Update app.py

app.py

@@ -229,7 +229,16 @@ def respond(message, history, model, temperature, num_calls, use_web_search):
         else:
             if model == "@cf/meta/llama-3.1-8b-instruct":
                 # Use Cloudflare API
-                for partial_response in get_response_from_cloudflare(message, num_calls=num_calls, temperature=temperature):
+                embed = get_embeddings()
+                if os.path.exists("faiss_database"):
+                    database = FAISS.load_local("faiss_database", embed, allow_dangerous_deserialization=True)
+                    retriever = database.as_retriever()
+                    relevant_docs = retriever.get_relevant_documents(message)
+                    context_str = "\n".join([doc.page_content for doc in relevant_docs])
+                else:
+                    context_str = "No documents available."
+                
+                for partial_response in get_response_from_cloudflare(prompt="", context=context_str, query=message, num_calls=num_calls, temperature=temperature):
                     first_line = partial_response.split('\n')[0] if partial_response else ''
                     logging.info(f"Generated Response (first line): {first_line}")
                     yield partial_response