Structured Response using Semantic functions

[Vinaysoni2310]

Hello,
I am trying to build a chatbot using semantic kernel, but llm response doesn't always follow the json format response mentioned in the prompt of each function.
I tried using semantic_kernel.pydantic library but that gives me another Validation error like input should be dict type.

class botresponse(KernelBaseModel):
BotResponse:str

req_settings = kernel.get_prompt_execution_settings_from_service_id(service_id=service_id)
req_settings.max_tokens = 2000
req_settings.temperature = 0.7
req_settings.top_p = 0.8
req_settings.function_choice_behavior = FunctionChoiceBehavior.Auto(filters={"excluded_plugins": ["chat"]})
req_settings.response_format = botresponse

Please if somebody can help to resolve this. I am using gpt-4o to make calls.

[moonbox3]

Hello @Vinaysoni2310, can you please have a look at these samples? I am able to run them both against OpenAI and Azure OpenAI: https://github.com/microsoft/semantic-kernel/tree/main/python/samples/concepts/structured_output