Initialize a bayesgrove Project — bg

Creates the directory structure and initial state for a new bayesgrove project.

Usage

bg_init(
  path = ".",
  project_name = NULL,
  config = list(),
  workflow_packs = NULL
)

path: Directory path where the project should be initialized.
project_name: Optional. Name of the project. Defaults to the basename of the path.
config: Optional. Configuration list.
workflow_packs: Optional list of active workflow packs. These are normalized and fixed at project initialization. The project starts empty by default. Built-in packs such as bayesgrove.default_bayesian provide computation review, branch-scoped fit criticism, candidate-comparison guidance, and explicit branch acceptance or rejection decisions. Optional built-in packs add prior rationale and prior predictive review (bayesgrove.prior_workflow), posterior predictive and SBC review (bayesgrove.model_checks), model-selection review with stacking weights (bayesgrove.model_selection), minimal causal framing (bayesgrove.causal_minimal), and PAD annotations (bayesgrove.pad_scaffold).

A bg_handle representing the open project.