MC_data_new.py

This simple script applies both the smearing due to diffusion and the electronic noise background to a MC sample track (GEANT4 output). To run, the script need to be in the same location where ConfigFile_new.txt is.

The output file is a .root file containing all the TH2F histograms generated.

INSTALLATION

Install python dependencies using:

pip install -r requirements.txt

In addition, you will need to have installed ROOT with python bindings

USAGE

The usage of the script is the following:

python MC_data_new.py ConfigFile_new.txt [options]

With the following options available:

 Identifier               Option                         Default value
 
     -I           `<path_to_inputfolder>`         `<path_to_current_directory>+src/`
     -O           `<path_to_outputfolder>`        `<path_to_current_directory>+out/`
     

Given the input folder, the script will run over all the .root files it can find in that location. The output file contains also a TDirectoryFile used as a storage for values imported from the ConfigFile_new.txt. They are stored in single-binned histogram, so you can easily access it using

param_dir->cd()
'histogram_name'->GetBinContent(1)

Also the type of particle, and its initial energy, for each event are stored in a subfolder. You can access it using

event_info->cd()
'histogram_name'->GetBinContent(1)

EXAMPLE

Here an example is provided.

• First of all, download the repository with git clone [email protected]:CYGNUS-RD/digitization.git
• You want to specify the folder in which your GEANT4 simulation output is. If you don't have any MC output file, you can download one here
• Run the script with the following command line: python MC_data_new.py ConfigFile_new.txt -I <path_to_input_folder>

You will find the output in the default out/ folder.

You can draw the image opening the output in an interactive ROOT session. To make the image similar to the experimental data, we advice to use the following commands

gStyle->SetPalette(kGreyScale)
gStyle->SetOptStat(0)

and to set properly the z-axis scale once the TH2F has been written with COLZ option.

Run in batch

To run in batch using PBS queue system you can use the script submit_digi_batch.py

Example command:

python scripts/submit_digi_batch.py `pwd` --inputdir /nfs/cygno/CYGNO-MC-data/pbs_outputs/CYGNO_60_40_ER_6_keV/ --outdir /nfs/cygno2/users/$USER/digitization-out/ --tag LIMEsaturation_test --conf ConfigFile_new_saturation.txt --ram 8000

If you want easily submit multiple jobs, you can use a similar script to scripts/run_batch.sh

You can check the status of the jobs you submitted with:

qstat | grep $USER

To print the status every 10 seconds, use the following command:

while true; do qstat | grep $USER; date; sleep 10; done

And you can cancel a job with:

qdel <job_number>.xmaster

Suggestions for debugging and contributing

If you have made minor changes to the code, and the physical model has not changed, the output should be the same (except for statistical fluctuations). Once you have set the same seed for random distributions, you can use the script compare_digitizations.py to easily compare the output of two simulations. For instance:

python3 compare_digitizations.py output1.root output2.root

Note, the two root file should have the same number of events.

Wiki page and documentation

You can find more information about the digitization and CYGNO Montecarlo on the Wiki-page

Work in progress

• Add an option in ConfigFile.txt to choose between different detectors and geometries, in order to simulate other setups without manually changing the parameters
• Parallelize background generation to make the script run faster
• Add an option in ConfigFile.txt to set x,y,z offsets
• Add random z,x,y options
• Add Vignetting effect

Possible improvements to reduce resource usage

• simulate 'exposure time' effect (to cut long tracks or miss spots in higher and lower part of the sensor)
• parallelize new saturation loop (speed up)
• parallelize PMT simulation (one core per PMT)
• reduce x-y dimension of single layer in saturation loop (save RAM, for oblique tracks)
• use sparse object for saturation (at the moment the numpy object is taking memory for zeros)
• use cython to compile code as C and define datatype (int16)