Noizu Labs: OpenAI v0.1.3

Noizu.OpenAI.Api.Audio (Noizu Labs: OpenAI v0.1.3)

Noizu.OpenAI.Api.Audio is a module that provides functionality for transcribing and translating audio using OpenAI's API.

This module offers two main functions:

  1. transcribe/3 - Transcribes the given audio file using a specified prompt and options.
  2. translate/3 - Translates the given audio file using a specified prompt and options.

usage

 Usage 

{:ok, transcript} = Noizu.OpenAI.Api.Audio.transcribe(audio, prompt, options)
{:ok, translation} = Noizu.OpenAI.Api.Audio.translate(audio, prompt, options)

Link to this section Summary

Types

transcribe_options()
transcribe_response()
transcribe_supported_models()
translate_audio_options()
translate_audio_response()
translate_audio_supported_models()

Functions

transcribe(audio, prompt, options \\ nil)

Transcribes audio using the specified model.

translate(audio, prompt, options \\ nil)

Translates audio using the specified model.

Link to this section Types

Link to this type

transcribe_options()

@type transcribe_options() ::
  %{
    optional(:model) =>
      Noizu.OpenAI.model_option(transcribe_supported_models()),
    optional(:response_format) => Noizu.OpenAI.response_format_option(),
    optional(:prompt) => String.t(),
    optional(:temperature) => Noizu.OpenAI.temperature_option(),
    optional(:language) => Noizu.OpenAI.language_option(),
    optional(:stream) => Noizu.OpenAI.stream_option()
  }
  | Keyword.t()
Link to this type

transcribe_response()

@type transcribe_response() :: map()
Link to this type

transcribe_supported_models()

@type transcribe_supported_models() :: :"whisper-1" | atom()
Link to this type

translate_audio_options()

@type translate_audio_options() ::
  %{
    optional(:model) =>
      Noizu.OpenAI.model_option(translate_audio_supported_models()),
    optional(:response_format) => Noizu.OpenAI.response_format_option(),
    optional(:prompt) => String.t(),
    optional(:temperature) => Noizu.OpenAI.temperature_option(),
    optional(:stream) => Noizu.OpenAI.stream_option()
  }
  | Keyword.t()
Link to this type

translate_audio_response()

@type translate_audio_response() :: map()
Link to this type

translate_audio_supported_models()

@type translate_audio_supported_models() :: :"whisper-1" | atom()

Link to this section Functions

Link to this function

transcribe(audio, prompt, options \\ nil)

@spec transcribe(
  audio :: String.t(),
  prompt :: String.t(),
  options :: transcribe_options()
) ::
  {:ok, transcribe_response()} | {:error, term()}

Transcribes audio using the specified model.

parameters

 Parameters

  • audio: The URL of the audio file to transcribe
  • prompt: The prompt or instruction for the transcription
  • options: A map or keyword list of options, such as model, response_format, temperature, language, and stream

returns

 Returns

Returns a tuple {:ok, response} on successful transcription, where response is a map containing the transcription result. Returns {:error, term} on failure, where term contains error details.

example

 Example 

audio_url = "https://your_audio_url_here"
prompt = "Transcribe this audio."
{:ok, response} = Noizu.OpenAI.transcribe(audio_url, prompt)
Link to this function

translate(audio, prompt, options \\ nil)

@spec translate(
  audio :: String.t(),
  prompt :: String.t(),
  options :: translate_audio_options()
) ::
  {:ok, translate_audio_response()} | {:error, term()}

Translates audio using the specified model.

parameters

 Parameters

  • audio: The URL of the audio file to translate
  • prompt: The prompt or instruction for the translation
  • options: A map or keyword list of options, such as model, response_format, temperature, and stream

returns

 Returns

Returns a tuple {:ok, response} on successful translation, where response is a map containing the translation result. Returns {:error, term} on failure, where term contains error details.

example

 Example 

audio_url = "https://your_audio_url_here"
prompt = "Translate this audio to English."
{:ok, response} = Noizu.OpenAI.translate_audio(audio_url, prompt)