CrI output issue

[nlinton]

Hi all, we have two EpiNow2 pipelines that analyze weekly data for two different diseases. In theory the pipelines are exactly the same. Same code (same notebooks), same compute, same everything except the input data. Both pipelines ran successfully, but one output the requested CrI of 0.95 while the other output the default EpiNow2 CrI values of 0.2 and 0.9.

Our epinow() call looks as below. We are trying to figure out whether the issue is related to the recent package updates or to something else. One possibility (though I don't know how) is the pipeline with the default CrI is running the updated EpiNow2 and the pipeline with our requested EpiNow2 is running an older package version somehow (we are rerunning them to try to verify if this is happening). At the same time, our copy of EpiNow2 is pulled from https://databricks.packagemanager.posit.co/cran/__linux__/noble/2025-03-20/ so also not sure if the package updates are even related?

In the meantime, any thoughts? Do we need to change the way we are selecting CrI?

EpiNow2::epinow(data=input_data,
      rt=rt_opts_vals,
      truncation=epinow2_truncation_module,
      gp=gp_opts_vals,
      generation_time=epinow2_gt,
      delays=epinow2_delay, 
      obs=obs_opts_vals,
      stan=stan_opts_vals,
      forecast=forecast_opts_vals,
      CrIs=c(0.95),
      return_output=T, 
      verbose=interactive()) 

[sbfnk]

This is almost certainly a bug that has arisen because of some recent changes in the package (thanks for reporting!) - which means that you're using the development version. If you're running this in production then you might consider switching to the CRAN version (which is of course a bit of a trade-off as then you don't always have access to the latest features / fixes.

[nlinton]

Thanks for sharing your thoughts. We should NOT be running the development version since we are doing our best to get our compute pointing to https://databricks.packagemanager.posit.co/cran/ for install. I have a suspicion of what might be going on so we'll try to get that separate issue sorted.

[seabbs]

Thanks for this @nlinton,

For completeness the bug report is now here: #1158

[nlinton]

Thanks for flagging!

[seabbs]

PR is merging in now so code on main should work for this shortly.

[nlinton]

We resolved the issue where one of our pipelines was pulling from dev. Glad we could help identify a bug along the way.