langchain-ai/langgraph
1
0
Fork 0
mirror of https://github.com/langchain-ai/langgraph.git synced 2026-07-15 12:35:43 -04:00
Code Issues 314 Packages Projects Releases 17 Wiki Activity

Custom tools(implemented in "Subclass Basetool" approach) cannot access to state via run_manager #238

New Issue
Closed
opened 2026-02-20 17:33:25 -05:00 by yindo · 2 comments
yindo commented 2026-02-20 17:33:25 -05:00
Owner
Copy Link
Copy Source

Originally created by @humbroll on GitHub (Sep 19, 2024).

Checked other resources

  • I added a very descriptive title to this issue.
  • I searched the LangGraph/LangChain documentation with the integrated search.
  • I used the GitHub search to find a similar question and didn't find it.
  • I am sure that this is a bug in LangGraph/LangChain rather than my code.
  • I am sure this is better as an issue rather than a GitHub discussion, since this is a LangGraph bug and not a design question.

Example Code

class GooglePlacesTool(BaseTool):
    """Tool that queries the Google places API."""

    name: str = "google_places"
    description: str = (
        "A tool that queries the Google Places API. "
        "Useful for finding places based on a search query. "
        "This tool can help validate addresses or discover locations "
        "from ambiguous text inputs. "
        "Input should be a search query string."
    )
    api_wrapper: GooglePlacesAPIWrapper = Field(default_factory=GooglePlacesAPIWrapper)  # type: ignore[arg-type]
    args_schema: Type[BaseModel] = GooglePlacesSchema


    def _run(
        self,
        query: str,
        run_manager: Optional[CallbackManagerForToolRun] = None,
    ) -> str:
        logger.debug(f"RunManager has state?: {hasattr(run_manager, 'state')}")

        # Run the original tool logic
        result = self.api_wrapper.run(query)

        return result

Error Message and Stack Trace (if applicable)

RunManager has state?: False

Description

How can I access state in a custom tool using the "Subclass Basetool" approach?

System Info

$ pip freeze | grep langchain langchain==0.3.0 langchain-anthropic==0.1.23 langchain-cohere==0.2.4 langchain-community==0.3.0 langchain-core==0.3.0 langchain-experimental==0.0.65 langchain-google-community==2.0.0 langchain-openai==0.2.0 langchain-text-splitters==0.3.0 langchainhub==0.1.21

Originally created by @humbroll on GitHub (Sep 19, 2024). ### Checked other resources - [X] I added a very descriptive title to this issue. - [X] I searched the [LangGraph](https://langchain-ai.github.io/langgraph/)/LangChain documentation with the integrated search. - [X] I used the GitHub search to find a similar question and didn't find it. - [x] I am sure that this is a bug in LangGraph/LangChain rather than my code. - [X] I am sure this is better as an issue [rather than a GitHub discussion](https://github.com/langchain-ai/langgraph/discussions/new/choose), since this is a LangGraph bug and not a design question. ### Example Code ```python class GooglePlacesTool(BaseTool): """Tool that queries the Google places API.""" name: str = "google_places" description: str = ( "A tool that queries the Google Places API. " "Useful for finding places based on a search query. " "This tool can help validate addresses or discover locations " "from ambiguous text inputs. " "Input should be a search query string." ) api_wrapper: GooglePlacesAPIWrapper = Field(default_factory=GooglePlacesAPIWrapper) # type: ignore[arg-type] args_schema: Type[BaseModel] = GooglePlacesSchema def _run( self, query: str, run_manager: Optional[CallbackManagerForToolRun] = None, ) -> str: logger.debug(f"RunManager has state?: {hasattr(run_manager, 'state')}") # Run the original tool logic result = self.api_wrapper.run(query) return result ``` ### Error Message and Stack Trace (if applicable) ```shell RunManager has state?: False ``` ### Description - I'm trying to access states in a custom tool implemented using the "[[Subclass Basetool](https://python.langchain.com/docs/how_to/custom_tools/#subclass-basetool)](https://python.langchain.com/docs/how_to/custom_tools/#subclass-basetool)" approach. - I've confirmed that I can access state with the [["@tool"](https://python.langchain.com/docs/how_to/custom_tools/#tool-decorator)](https://python.langchain.com/docs/how_to/custom_tools/#tool-decorator) decorator method. How can I access state in a custom tool using the "Subclass Basetool" approach? ### System Info `$ pip freeze | grep langchain langchain==0.3.0 langchain-anthropic==0.1.23 langchain-cohere==0.2.4 langchain-community==0.3.0 langchain-core==0.3.0 langchain-experimental==0.0.65 langchain-google-community==2.0.0 langchain-openai==0.2.0 langchain-text-splitters==0.3.0 langchainhub==0.1.21`
yindo closed this issue 2026-02-20 17:33:25 -05:00
yindo commented 2026-02-20 17:33:26 -05:00
Author
Owner
Copy Link
Copy Source

