- The Shiny app now loads reviewed species/endpoint display metadata and file-upload guidance from packaged YAML once per session, shows friendly labels and units while preserving raw prior-bundle keys for bundle lookup, and safely falls back to raw codes for valid custom prior bundles without packaged metadata.
- Documentation now describes the CSV/TSV upload contract and the user's responsibility to confirm species, endpoint-definition, and measurement-unit compatibility.
- Added coverage for packaged YAML parsing, display-metadata fallback behavior, Shiny selector raw-value preservation, bundled-prior metadata rendering, and file-upload guidance rendering.
- Added
list_prior_bundle_endpoints(prior_file = NULL), a read-only API that reports the available species/endpoint combinations and convergence status in a validated prior bundle.
- Added packaged endpoint-description metadata for reviewed display labels and units.
- Target-study loading and the File Upload UI support CSV and TSV input only; Excel workbook input is not supported.
- Added Results-sidebar controls for figure typography and aspect ratio so credible-interval and probability plots can be formatted for publication.
- Consolidated the maintained workflow documentation into
vignettes/bayesian-priors-workflow.Rmd. - Documented the Shiny application as a runtime that uses a reviewed bundled prior artifact by default and does not build priors from raw historical data during an app session.
- Clarified synthetic demonstration-data use, prior compatibility responsibilities, convergence-evidence limitations, and supported CSV/TSV target-study input.
load_study_data()accepts an optionalPriorRhythmConfigobject for configuration-backed column mapping and animal-string resolution while preserving explicit-argument precedence.
- Removed the unreleased experimental Stan backend. Stage 1 prior construction uses JAGS.
- Added convergence-policy validation and compact endpoint-level convergence evidence to runtime prior bundles.
- Improved time-interval parsing, study-data validation, configuration support, and user-facing model terminology.
Earlier internal development history is retained in the private development repository and is intentionally omitted from this public-source candidate.