Create a Model Configuration Schema — create_config

Creates a JSON Schema for validating model configuration. The schema defines what configuration options a model accepts, their types, constraints, and default values.

Usage

create_config_schema(
  title = NULL,
  description = NULL,
  properties = list(),
  required = character(0),
  additional_properties = TRUE
)

Arguments

title: Title for the schema (typically the model name)
description: Description of the configuration
properties: Named list of property definitions created with schema_integer(), schema_number(), schema_string(), schema_boolean(), schema_enum(), or schema_array()
required: Character vector of required property names
additional_properties: Whether to allow properties not defined in schema (default: TRUE for flexibility)

Value

A list representing a complete JSON Schema object

Examples

# Define a configuration schema
my_schema <- create_config_schema(
  title = "My Forecasting Model",
  description = "Configuration for disease case forecasting",
  properties = list(
    n_samples = schema_integer(
      description = "Number of Monte Carlo samples for predictions",
      default = 100L,
      minimum = 1L,
      maximum = 10000L
    ),
    learning_rate = schema_number(
      description = "Learning rate for model optimization",
      default = 0.01,
      minimum = 0,
      maximum = 1
    ),
    method = schema_enum(
      values = c("arima", "ets", "prophet"),
      description = "Forecasting method to use",
      default = "arima"
    ),
    use_covariates = schema_boolean(
      description = "Whether to include additional covariates",
      default = TRUE
    )
  ),
  required = c("n_samples")
)

# Use with create_chap_cli
if (FALSE) { # \dontrun{
create_chap_cli(train_fn, predict_fn, model_config_schema = my_schema)
} # }