Langchain print sources

How are you setting up the tool and agent now?

Im using an agent with, for now, a single tool (where I specified the gpt_index). Furthermore Im using agent="conversational-react-description" and the ConversationBufferMemory as memory

Right! And I'm guessing you are using a lambda function for the tool func?

You can replace the lambda with our own function the calls the index query, prints the source nodes, and returns the response

Can you give me an example? I created a function which print “response.response” and “response.get_formatted_sources” and return “str(response.response)”

However when i use it in the func of Tool i get the following error:
ValidationError: 1 validation error for Tool func
It also tells me the error type: type=type_error.callable

I cannot figure out what is it

Ohhh, you need to set func to a func still. So either a lambda that calls your new function or pass the function itself

Here's an example I have

Here is the tool definition
https://github.com/logan-markewich/llama_index_starter_pack/blob/main/streamlit_sql_sandbox/streamlit_demo.py#L52

Here is the function definition
https://github.com/logan-markewich/llama_index_starter_pack/blob/main/streamlit_sql_sandbox/utils.py#L5