Dev branch local ollama suggestion

[Super-six-java]

hi，bro
When I run the dev branch using the local ollama qwen2:72b model, should I use OllamaFunctions when loading the ollama model with the get_llm() method in the llm.py file?

    elif "ollama" in model_version:
        model_name, base_url = env_value.split(",")
        llm = OllamaFunctions(base_url=base_url, model=model_name)

The OllamaFunctions class inherits from the ChatOllama class and looks like has the with_stuctured_output() , but there is another issue where with_stuctured_output() seems to only output in JSON format.
If you use this method, you also need to change the schema_extraction_from_text() method in the schema_extraction.py file , for example

    prompt = ChatPromptTemplate.from_messages(
    [("system", schema_prompt), ("user", "{text}")]
    )
    

    if llm.get_name().startswith("Ollama"):
        runnable = prompt | llm.with_structured_output(schema=Schema, include_raw=False)
        
    else:

        runnable = prompt | llm.with_structured_output(
            schema=Schema,
            method="function_calling",
            include_raw=False,
        )

and change the get_graph_document_list() method in the llm.py file

def get_graph_document_list(
    llm, combined_chunk_document_list, allowedNodes, allowedRelationship
):
    futures = []
    graph_document_list = []
    # if llm.get_name() == "ChatOllama":
    if llm.get_name().startswith("Ollama") :
        node_properties = False
    else:
        node_properties = ["description"]