This repository contains scripts for generating datasets using WCSim for WatChMaL. (only works for sukap)

Please set your own paths by modifying the following lines:

• e_temp.sh: Line 471
• wcsim_submit_temp.sh: Lines 4 and 5
• root2h5_temp.sh: Lines 4, 5, 10, 11
• setup_env.sh: Line 3

Step 1: Generate root files using WCSim

I wrapped them into a shell script, which you can execute by running:

./sub_wcsim_masspro.sh

to perform mass WCSim data production.

We should modified the following parameters:

• START_EVENT: the starting event index.
• TOTAL_JOBS: the total number of jobs you want to submit.
• MAX_ACTIVE_JOBS: the maximum number of active jobs allowed simultaneously.
• SLEEP_TIME: time interval between checks (in seconds).

The script will submit jobs until the total number (TOTAL_JOBS) is reached. Every SLEEP_TIME seconds, it checks whether the number of currently active jobs is less than MAX_ACTIVE_JOBS. If yes, it automatically submits new jobs up to this limit, continuing this process until all jobs are submitted and completed. Each job generates one corresponding ROOT file.

The script calls submit_WCSim.sh, which copies templates (tempscript/e_500_temp.mac and tempscript/wcsim_submit_temp.sh) and then submits the job. (I'm not sure if there's a better way, but at least this method is straightforward and functional.)

If you need to modify the macro file, edit e_temp.mac.

Step 2: Convert root files into h5 files

I also wrapped them into a shell script,

./root2h5_masspro.sh

It's similar with the sub_wcsim_masspro.sh.

This script calls root2h5_sukap_batch.sh, which copies the template script (tempscript/root2h5_temp.sh) and submits a job using pjsub. Within each submitted job, the scripts FCroot2npz.py and FCnpz2h5.py are executed to convert ROOT files into h5 format.

Step 3: Merge h5 files

Merge individual h5 files into a single npz file: ./merge_h5.py /your/h5/file1 /your/h5/file2 /your/h5/file3 ... -o /your/output/path

Assigning labels to identify their train/validation/test splits. ./gen_split.py (modify Line 12)

Then you can have h5 files and npz files, which can be used for WatChMaL.

Others:

./sub_fitqun_masspro.sh used for submitting jobs to sukap nodes to run fiTQun.

./sub_ana_npz_fitqun_masspro.sh used for extracting the reconstruction result from the fiTQun result root file and write the result into a npz file which is easier for the later analysis.