Neutrino simulation automation

Files to run SND@LHC and AdvSND neutrino simulation (and maybe other simulations later?) in a reliable and reproducible way.

Goals

• Reliable
• Reproducible
• Runnable by non-experts
• Only rerun failed jobs
• Only upload output of successful jobs

How to run

• run_submit.py creates the HTCondor DAG instruction with a lot of options. See the possible keys in the script.
• To set up parameters of the simulation, check simulation_configs/default_target_2024.yaml file.
• To launch simulation, one has to specify $SNDSW_DIR and $CONDOR_FOLDER in system_config.sh file and launch it source system_config.sh
• To generate .gdml geometry file, use python create_geo_config.py --from_config --year 2024. See flags in the file.
• python run_submit.py --yaml simulation_configs/default_target_2024.yaml or manually set info: python3 run_submit.py --tag 2022/sndlhc_1500fb-1_up --topvol volMuFilter --neutrino 14 --year 2022 --flukaflux /eos/experiment/sndlhc/MonteCarlo/FLUKA/neutrino_up_13TeV/all13TeVK0_gsimple.root
• This script will generate a dedicated folder with DAG instructions and logs for HTCondor DAG run. Logs for each subjob is stored in logs/ directory.
• If you want to simulate SND@HL-LHC, then just add --advsnd flag to the run_submit.py: run_submit.py --yaml simulation_configs/advtarget_14_2024.yaml --advsnd.
• Set the steps you want to launch by --flow key: run_submit.py --yaml simulation_configs/your_config.yaml --flow generate_input_file transport digitise. Write step by step the names of the .sub files you would like to launch. To launch the hardcoded pipeline of ["generate_input_file", "transport_neutrinos", "digitise", "analysis"] use --flow all. Default pipeline is ["generate_input_file", "transport_neutrinos", "digitise"].
• You can choose the exact number of events to simulate or the number of pp collisions to simulate. See yaml configs.

Possible inputs

Flux for	Input GENIE flux
snd@lhc 2022 nue-numu vertical down -160 urad	/eos/experiment/sndlhc/MonteCarlo/FLUKA/neutrino_down_13TeV/SND_neutrinos_13TeV_down_19p95M_z481p22m_gsimple.root
snd@lhc 2022 vertical up 160 urad	/eos/experiment/sndlhc/MonteCarlo/FLUKA/neutrino_up_13TeV/all13TeVK0_gsimple.root
snd@lhc 2024 vertical up 160 urad	/eos/user/u/ursovsnd/neutrino_production_sndlhc_june_2025/ALL_lhc_ir1_coll_2024_1p585mm_xrp_exp001_fort.30_FIXED.gsimple.root
advsnd@lhc no_exc nue-numu horizontal 250 urad	/eos/experiment/sndlhc/MonteCarlo/FLUKA/AdvSND/FAR/neutrino/HL-LHC_neutrinos_TI18_20e6pr.gsimple.root
advsnd@lhc no_exc nutau horizontal 250 urad	/eos/user/u/ursovsnd/neutrino_production_sndlhc_june_2025/no_bias_5000m_eta_g_6-0_forward_09042025_1.gsimple.root

Verifying the number of files

• ls -la $OUTPUT_FOLDER$/nu12/volume_volMuFilter/*/sndLHC.Genie-TGeant4_dig.root | wc -l
• Check if the GENIE simulation works: source local_generate_input.py.

Current limitations

• A lot of variables need to be kept in synch across jobs an scripts. How to avoid this?
• DAG needs the entire step to complete. Any way to have Child B of Parent A run on subjobs?