Function Calling in ContextChatEngine

[Calabrone76]

We are using a chat engine with openai and GPT-4o-mini.
The engine is built using the ContextChatEngine class.
Here the snippet of code for the chat engine creation:

        v_store = db.db_vector_store(CONFIG['database']['view'], True)
        self.index = VectorStoreIndex.from_vector_store(
            vector_store=v_store,
            embed_model=OpenAIEmbedding(
                model=CONFIG['llm']['embedding']['model']            
            )
        )

        filters = MetadataFilters(
            filters=[
                MetadataFilter(key="id", value=list_embeddings, operator=FilterOperator.IN)
            ]
        )
        retriever = self.index.as_retriever(
            filters=filters,
            similarity_top_k=CONFIG['llm']['embedding']['limit'],            
            verbose=True,
            vector_store_query_mode=VectorStoreQueryMode.DEFAULT
        )
        chat = ContextChatEngine(
            retriever=retriever,
            llm=self.__llm,
            memory=self.__memory,
            prefix_messages=[]
        ).from_defaults(
            retriever=retriever,
            system_prompt=ConversationLLamaIndex.__MISSION        
        )

        self.lama_chat = chat

All work fine, but now we would like to implement also function_calling. Is it possible to use this feature in the structure above?

EDIT: i've tryed to use tools in llm configuration passing as additional_kwargs but it doesn't work

        tools = [
            {
                "name": "get_weather",
                "description": "Ottieni le informazioni meteo per una specifica località",
                "parameters": {
                    "type": "object",
                    "properties": {
                        "location": {
                            "type": "string",
                            "description": "La località di cui vuoi sapere il meteo"
                        }
                    },
                    "required": ["location"]
                }
            }
        ]
self.__llm = OpenAI(model=model, api_key=CONFIG['openai']['api_key'],additional_kwargs={"tools":tools})