new cache

app.py

@@ -6,6 +6,7 @@ from langchain.text_splitter import RecursiveCharacterTextSplitter
 from langchain.chains.question_answering import load_qa_chain
 import streamlit as st
 from langchain.chains import RetrievalQA
+query = st.text_input("Ask a question: ")
 @st.cache_resource
 def llm():
     return HuggingFaceHub(repo_id="google/flan-t5-small", model_kwargs={"temperature":0,"max_length":200}, huggingfacehub_api_token="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") # type: ignore
@@ -18,8 +19,8 @@ def qa(query):
     texts = text_splitter.split_documents(docs)
     embeddings = HuggingFaceEmbeddings()
     vectordb = Chroma.from_documents(documents=texts,embedding=embeddings)
-    return RetrievalQA.from_chain_type(llm=llm(), chain_type="stuff",retriever=vectordb.as_retriever(search_type="mmr", search_kwargs={'fetch_k': 30}), return_source_documents=True)
-query = st.text_input("Ask a question: ")
-result = qa({"query": query})
-st.write(result["result"])
-st.write(result["source_documents"][0])
+    qa = RetrievalQA.from_chain_type(llm=llm(), chain_type="stuff",retriever=vectordb.as_retriever(search_type="mmr", search_kwargs={'fetch_k': 30}), return_source_documents=True)
+    result = qa({"query": query})
+    st.write(result["result"])
+    st.write(result["source_documents"][0])
+qa(query)