Create and Structure LLM Response Object

The new_LlmResponse() function sends a prompt to a Large Language Model (LLM) API and returns a structured response object. It integrates the credentials from an LlmApi object and the prompt configuration from an LlmPromptConfig object, handles request errors gracefully, and returns the model-generated content along with associated metadata.

Usage

new_LlmResponse(api, prompt_config)

Arguments

api

An object of class RemoteLlmApi or LocalLlmApi.

prompt_config

An object of class LlmPromptConfig, containing prompt content, model, and tuning parameters (e.g., temperature, max tokens).

Value

An object of class LlmResponse, which includes the following components:

- content: The raw API response returned from the model. - provider: The name of the API provider (e.g., "OpenAI", "DeepSeek"). - prompt_config: The unclassed list representation of the original prompt settings. - generated_text: The primary response text content from the model.

If an error occurs during validation or request sending, an empty list is returned with an error attribute containing the error message.

See also

[new_RemoteLlmApi()], [new_LlmPromptConfig()]]

Examples

api <- new_RemoteLlmApi(api_key_path = "path/to/key.txt", provider = "OpenAI")
prompt <- new_LlmPromptConfig(
  prompt_content = "Explain entropy in simple terms.",
  model = "gpt-3.5-turbo",
  temperature = 0.7
)
response <- new_LlmResponse(api, prompt)

if (!is.null(attr(response, "error"))) {
  cat("Error:", attr(response, "error"), "\n")
} else {
  cat("Model response:", response$generated_text, "\n")
}
#> Error: LLM API not valid, must be an RemoteLlmApi or LocalLlmApi object.