streaming response from haystack

[RazaGR]

I am using haystack's OllamaGenerator which takes a streaming callback function for streaming message directly from llm, I am able to print this on server by below code but I am not sure how can I return that stream to caller instead of sync request.

@get("/query")
async def query_db(q: str) -> str:
    result = query_documents(q, streaming_callback)
    return result

def streaming_callback(chunk):
    print(chunk.content, end="", flush=True)  

I have tried to hack and store response in content and use a stream but no success, it still return me response at the end instead of streaming

content = []

# Global event to notify when content is updated
content_event = asyncio.Event()    

async def my_generator_two() -> AsyncGenerator[bytes, None]:
    while True:
        # Wait until content is updated
        await content_event.wait()
        content_event.clear()  # Reset event
        # Stream the updated content to the client
        yield encode_json({"content": content})        

@get("/query", sync_to_thread=False)
async def query_db(q: str) -> Stream:
     query_documents(q, streaming_callback))
    return Stream(my_generator_two())       

how can I make it happen?
Thanks