Stop spurious NA R-hat warnings from estimate_secondary

[sbfnk-bot]

Description

This PR closes #1497.

The remaining case of that issue was estimate_secondary(), which still emitted the spurious The largest R-hat is NA, indicating chains have not mixed. warning (plus the accompanying near-zero-ESS warnings) for both chains. estimate_infections() (#1499) and estimate_truncation() (#1501) were already handled; secondary was never covered because the monitoring-exclusion in create_stan_args() was gated on identical(model, "estimate_infections").

Reproducing fit_secondary_fixed from the prior choice guide, the only parameters with NA R-hat are delay_np_pmf_use[1..31], the structurally-constant nonparametric delay PMF, which is the same quantity already excluded for estimate_infections. Every real parameter is well converged. This PR extends the exclusion to estimate_secondary so that quantity is no longer monitored, which removes the spurious NA R-hat (verified: 0 NA-R-hat parameters afterwards).

It also bumps the prior choice guide's MCMC settings from samples = 100, warmup = 100 to samples = 500, warmup = 250. At 100 samples the genuine Bulk/Tail ESS warnings appear across the guide's fits simply because 100 post-warmup draws is too few (for estimate_secondary, 54 to 85 of 115 real parameters fell below an ESS of 100). At 500/250 those fits converge cleanly (0 real parameters below ESS 100, max R-hat 1.01) at roughly 8 s per secondary fit. The baked prior_choice_guide.Rmd is regenerated by the render-prior_choice_guide.yaml workflow, which fires on push to main when the .orig changes, so it is not committed here.

Initial submission checklist

• My PR is based on a package issue and I have explicitly linked it.
• I have tested my changes locally.
• I have added or updated unit tests where necessary.
• I have updated the documentation if required and rebuilt docs if yes.
• I have followed the established coding standards.
• I have added a news item linked to this PR.

[sbfnk]

merging as minimal change to output