EpiNow2 1.8.0

This is a major release with significant API improvements, new features, and bug fixes.

Highlights

• Unified return structure: All main modelling functions (estimate_infections(), estimate_secondary(), estimate_truncation(), epinow()) now return consistent S3 objects with fit, args, and observations elements
• New accessor methods: get_samples(), get_predictions(), get_parameters(), and summary() provide a consistent interface across all model types
• scoringutils integration: get_predictions() supports "sample" and "quantile" output formats for direct use with the scoringutils package
• Improved discretisation: Now uses the primarycensored package for more accurate PMF calculations
• Population-adjusted Rt: Accounts for susceptible depletion when pop is specified in rt_opts()

Breaking changes

• Changed discretise() to use the primarycensored package for double censored PMF calculations
• Changed population-adjusted Rt estimates to account for susceptible depletion
• Removed deprecated functions and arguments from v1.5.0/v1.6.0
• The variable column in get_samples() and summary() output now contains semantic parameter names

New features

• Added get_samples(), get_predictions(), and get_parameters() S3 generics
• Added support for fitting the susceptible population size
• Added growth rate estimation using Parag et al. (2022) method via growth_method argument
• Added style argument to plot methods for error bar display
• Added comprehensive prior choice guide vignette

Bug fixes

• Fixed incorrect Rt prior in vignettes causing divergent transitions
• Fixed report_log_lik using wrong overdispersion parameter
• Fixed CrIs parameter not being passed through in epinow()
• Fixed forecast_infections() failing with samples = 1
• Fixed Matern 5/2 Gaussian process kernel implementation
• Many other fixes (see full changelog)

Full Changelog: See NEWS.md for complete details.