Pipelines for GRB/GW simulations task

• yaml: pip install PyYAML
• astropy: pip install astropy
• environs: pip install environs
• pandas: pip install pandas
• wget https://raw.githubusercontent.com/HESOFTS/sexten_2017/master/model_creator/scriptModel_variable.py

The scripts are developed and tested inside a ctools Anaconda environment in order ot use gammalib/ctools functionalities. I personally advise to use the Anaconda package. It can be also easily used in a job submitting system (such as LSF or others), provided that the right path are given. This set of scripts is built in order to handle this correctly.

NOTE: remember to load the IRFs in the ctools environment. Just download them from the webpage https://www.cta-observatory.org/science/cta-performance/ and unpack the FITS version in the parent folder of the $CALDB folder (visible from inside the environment). Other supported IRFs are those that can be downloaded from redmine.

Simulation of background to be saved and reused afterwards. The configuration file has to be prepared according to the job submitting system.

python ctools_pipe.py --background background.yaml --jobs jobs*.yaml

or

python ctools_pipe.py -b background.yaml -j jobs*.yaml

Creation of models (ctools compliant) for sources with variable spectra over time. The idea is to put all the sources in the same point in space, but each of them have also a lightcurve attached, which is shaped as a squared wave so that it's zero everywhere (the source is OFF), expect for a time window in which the source is ON. The creation of the model have to be done AFTER having sourced the ctools environment because the scrit is using ctools and gammalib, that are installed inside that environment.

python ctools_pipe.py --models model_input.yaml --jobs jobs_local.yaml

or

python ctools_pipe.py -m model_input.yaml -j jobs_local.yaml

Every source realization is attached to a background which was previously simulated: this saves time when thousands of simulations have to be done. This part performs the time selection and the detection of the source using 3 different methods:

1) mode_1: counts: onoff counts extracted manually: Li&Ma significance computed using gammapy.stat.significance_on_off (from here). Use ctskymap to get a significance sky map from which we can take the maximum/value at the position of the source to get the significance directly.

The next two methods uses the same approach but are implemented in a slightly different way: a fit at low energy (30-80 GeV) with a Power Law is used to determine the slope, prefactor and pivot energy which are then used, together with a cutoff energy, to create input model for the likelihood over the whole energy range.

2) mode_2: ctlike-onoff: calculates the TS with ctlike in the onoff mode

3) mode_3: ctlike-std: calculates the TS with ctlike in the standard mode

python ctools_pipe.py --simulation simulation.yaml --jobs jobs*.yaml

or

python ctools_pipe.py -s simulation.yaml -j jobs*.yaml

See the simulation_analysis.yaml file for mode informations (time cuts, energy cuts, ...).

You can always type:

python ctools_pipe.py --help

to get the help message on how to use the script.

• change the batch_example_lapp.sh to your batch_LAPP.sh after having set the LAPP_APP_SHARED folder properly.
• source batch_LAPP.sh in order to se the proper environmental variables
• change the variables_example.sh properly and create your variable.sh (see a variable_LAPP.sh example file).
• source variable.sh

batch_LAPP.sh and variable.sh have to be created once and then they just have to be sourced at the beginning of each new session.

How to launch a job on MUST:

• activate the environment: source activate ctools_p3
• launch command, i.e. python ctools_pipe.py --background background.yaml --jobs jobs_qsub.yaml. This will NOT launch a job but it will print on screen the string to be used for submitting the job (try with few files...you'll see). you can pipe it to a launch_multiple.sh, for example, like this:

python ctools_pipe.py --background background.yaml --jobs jobs_qsub.yaml > launch_multiple.sh

• EXIT from the environment: source deactivate
• submit the jobs, i.e. source launch_multiple.sh

...et voilà!!!

.
└── $MAIN_FOLDER                        # This is created from the variable_*.sh file
    ├── back_sim                        # One folder per each IRF (N IRF here).
    |   ├── IRF1                        # i.e.: North_z20_30m
    |   |   ├── background_1_1.fits
    |   |   ├── ...        
    |   |   └── background_M_1.fits     # M files per each IRF
    |   ├── IRF2                        # i.e.: North_z40_30m
    |   |   ├── background_1_2.fits
    |   |   ├── ...        
    |   |   └── background_M_2.fits        
    |   ├── ...  
    |   └── IRFN                        # i.e.: South_z20_30m
    |       ├── background_1_N.fits
    |       ├── ...        
    |       └── background_M_N.fits    
    ├── input                           # input of the model creation. THIS HAS TO BE CREATED BY-HAND 
    |   ├── GRB_1.fits
    |   ├── ...
    |   └── GRB_X.fits                        
    ├── models                          # output of the model creation 
    |   ├── GRB_1
    |   |   ├── ...
    |   |   └── model_GRB_1.xml         # XML model for ctools for GRB_1
    |   ├── ...
    |   └── GRB_X.fits                        
    |       ├── ...
    |       └── model_GRB_X.xml         # XML model for ctools for GRB_X
    └── results