Hi I'm currently trying to start to use langgraph and it's really awesome and incredible. I'm slightly confused on a certain subject though, how does langgraph identify tools? How does it know what a tool does? How can I convey the purpose of the tool in order to be clearer? From what I understand, its only means is to read the function and the input it gets. Does this mean I should be making super long function names in order to explain the finer details? This might be a very simple solution but I'm confused as to how it's able to fetch further information on the tool provided. From the tutorial they use the following example: @tool def human_assistance(query: str) -> str: """Request assistance from a human.""" human_response = interrupt({"query": query}) return human_response["data"] I don't see how it's able to understand from the function name that human_assitance is capable to interrupt a query and tell the ai graph that it will get a response from a human. For all I know, it could interpret it as "human_assistance" -> "human helps me" -> "human does the work" and not that it will have to continue querying. This is a bit of a far fetch but it's to better understand more complex tools and how to properly convey to an LLM its function. Continue reading...
