Evaluates the active workflow packs against a derived workflow context and
returns obligations, suggested actions, and planner-ready external holds.
Callers can pass result$metadata$external_holds into bg_plan() to keep
workflow holds distinct from structural blockers.
Usage
bg_next_actions(project, scope = c("project", "branch"), branch_id = NULL)Value
A bg_next_actions_result plain-data list with context,
obligations, actions, and metadata$external_holds.
Details
For the built-in bayesgrove.default_bayesian pack, the returned
obligations and actions can include:
review_computation_validityand matchingcomputation_reviewactions for fresh warning/error summaries,review_fit_criticismplusfit_criticismandbranch_and_modifyactions for branch-scoped fit or diagnostic problems,compare_candidate_branchesplus comparison-node creation ormodel_comparisondecision actions when multiple clean fit candidates exist, andaccept_or_reject_branchplusbranch_dispositionactions after a current comparison exists for an active candidate set.
When the optional packs are active, the result can also include
prior-rationale recording, prior and posterior predictive review,
simulation-based calibration review, model-selection review keyed to
model_comparison and stacking_weights summaries, causal-question
prompts, and PAD annotation review.