Classify a character vector into predefined categories using an LLM — classify_llm • classifyLLM

Designed for tidyverse pipelines. Sends each element of x to the model with a strict instruction to return exactly one label from categories.

Usage

classify_llm(
  x,
  categories,
  model = "gpt-4o-mini",
  temperature = 0,
  batch_size = 1L,
  delay = 0,
  verbose = FALSE
)

Arguments

x: Character vector to classify.
categories: Character vector of allowed categories (labels).
model: OpenAI model name. Default: "gpt-4o-mini".
temperature: Numeric in 0,1; default 0 for deterministic outputs.
batch_size: Integer; number of elements per API call batch. Default 1.
delay: Seconds to sleep between batches to avoid rate limits. Default 0.
verbose: Logical; print progress info. Default FALSE.

Value

A factor with levels = categories, aligned to input length.

Examples

if (FALSE) { # \dontrun{
Sys.setenv(OPENAI_API_KEY = "sk-...")
classify_llm(
  x = c("siamese kitty","golden retriever","parakeet"),
  categories = c("cat","dog","bird")
)
} # }