Convert a recipe to an extraction plan

Compiles the recipe into an omop_plan suitable for server-side execution via ds.omop.plan.execute. The conversion maps population-level filters to cohort specifications, groups variables by output and table, selects the appropriate plan builder (person_level, features, events, survival, intervals, baseline, temporal_covariates, or person_period) for each output type, and attaches row-level filter trees.

Usage

recipe_to_plan(recipe)

Arguments

recipe

An omop_recipe object.

Value

An omop_plan object ready for execution.

Details

Multiple populations and recipe-level scope are both serialized into the plan for the server to execute:

• plan$populations carries every recipe population. A criteria population serializes as list(id, label, kind = "criteria", filter_tree, cohort_definition_id); a set-op population as list(id, label, kind = "setop", setop = list(op, members)). The base population is always included so its cohort drives single-population recipes exactly as before.

• every plan$outputs[[name]] carries the population_id it was authored against (default "base"), so the server materializes and gates each output against the right population.

• plan$scope carries the recipe-level scope (list(cohort, tables, combine)) the server folds and intersects into every population. It is omitted when no scope was set.

Recipes are the recommended interface for ordinary analysis code. Plans are retained as an explicit lower-level contract so advanced users, tests, and the server can inspect the exact payload before it is executed.

See also

recipe_execute, omop_recipe, ds.omop.plan

Examples

if (FALSE) { # \dontrun{
recipe <- omop_recipe(
  blocks = omop_variable_block(
    table = "condition_occurrence",
    concept_ids = c(201820), format = "binary"),
  outputs = omop_output(type = "wide"))
recipe_execute(recipe)
plan <- recipe_to_plan(recipe)  # advanced: inspect the server payload
} # }