2/ there's so much redundancy with run_stan. For example, it's just very confusing to have sensitivity analysis switchs there. If the figures created do not compare sensitivity analysis, but are for one given analysis, we should transform the script as a postprocessing script (like src/transmission_flows/postprocessing_figures). In the latter, the env from rstan is loaded so there's no need to redefined the switchs and data every time...
concerning the script similar to run_stan.R , I function-ized the pair-selection so if any changes are made, they are loaded in both instances, and took away some unnecessary definitions.
concerning the "postrpoc' ones, I did remove the args_line, but I am not sure whether this is really ideal (why not use the HPC if posteriors are involved)
Hi @abriz97,
There is a need to clean the scripts in figs https://github.com/MLGlobalHealth/phyloSI-RakaiAgeGender/tree/main/figs...
Do you think you could have a look at this?
In particular: 1/ We do not want to have a R reader, i.e.,
without a pbs job to call it
2/ there's so much redundancy with run_stan. For example, it's just very confusing to have sensitivity analysis switchs there. If the figures created do not compare sensitivity analysis, but are for one given analysis, we should transform the script as a postprocessing script (like src/transmission_flows/postprocessing_figures). In the latter, the env from rstan is loaded so there's no need to redefined the switchs and data every time...
3/ the paths fo the function should be updated