To log the model calls and answers into a variable using LangChain Callbacks, you can create a custom Callback Handler.
from typing import Any, Dict
from langchain import PromptTemplate
from langchain.callbacks.base import BaseCallbackHandler
from langchain.chains import LLMChain
from langchain.llms import OpenAI
llm = OpenAI()
prompt = PromptTemplate.from_template("1 + {number} = ")
handler = MyCustomHandler()
chain = LLMChain(llm=llm, prompt=prompt, callbacks=[handler])
chain.run(number=2)
class MyCustomHandler(BaseCallbackHandler):
def on_text(self, text: str, **kwargs: Any) -> Any:
print(f"Text: {text}")
self.log = text
def on_chain_start(
self, serialized: Dict[str, Any], inputs: Dict[str, Any], **kwargs: Any
) -> Any:
"""Run when chain starts running."""
print("Chain started running")
This custom handler logs the model’s text output and stores it in the handler.log variable.
Q&A
How can I remove the special characters/codes that appear in the variable logging the output of a LangChain LLM call?
—
LangChain adds ANSI color codes to the ‘formatted_prompt’, these are not passed to the LLM. Consider logging the raw ‘prompt’ text instead.
Can callbacks be used to log the model’s outputs?
—
Yes, callbacks can be used to log model outputs. Custom callback handlers can be created to handle text and other events.
What is the purpose of the on_chain_start method in the custom callback handler?
—
The on_chain_start method is called when the chain begins running. It can be used to perform tasks such as initializing variables or printing messages.
Video Explanation:
The following video, titled "Create a PDF Search and Summarization Tool in less than 100 ...", provides additional insights and in-depth exploration related to the topics discussed in this post.
In this tutorial, you'll learn how to build a powerful semantic search application using Streamlit and GPT-Index (Llama Index).
