I’m using LangGraph to control the chain behaviour and path.
I have defined 3 functions (get_time, get_content, get_information). The three functions returns the state.
After I convert them using convert to openai function and invoke the graph as following:

def graph_core(state: Graph) -> Graph:
  template = get_template()
      prompt = ChatPromptTemplate.from_template(template)
      model = ChatOpenAI(model='gpt-4o')
      graph = prompt | model.bind_functions(functions=[convert_to_openai_function(tool) for tool in tools])
      graph_response = graph.invoke(
          {
              'user_question': state['user_question'],
              'time': state['time'],
              'content': state['content'],
              'information': state['information'],
              'chat_history': state['chat_history'],
              'response': state['response'],
              'messages': state['messages']
              })
      print(graph_response)

The graph calls only 2 of the three bound tools. Knowing that the template contains keys for the three variables.
This is the router node, invoke tool node, and invoke_tool function.

I even believe that the invoke tool should be implemented like this. I cannot see a difference between using tools as nodes and the implementation of this function.

def should_continue(state: Graph) -> Literal['tool_node', '__end__']:
    """Returns the next node's name to be executed"""
    last_message = state['messages']
    if name in last_message.additional_kwargs['function_call']:
        return 'tool_node'
    else:
    return '__end__'

def invoke_tool(state: Graph) -> Graph:
    last_message = state['messages']
    tool_name = last_message.additional_kwargs['function_call']['name']

    if tool_name == 'get_time':
        if not state['time']:
            updated_time = get_time(state)
            state['time'] = updated_time['time']

    if tool_name == 'get_content':
        if not state['content']:
            updated_content = get_content(state)
            state['content'] = updated_content['content']

    if tool_name == 'get_information':
        if not state['information']:
            updated_information = get_information(state)
            state['information'] = updated_information['information']

    print('nupdated state in invoke tool')
    print(state)
    print('n')
    
    return state


def tool_node(state: Graph) -> Graph:
    updated_state = invoke_tool(state)
    return updated_state

Your guidance is much appreciated!