[FEATURE] Add support for Chat Completion Input Audio

[sergioverde90]

Is your feature request related to a problem? Please describe.
As discussed in langchain4j/langchain4j#2434 (comment), the base code for Content.java has to be updated in order to support the input_audio content type.

Describe the solution you'd like

public enum ContentType {

    @JsonProperty("text")
    TEXT,
    @JsonProperty("image_url")
    IMAGE_URL,
    
   // NEW!
    @JsonProperty("input_audio")
    AUDIO
}

...

@JsonDeserialize(builder = InputAudio.Builder.class)
@JsonInclude(JsonInclude.Include.NON_NULL)
@JsonNaming(PropertyNamingStrategies.SnakeCaseStrategy.class)
public class InputAudio {

    private String data;
    private String format;

...

[sergioverde90]

Hi guys,

I'm trying to implement this but as far as I see the unit tests are using the real OpenAI API, so I need a "subscription" (add the payment method and take into account the tokens) in order to run the unit tests locally, am I right?

[dliubarskyi]

Hi @sergioverde90, yes, you need an API key to test. I can provide a temporary one if you'll ping me via email: [email protected].
How do you plan to use this feature though? Which LLM provider do you plan to use?

[sergioverde90]

Thanks @dliubarskyi,

Because I want to contribute with something useful to this project, my aim is to solve langchain4j/langchain4j#2434 because it looks a good starting point. So answering your question I only plan to use the key to pass the tests locally, but maybe there are some alternatives to that. And becasue the original issue points to OpenAI that's the provider I would use.

It's almost implemented btw, but I want (and have) to run the tests to verify that everything is working as expected.