Hi everyone,
I am facing a little challenge when dealing with metadata.
Long story short, I first run a RAG that returns a bunch of documents. Each document is associated with metadata.
Then I build a prompt & combine it to all my documents (as context) before sending every to GPT. The goal of the prompt is to select the most relevant pieces of text from the different documents in order to answer a specific question. So basically I might start with 10 documents and end up with text extract from document 1, 5 and 8 for instance.
Blockquote
def gpt4_query(prompt, input_documents):
from langchain.chat_models import ChatOpenAI
from langchain.chains.question_answering import load_qa_chain
import time
time.sleep(3) #Pause to avoid overloading open ai API
#Initiate the LLM model access
llm = ChatOpenAI(openai_api_key="xxxx", model = "gpt-4-1106-preview")
#Initiative query chain mode.
chain = load_qa_chain(llm, chain_type="stuff",verbose=True) #"Stuff" chain type is used because we will only submit a small piece of text that is prefiletered with the semantic search
answer = chain.run(input_documents=input_documents, question=prompt, return_only_outputs=True) #Submit the analysis to GPT4 for final check / relevancy verification / Semantic cleaning
return answer
Is there a way I could keep the metadata from the documents that were selected (1,5,8 in my example) by GPT? Langchain input_document parameter does not seem to consider the metadata part so I’m not sure what to do to keep it.
Thanks a lot for the help