Update app.py

app.py

@@ -85,17 +85,16 @@ if st.button('Crawl CUDA Documentation'):
     with st.spinner('Crawling CUDA documentation...'):
         crawled_data, crawled_urls = crawl("https://docs.nvidia.com/cuda/", max_depth=1, delay=0.1)
         st.write(f"Processed {len(crawled_data)} pages.")
-
+        
         texts = []
         for url, text in crawled_data:
             chunks = chunk_text(text, max_chunk_size=1024)
             texts.extend(chunks)
         st.success("Crawling and processing completed.")
-
+        
         # Create embeddings
-        embeddings = HuggingFaceEmbeddings(model_name='sentence-transformers/all-MiniLM-L6-v2',
-                                           model_kwargs={'device': 'cpu'})
-
+        embeddings = HuggingFaceEmbeddings(model_name='sentence-transformers/all-MiniLM-L6-v2', model_kwargs={'device': 'cpu'})
+        
         # Store embeddings in FAISS
         st.session_state.vector_store = FAISS.from_texts(texts, embeddings)
         st.session_state.documents_loaded = True
@@ -109,16 +108,13 @@ if query and st.session_state.documents_loaded:
         llm = GoogleGenerativeAI(model='gemini-1.0-pro', google_api_key="AIzaSyC1AvHnvobbycU8XSCXh-gRq3DUfG0EP98")
 
         # Create a PromptTemplate for the QA chain
-        qa_prompt = PromptTemplate(
-            template="Answer the following question based on the context provided:\n\n{context}\n\nQuestion: {question}\nAnswer:",
-            input_variables=["context", "question"])
+        qa_prompt = PromptTemplate(template="Answer the following question based on the context provided:\n\n{context}\n\nQuestion: {question}\nAnswer:", input_variables=["context", "question"])
 
         # Create the retrieval QA chain
-        qa_chain = RetrievalQA.from_chain_type(
-            chain_type="map_rerank",
+        qa_chain = RetrievalQA(
             retriever=st.session_state.vector_store.as_retriever(),
-            combine_documents_chain=qa_prompt,
-            llm=llm
+            llm=llm,
+            prompt=qa_prompt
         )
 
         response = qa_chain({"question": query})