openai_helper_gem

OpenAIHelperGem

OpenAIHelperGem is a Ruby module designed to simplify interactions with OpenAI's API. It provides functionality to upload files, send messages with structured outputs, generate speech from text, and automatically play the resulting audio on macOS.

Features

• File Upload: Upload files to OpenAI and retrieve file_id for further use.
• Text-to-Speech (TTS): Convert text to high-quality speech and save the audio locally.
• Autoplay Audio: Automatically play the generated audio after saving (configurable).
• Structured Output Extraction: Send messages and extract structured outputs from OpenAI responses.
• Custom Notifications: Handle errors with notifications (macOS, terminal output, or none).

Installation

Add the module to your project:

require_relative 'path_to_openai_helper_gem/openai_helper_gem'

Ensure you have the following environment variables set:

• OPENAI_API_KEY: Your OpenAI API key.
• GPT_MODEL (optional): The OpenAI model you wish to use (defaults to gpt-3.5-turbo).
• OPENAI_AUDIO_PATH (optional): Custom path to save generated audio (defaults to the user's home directory).

Usage

Basic Example

Send a message and extract structured output:

require_relative 'openai_helper_gem'

messages = [{ role: 'user', content: 'Tell me a joke.' }]
schema = {
  type: "object",
  properties: {
    joke: { type: "string" }
  },
  required: ["joke"]
}

response = OpenAIHelperGem::Client.perform_request(messages, schema)
puts response['joke']

File Upload

Upload a file to OpenAI and retrieve the file_id:

file_id = OpenAIHelperGem::Client.upload_file('path_to_your_file.pdf')

Generate Speech

Generate speech from text and automatically play the audio:

OpenAIHelperGem::Client.generate_speech("This is a test message.")

Generate speech without playing it immediately:

OpenAIHelperGem::Client.generate_speech("This is a test message.", "output.mp3", autoplay: false)

Notifications

Customize how errors and other notifications are handled:

Use terminal output:

OpenAIHelperGem::Client.notification_method = :puts

Disable notifications:

OpenAIHelperGem::Client.notification_method = :none

Default Paths

The module saves generated audio files to a default directory in the user's home folder:

OpenAIHelperGem::Client.default_audio_path = ENV['OPENAI_AUDIO_PATH'] || File.join(Dir.home, 'Local Resources', 'API Logs', 'OpenAI Whisper')

Complex Example

Here’s a more comprehensive example that combines file upload, schema validation, and text-to-speech generation:

require_relative 'openai_helper_gem'

# Step 1: Define messages and schema
messages = [
  { role: 'user', content: 'Please summarize the attached document.' }
]

schema = {
  type: "object",
  properties: {
    summary: { type: "string" }
  },
  required: ["summary"]
}

# Step 2: Send message with the file and extract structured output
response = OpenAIHelperGem::Client.perform_request(messages, schema, 'path_to_your_document.pdf')

# Step 3: Output the summary
puts response['summary']

# Step 4: Convert the summary to speech and save it as an audio file
OpenAIHelperGem::Client.generate_speech(response['summary'], "document_summary.mp3")

License

This module is free to use and modify under the MIT License. See the LICENSE.txt file for more details.