Plot SSNet

SSNet analysis scripts

WARNING: these are very quick and dirty.

Instructions

Building

First, setup your copy of LArCV (or dllee_unified) by running their respective configure scripts.

Then return this folder and run 'make'.

In principle it should build.

Running

These scripts were designed to be run on the Tufts grid. We use the Tufts PUBs database to provide us with a set of input files to run. We process each file one-by-one, but process many at once using the Tufts grid.

The following are the steps to go from filelist (from PUBs), to output TROOT file.

Step 1: Get the filelist

• First, remember to setup the environment for the PUBs production database. (source setup_dlleepubs_procdb.sh in the Tufts PUBs folder.)

• Next, use get_filelist.sh to generate a list of files that have been processed successfully by serverssnetv2.

• Go into get_filelist.sh and set the name of the file table (e.g. mcc9tag2_nueintrinsic_corsika).

• source get_filelist.sh: as a result you should see a list of files in filelist/${table}_ssnet.list (and one for supera files.)

Step 2: compile the script

• For the Tufts grid, we provide script to submit a job for compiling the code. It basically loads a singularity container containing a copy of dllee_unified, sets it up, then runs make

  source compile.sh
  

Step 3: launch the jobs

• Go into submit.sh and change the TAG variable to reflect the file table name used for the sample.
• Still inside submit.sh, adjust the number of jobs (and which subset of the files) to run using the --array flasg
• LAUNCH sbatch submit.sh

Step 4: hadd

• The files automatically stored in output/${tagname}/.
• To merge them into one file for convience using hadd.sh.
• Go to hadd.sh and change the table name (tag name)
• source hadd.sh (if the number of files are not crazy large)

To do for scripts

(to do)