mrc-cso-sphsu / simpaths_parallel Goto Github PK

A suggestion for this script would be to allow for optional running/re-running of SimPathsStart to generate the input.mv.db database before multi-runs. This is necessary and so far I only know how to do it using the GUI. SimPathsStart I think should be able to run headless, but at the moment those options seem disabled and have perhaps been deprecated.

Benefits of optional re-running single run:

• Can stabilise all run parameters by modifying/archiving input files - no longer have to re-enter GUI options
• Easier to run/describe run for reproducibility
• Can all be done headless
• Can force do_full_run.sh to re-run the setup to absolutely remove all traces of previous runs and ensure correct population being used

Some command line parameters stay roughly the same across the simulations, there is no need to enter them over and over again.

I'd like you to check the readme and correct/expand it.

• Add all elements required to conduct multiruns of SimPaths in parallel. #9
• Add scripts for data conversion

Doing:

./do_full_run.sh -n 20 -p 150000 -s 2017 -e 2023 -g false

returns:

All parameters must be provided
Usage: sh ././do_full_run.sh <[options]>
        Options:
                -n    --batch_size        The number od simulations to run in one batch, strictly integer and positive
                -p    --population_size   The size of a population, strictly integer and positive
                -s    --start_year        The year simulation starts, from 2010 to 2023
                -e    --end_year          The year simulation ends, from 2010 to 2023, greater or equal than `-s`
                -g    --gui               GUI flag, `true` or `false` to enable/disable GUI support

@andrewbaxter439

I'd like to know what certain parts of the bash script mean:

https://github.com/MRC-CSO-SPHSU/simpaths_parallel_r/blob/c3bac440058ab8fe65cfd9857fa68055b05fa0b6/do_full_run.sh#L2

1. Why use seq, parallel has its own inbuilt iterator? Also, why such a long range, 1-50 should be more than enough?
2. -r, -n, -p, -s, -e are self-explanatory, what about -g & -f?