2005 (end of the E01015 SRC experiment).

To learn more about the goals of this specific test, you can look
into the beam time request for beam development time:
http://hallaweb.jlab.org/experiment/E02-013/expdocs/proposal_for_n20.txt
A description of the setup is included in 
http://hallaweb.jlab.org/experiment/E02-013/expdocs/N20_AdcCal_Usersguide.ps

All analysis software for this project is based on Ole Hansens 
 Hall A root analyzer, for which we added a library dedicated to non standard 
equipment used during E02013 (e.g. drift chambers for BigBite, Shower for 
BigBite, HAND ( = Hall A neutron detector ). It is still
in a developement state, so dont be surprised about frequent changes.
It is strongly recommended to use the CVS version of both the analyzer
and the AgenLib package, and frequently update both.

Instructions of how to obtain the software can be found at
http://hallaweb.jlab.org/experiment/E02-013/expdocs/agen_install_howto.txt

Besides the successfully compiled analysis software (consiting of 
root (preinstalled on most Jlab machines), the hall A root analyzer,
and the AgenLib package) you will need some more files (here again
for the N20 tests, ask the experts for similar tables for other tests):

1) list of runs, including run numbers and running conditions:

   /work/halla/e02013/disk1/runinfo/*

2) raw data files:

   the are on the mass storage system, together with the raw data from 
   the SRC experiment:

   /mss/halla/e01015/raw/e01015_####.dat.X

   you can copy them onto a cache disk, using the jcache command
   jcache /mss/halla/e01015/raw/e01015_####.dat.X
   and you will find the file after some time and only for a limited period at
   /cache/ /mss/halla/e01015/raw/e01015_####.dat.X

   or you can copy them with jget onto a local disk.

   You might also be lucky and find a copy on the work disk in
   /work/halla/e02013/disk1/raw
   or 
   /work/halla/e02013/disk1/raw

   (If you need them frequently, it is fine to copy a few files there,
    but please dont copy them additionally into your private subdirectory
    on the workdisk) 
   
3) Database files

   The root analyzer and its subclasses need a lot of additional 
   information, which is not contained in the raw data files, such 
   as running conditions, detector setups, calibration data .....
   All this information is located in a "database" a huge collection 
   of - at present - ascii text files. 
   These files have a tight naming scheme: db_APPARATUS.DETECTOR.dat
   For each detector which is used in the analysis such a file is needed,
   the software first looks in your local directory (from where you 
   started the analyzer) if such a file exists. If not it searches 
   in the directories defined by the environment variable DB_DIR and 
   the time the raw data was taken. E.g. the N20 data was taken on 
   February 2, if beyond DB_DIR the following three directories exist:
   > echo $DB_DIR
   /work/halla/e02013/disk1/db
   > ls $DB_DIR
   20050123  20050710  20050823
   Then the analyzer will use the files in 20050123, whereas if the 
   data was taken on July 30, the files in 20050710 will be used.

   For all gen tests, a working set of database files should be in 
   /work/halla/e02013/disk1/db
   If you want to optimize a database file, it is recommended to 
   first put it into your local directory, and only after your 
   optimization is finished, you should replace it in the public 
   database directory. Changes there should be logged in the 
   AgenDb_log file. Dont forget to remove experimental database files 
   in your local directory, once they are not needed anymore.

4) Steering Scripts

   The programming language as well as the macro language for the 
   root and our derived software is C/C++. Therefore all scripts are 
   written in C. Usually the analysis is separated into (at least) two
   passes. The first pass, the so called replay, requires the analyzer,
   and reads in the raw data file, and produces a "rootfile", which 
   can contain either histograms of variables, and/or variables can 
   be put in a tree. An example for such a script can be found 
   in /work/halla/e02013/disk1/replay/replay.C
   These scripts usually require two additional input files, one 
   which defines variable to be put into the root file
   (output_N.def) and one which defines cuts (cutdef_N.dat).
   In the second pass various scripts can be used to actually 
   look at the data in a rootfile (look e.g. Adc_Cal.C )

   Further information about these scripts can be found in 
   Hall A C++ Analyzer Documentation
         http://hallaweb.jlab.org/root/doc/index.html
   ROOT Documentation		    
         ftp://root.cern.ch/root/doc/Users_Guide_4_08.pdf


To be continued ...