Skip to contents

Converts the nested list response from an OpenAI embedding API request into a tidy tibble with embedding vectors as columns.

Usage

tidy_oai_embedding(response)

Arguments

response

An httr2 response object or the parsed JSON response from OpenAI's embedding API

Value

A tibble containing the embedding vectors as columns (V1, V2, etc.) and optionally an oai_index column if present in the response

Details

This function handles both single document and batch embedding responses. It extracts the embedding vectors and converts them into a wide format tibble where each column (V1, V2, ..., Vn) represents one dimension of the embedding vector. If the response includes index information, it adds an oai_index column to preserve the ordering.

Examples

if (FALSE) { # \dontrun{
  # Process response from httr2 request
  req <- oai_build_embedding_request("Hello world")
  resp <- httr2::req_perform(req)
  embeddings <- tidy_oai_embedding(resp)

  # Process already parsed JSON
  resp_json <- httr2::resp_body_json(resp)
  embeddings <- tidy_oai_embedding(resp_json)
} # }