@dikla-avigad commented on GitHub (May 21, 2025):

I had the same issue when trying to access the state from my custom tool, and I was able to resolve it using the following approach - I included the state as part of the tool’s input schema like this:

from typing import Annotated, Callable
from pydantic import BaseModel, Field
from langchain_core.tools import BaseTool
from langgraph.prebuilt import InjectedState
from langchain_core.callbacks import (
    CallbackManagerForToolRun,
    AsyncCallbackManagerForToolRun,
)
import inspect
import logging


class MyAgentState(BaseModel):
    customer: dict

class MyInput(BaseModel):
    query: str = Field(description="A string representing the search query")
    state: Annotated[MyAgentState, InjectedState]

class SearchQueryTool(BaseTool):
    name = "search_query"
    description = "Search for relevant results based on a query"

    def __init__(self, search_client: Callable[[str, str], list[str]]):
        super().__init__()
        self._search_client = search_client
        object.__setattr__(self, "args_schema", MyInput)

    async def _arun(
        self,
        *args: Any,
        run_manager: AsyncCallbackManagerForToolRun | None = None,
        **kwargs: Any
    ) -> Any:
        logger.info(f"[Tool Called] search_query with kwargs: {kwargs}")
        try:
            query = kwargs["query"]
            state: MyAgentState = kwargs["state"]

            # Access injected state
            customer_index = state.customer["index_name"]
            results = await self._search_client(
                index_name=customer_index,
                query=query,
            )
            return results
        except Exception as e:
            logger.exception(f"[Tool Error] search_query")
            raise Exception(f"Tool failed: {e}") from e
@dikla-avigad commented on GitHub (May 21, 2025): I had the same issue when trying to access the state from my custom tool, and I was able to resolve it using the following approach - I included the state as part of the tool’s input schema like this: ```python from typing import Annotated, Callable from pydantic import BaseModel, Field from langchain_core.tools import BaseTool from langgraph.prebuilt import InjectedState from langchain_core.callbacks import ( CallbackManagerForToolRun, AsyncCallbackManagerForToolRun, ) import inspect import logging class MyAgentState(BaseModel): customer: dict class MyInput(BaseModel): query: str = Field(description="A string representing the search query") state: Annotated[MyAgentState, InjectedState] class SearchQueryTool(BaseTool): name = "search_query" description = "Search for relevant results based on a query" def __init__(self, search_client: Callable[[str, str], list[str]]): super().__init__() self._search_client = search_client object.__setattr__(self, "args_schema", MyInput) async def _arun( self, *args: Any, run_manager: AsyncCallbackManagerForToolRun | None = None, **kwargs: Any ) -> Any: logger.info(f"[Tool Called] search_query with kwargs: {kwargs}") try: query = kwargs["query"] state: MyAgentState = kwargs["state"] # Access injected state customer_index = state.customer["index_name"] results = await self._search_client( index_name=customer_index, query=query, ) return results except Exception as e: logger.exception(f"[Tool Error] search_query") raise Exception(f"Tool failed: {e}") from e ```
yindo commented 2026-02-20 17:33:28 -05:00
Author
Owner
Copy Link
Copy Source

@sydney-runkle commented on GitHub (Jun 10, 2025):

Going to close for now given the above workaround, thanks @dikla-avigad!

@sydney-runkle commented on GitHub (Jun 10, 2025): Going to close for now given the above workaround, thanks @dikla-avigad!
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
agents
awaiting author revision
bug
change
chat models
checkpointer
ci
code quality
code quality
dependencies
documentation
drawing
duplicate
enhancement
good first issue
help wanted
invalid
maintainer
meta
missing information
not maintained by langchain
open-swe-auto
open-swe-max
open-swe-max-auto
pending
performance
pull-request
Mirrored from GitHub Pull Request
 question
requires design
sdk
server
stale
studio
tests
type checking
usage question
v0.5
v1.1
No Label
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
yindo
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: langchain-ai/langgraph#238
Delete Branch "%!s()"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?