Plan a multi-model, multi-prompt grid

Validate model and prompt configuration tables and estimate the number of model responses before making any calls. This is the dry-run companion to run_prompt_grid(). Prompt variants are independent calls, not sequential turns in one conversation.

Usage

plan_prompt_grid(data, prompt_variants, model_config, smoke_n = NULL)

Arguments

data

A data frame accepted by run_structured_responses().

prompt_variants

A named character vector of independent prompt templates, or a data frame with prompt_id and prompt columns. An optional logical active column can disable prompt variants without deleting them.

model_config

A character vector of model names, or a data frame with a required model column. Supported optional columns are model_id, family, integration, virtual_key, base_url, temperature, output_mode, seed, max_active, rpm, n_completions, and active. Missing settings use the same defaults as run_structured_responses().

smoke_n

Optional positive integer limiting the workflow to the first smoke_n input rows. smoke_n = 1 is useful before an expensive run.

Value

A tibble with one row per active model-prompt combination and columns for input rows, completion batches, and estimated response calls.