chatgpt: ChatGPT Wrapper (Deprecated)

Description

Provides a wrapper for the openai() function to facilitate migration from the deprecated chatgpt() function. This ensures backward compatibility while allowing users to transition to the updated features.

Usage

chatgpt(
  .llm,
  .model = "gpt-4o",
  .max_tokens = 1024,
  .temperature = NULL,
  .top_p = NULL,
  .top_k = NULL,
  .frequency_penalty = NULL,
  .presence_penalty = NULL,
  .api_url = "https://api.openai.com/",
  .timeout = 60,
  .verbose = FALSE,
  .json = FALSE,
  .stream = FALSE,
  .dry_run = FALSE
)

Value

An LLMMessage object with the assistant's reply.

Arguments

.llm: An LLMMessage (passed directly to the openai() function)
.model: A character string specifying the model to use.
.max_tokens: An integer specifying the maximum number of tokens (mapped to .max_completion_tokens in openai())
.temperature: A numeric value for controlling randomness. This is
.top_p: A numeric value for nucleus sampling, indicating the top
.top_k: Currently unused, as it is not supported by openai().
.frequency_penalty: A numeric value that penalizes new tokens based on their frequency so far.
.presence_penalty: A numeric value that penalizes new tokens based on whether they appear in the text so far.
.api_url: Character string specifying the API URL. Defaults to the OpenAI API endpoint.
.timeout: An integer specifying the request timeout in seconds. This is
.verbose: Will print additional information about the request (default: false)
.json: Should json-mode be used? (detault: false)
.stream: Should the response be processed as a stream (default: false)
.dry_run: Should the request is constructed but not actually sent. Useful for debugging and testing. (default: false)

Details

This function is deprecated and is now a wrapper around openai(). It is recommended to switch to using openai() directly in future code. The chatgpt() function remains available to ensure backward compatibility for existing projects.

Examples

Run this code

if (FALSE) {
# Using the deprecated chatgpt() function
result <- chatgpt(.llm = llm_message(), .prompt = "Hello, how are you?")
}