README.md

OpenAI Assistants API V2

A Swift package for integrating OpenAI Assistants API V2, supporting both streaming and standard data transmission. This package allows you to create threads, manage messages, and utilize advanced features like speech-to-text and text-to-speech.

Getting Started

Prerequisites

• Swift 5.5 or later
• Xcode 13.0 or later
• An OpenAI API Key, which you can obtain from OpenAI Platform.

Installation

To install the package, add the following dependency to your Package.swift file:

dependencies: [
    .package(url: "https://github.com/IAMTHEBURT/OpenAIAssistantsAPI.git", from: "0.1.0")
]

Usage

1. Obtain API Key

   First, go to OpenAI Platform and obtain your API key.

2. Initialize AssistantsAPI

   let apiKey = "YOUR_API_KEY"
   let assistantsAPI = AssistantsAPI(apiKey: apiKey)

3. Create an Assistant

To create a new assistant:

1. Visit OpenAI Assistants.

2. Follow the prompts to create a new assistant.

3. Once created, note the Assistant ID; you will need this ID to interact with the assistant through the this ID

4. Create a New Thread

assistantsAPI.createThread { result in
    switch result {
    case .success(let threadResponse):
        let thread = threadResponse.thread
        // Use the thread for further operations
    case .failure(let error):
        print("Failed to create thread: \(error)")
    }
}

5. Add a Message to a Thread You can add a message while creating a thread or separately using the createMessage method.

assistantsAPI.createMessage(
    to: thread,
    role: .user,
    content: [ContentPart(text: "Hello, how can I assist you today?")]
) { result in
    switch result {
    case .success(let message):
        print("Message created: \(message)")
    case .failure(let error):
        print("Failed to create message: \(error)")
    }
}

6. Create a Run (Non-Streaming) If you are not using streaming, create a run by specifying the assistant ID:

assistantsAPI.createRun(
    for: thread,
    assistantId: "ASSISTANT_ID"
) { result in
    switch result {
    case .success(let run):
        print("Run created: \(run)")
    case .failure(let error):
        print("Failed to create run: \(error)")
    }
}

7. Check Run Status Periodically check the run status:

assistantsAPI.getRun(threadID: thread.id, runID: run.id) { result in
    switch result {
    case .success(let run):
        print("Run status: \(run.status)")
    case .failure(let error):
        print("Failed to get run status: \(error)")
    }
}

8. List Messages in a Thread Once the run is complete, list all messages in the thread:

assistantsAPI.listMessages(from: thread) { result in
    switch result {
    case .success(let messages):
        print("Messages: \(messages)")
    case .failure(let error):
        print("Failed to list messages: \(error)")
    }
}

9. Create a Run with Streaming For streaming, create a run with the stream parameter set to true and handle partial responses:

assistantsAPI.createRunStream(
    for: thread,
    assistantId: "ASSISTANT_ID",
    stream: true,
    onPartialResponse: { partialResponse in
        print("Partial response: \(partialResponse)")
    },
    completion: { result in
        switch result {
        case .success(let run):
            print("Run completed: \(run)")
        case .failure(let error):
            print("Failed to complete run: \(error)")
        }
    }
)

10. Speech-to-Text and Text-to-Speech

Text-to-Speech

assistantsAPI.createSpeech(
    input: "Hello, how can I assist you?",
    voice: Voice(name: "default")
) { result in
    switch result {
    case .success(let data):
        print("Speech created successfully")
    case .failure(let error):
        print("Failed to create speech: \(error)")
    }
}

Speech-to-Text

assistantsAPI.createTranscription(
    file: audioData
) { result in
    switch result {
    case .success(let transcription):
        print("Transcription: \(transcription.text)")
    case .failure(let error):
        print("Failed to create transcription: \(error)")
    }
}

Contributing

Demo projects and tests will be added soon. I welcome pull requests from the community. Please feel free to contribute!

License

This project is licensed under the MIT License.

For any issues or questions, feel free to open an issue or reach out.

Happy coding!

Ivan Lvov