Usage
The workflow is organized as a stepwise procedure, where the steps can be turned on/off through Boolean flags. It is configured by one input file in JSON format, named workflow_input.json, which must be properly filled by the user before starting.
The execution is managed by the Python script workflow_main.py in the main folder. All the other scripts are in the folder py, while bash scripts for submitting jobs to the computational
nodes are in the folder sh.
To run, activate the software environment as explained here, or using a different package manager, and type (assuming the JSON input file is in the working folder):
python /path-to-software/cheese-ptha-master/workflow_main.py --sitename SITENAME --input_workflow /path-to-wdir/workflow_input.json
The steps marked as True in the JSON input file will be executed. Of course, if one single step is required, the previous ones are assumed to be already executed.
The word SITENAME in the command is a placeholder and must be replaced with a name chosen by the user (typically the name of the target site, but it could be the name of a project or whatever). It is worth noting that a new folder called SITENAME is created within the working folder, and all the outputs will be directed there. Moreover, the grid files, required for the tsunami simulations and located in the folder specified in the JSON input file (/path-to-grids/), must be named as SITENAME_grid0.grd, SITENAME_grid1.grd etc., from the coarsest to the finest resolution (more details).
In production mode, it is strongly recommended to run the workflow preventing the process from receiving the SIGHUP (Signal Hang UP) signal upon closing or exiting the terminal: