Passing Config to langgraph tool doesnt work with 0.2.x version upgrades. It was working fine till 0.1.5.

[akash9771]

Checked other resources

• I added a very descriptive title to this question.
• I searched the LangChain documentation with the integrated search.
• I used the GitHub search to find a similar question and didn't find it.

Commit to Help

• I commit to help with one of those options 👆

Example Code

Hello Team, i am using BaseTool where i have defined config like below

    def _run(
        self,
        run_manager: Optional[CallbackManagerForToolRun] = None,
        config: Optional[RunnableConfig]=None,
        
        
        **kwargs: Any,
        


        
    ) -> Union[Dict, str]:
        """Synchronous tool execution."""
        if self.lambda_wrapper:
            try:
                print("PAYLOAD AND CONFIG===================================+++++++++++++++++++++++++",config)
Calling the langgraph like below:

        runnable_config = RunnableConfig(configurable=self.base_config)
        state = {"messages": ("user", self.user_query)}
        logger.debug(f"runnable_config: {runnable_config}")
        messages = self.workflow.invoke(input=state, config=runnable_config)
SAME SETUP WORKS TILL langgraph 0.1.5, its stoped working and printing the config as None in my basetool above print statement. Any reason or do u suggest i need to modify my exixting code? any documentation for this change?

Description

langgraph 2.0

System Info

latest langchain and langraph

[dosubot[bot]]

The configuration passing to the langgraph tool stops working and prints None in version 0.2.x because, as of release 0.2.0, langchain is required to be integration-agnostic. This means that code in langchain should not by default instantiate any specific chat models, LLMs, embedding models, vectorstores, etc.; instead, the user is required to specify those explicitly.

To fix this issue, you need to modify your existing code to explicitly pass the required models as arguments. For example, if you were using a function or class that previously did not require an explicit LLM or embedding model, you now need to pass those explicitly.

Here is an example of how you might modify your code:

Before 0.2.0:

from langchain import SomeFunction

result = SomeFunction(config)

After 0.2.0:

from langchain import SomeFunction
from some_llm_library import SpecificLLM

llm = SpecificLLM()
result = SomeFunction(config, llm=llm)

Make sure to review the specific functions and classes you are using and update them accordingly based on the breaking changes listed in the documentation.

Additionally, the 0.2.0 release removed the dependency of langchain on langchain-community, which means that any imports relying on langchain-community will need to be updated. User code that still relies on deprecated imports from langchain will continue to work as long as langchain_community is installed, but these imports will start raising errors in release 0.4.x. Therefore, it is advisable to update your imports to align with the new package structure.

To assist with this migration, you can use the langchain-cli tool to help replace old imports with the new ones. Here is how you can use the tool:

1. Install the langchain-cli tool:

   pip install langchain-cli
   langchain-cli --version # <-- Make sure the version is at least 0.0.22

2. Run the migration script twice to apply all necessary import replacements:

   # Run a first time
   langchain-cli migrate --diff [path to code] # Preview
   langchain-cli migrate [path to code] # Apply
   
   # Run a second time to apply more import replacements
   langchain-cli migrate --diff [path to code] # Preview
   langchain-cli migrate [path to code] # Apply

3. Manually resolve any remaining deprecation warnings and re-run your unit tests to ensure everything works correctly.

Additionally, some tools have been moved to langchain_experimental, and you will need to update your imports accordingly. For example, tools like PythonAstREPLTool and PythonREPLTool have been moved, and you should replace their imports from langchain to langchain_experimental:

# Before 0.2.0
from langchain import PythonAstREPLTool

# After 0.2.0
from langchain_experimental import PythonAstREPLTool

By following these steps, you can update your code to be compatible with LangChain version 0.2.x ^[1][2][3][4].

To continue talking to Dosu, mention @dosu.

[nath-byond]

any update in this issue?