Version 0.2.0 breaks example with completion

[jean-bovet]

Describe the bug
The latest version, 0.2.0, break the completion example, likely due to the stream being true and the decoding not expecting the JSON format which OpenAI returns.

To Reproduce
Steps to reproduce the behavior:

1. Create a new Xcode project
2. Use the example of the prompt in the README:
3. let query = CompletionsQuery(model: .textDavinci_003, prompt: "What is 42?", temperature: 0, max_tokens: 100, top_p: 1, frequency_penalty: 0, presence_penalty: 0, stop: ["\\n"]) openAI.completions(query: query) { result in //Handle result here } //or let result = try await openAI.completions(query: query)
4. Run the example and observe the following error:
5. failure(Swift.DecodingError.dataCorrupted(Swift.DecodingError.Context(codingPath: [], debugDescription: "The given data was not valid JSON.", underlyingError: Optional(Error Domain=NSCocoaErrorDomain Code=3840 "Invalid value around line 1, column 0." UserInfo={NSDebugDescription=Invalid value around line 1, column 0., NSJSONSerializationErrorIndex=0}))))

Expected behavior
Reverting to the version before 0.2.0 works fine with an expected result without error.

Desktop (please complete the following information):

• OS: macOS 13.3.1

[Krivoblotsky]

Hey, @jean-bovet.
Thanks for reporting.
My bad, forgot to switch the streaming off by default.
Just fixed it with https://github.com/MacPaw/OpenAI/releases/tag/0.2.1

Thanks!

[jean-bovet]

Thanks @Krivoblotsky for the quick turn-around. Much appreciated! May I suggest adding unit tests around this in the future?

[Krivoblotsky]

Sure, I’ll add it with next release 👌