Update main.py

main.py

@@ -86,6 +86,7 @@ def to_cache(file):
 
 @cl.cache
 def retriever_to_cache():
+    from langchain.cache import InMemoryCache
     index_name = os.environ['PINECONE_INDEX_NAME']
     embeddings = HuggingFaceEmbeddings()
     pinecone.init(
@@ -96,6 +97,7 @@ def retriever_to_cache():
         index_name=index_name, embedding=embeddings
     )
     retriever = vectorstore.as_retriever(search_type="similarity_score_threshold", search_kwargs={"score_threshold": .7, "k": 30,"filter": {'categorie': {'$eq': 'OF'}}})
+    retriever = InMemoryCache()
     return retriever
 
 @cl.set_chat_profiles
@@ -169,11 +171,11 @@ async def start():
         )
         #question_generator = LLMChain(llm=llm, prompt=CONDENSE_QUESTION_PROMPT)
         #doc_chain = load_qa_chain(streaming_llm, chain_type="stuff")
-        relevant=retriever_to_cache()
+        #relevant=retriever_to_cache()
         qa = ConversationalRetrievalChain.from_llm(
             streaming_llm,
             chain_type="stuff",
-            retriever=relevant,
+            retriever=retriever_to_cache(),
             #combine_docs_chain=doc_chain,
             #question_generator=question_generator,
             memory=memory,