Basically, I have one agent and it takes lot of tools.

When runninng perticular tool, inside the tool, i want to know the state of the langraph state, but I am not able to access it.

My main agenrt code as follows:

#thread.py
#Create chatbot node
tools = [taxes_federal_internal_revenue_code, taxes_federal_court_cases, taxes_federal_treasury_regulations, taxes_international, taxes_federal_forms, taxes_states]
llm = ChatOpenAI(model='gpt-4o')
llm_with_tools = llm.bind_tools(tools)
def chatbot(state: MessagesState):
    toolResp = llm_with_tools.invoke(state["messages"])
    return MessagesState(
        messages=[toolResp], 
        log_stream_name=state["log_stream_name"],  # Corrected access
        next_step=state["next_step"],  # Ensure next_step is passed
)
graph_builder.add_node("chatbot", chatbot)

In this context, my taxes_states tool looks like following:

@tool()
def taxes_states(query: str) -> Tuple[List[str]]:
"""
Purpose: Retrieve relevant U.S. state tax law documents, including specific state tax codes, 
regulations, and guidance (e.g., California tax laws)
"""

print("-----------------taxes_states-----------")
print(query)
print("-----------------taxes_states-----------")

#Perform RAG here and get retrieved_docs and return it
return retrieved_docs

The documentation tells that how we can do that inside the node (/), but it does not mention how we can achive that inside the tool.

Basically, I have one agent and it takes lot of tools.

When runninng perticular tool, inside the tool, i want to know the state of the langraph state, but I am not able to access it.

My main agenrt code as follows:

#thread.py
#Create chatbot node
tools = [taxes_federal_internal_revenue_code, taxes_federal_court_cases, taxes_federal_treasury_regulations, taxes_international, taxes_federal_forms, taxes_states]
llm = ChatOpenAI(model='gpt-4o')
llm_with_tools = llm.bind_tools(tools)
def chatbot(state: MessagesState):
    toolResp = llm_with_tools.invoke(state["messages"])
    return MessagesState(
        messages=[toolResp], 
        log_stream_name=state["log_stream_name"],  # Corrected access
        next_step=state["next_step"],  # Ensure next_step is passed
)
graph_builder.add_node("chatbot", chatbot)

In this context, my taxes_states tool looks like following:

@tool()
def taxes_states(query: str) -> Tuple[List[str]]:
"""
Purpose: Retrieve relevant U.S. state tax law documents, including specific state tax codes, 
regulations, and guidance (e.g., California tax laws)
"""

print("-----------------taxes_states-----------")
print(query)
print("-----------------taxes_states-----------")

#Perform RAG here and get retrieved_docs and return it
return retrieved_docs

The documentation tells that how we can do that inside the node (https://langchain-ai.github.io/langgraph/how-tos/state-reducers/), but it does not mention how we can achive that inside the tool.

• langchain
• langgraph

1 Answer 1

Accessing the graph state within a LangGraph tool isn't straightforward due to the framework's design. However, you can pass the state to your tool by annotating the tool's parameters with InjectedState. Here's how:

from typing_extensions import Annotated
from langchain_core.tools import tool
from langgraph.prebuilt import InjectedState

@tool
def taxes_states(
    query: str,
    state: Annotated[dict, InjectedState],
) -> Tuple[List[str]]:
    """
    Purpose: Retrieve relevant U.S. state tax law documents, including specific state tax codes,
    regulations, and guidance (e.g., California tax laws)
    """
    # Access state information
    user_info = state.get("user_info", {})
    # Perform RAG here and get retrieved_docs
    return retrieved_docs

By using the InjectedState annotation, LangGraph injects the current state into the tool when it's called. This approach allows your tools to access and utilize the graph's state effectively.