Update app.py

app.py

@@ -200,14 +200,11 @@ def ask_question(question, temperature, top_p, repetition_penalty, web_search):
     if not question:
         return "Please enter a question."
     
-    if question in memory_database:
+    if question in memory_database and not web_search:
         answer = memory_database[question]
     else:
-        embed = get_embeddings()
         model = get_model(temperature, top_p, repetition_penalty)
         
-        history_str = "\n".join([f"Q: {item['question']}\nA: {item['answer']}" for item in conversation_history])
-        
         if web_search:
             search_results = google_search(question)
             context_str = "\n".join([result["text"] for result in search_results if result["text"]])
@@ -219,27 +216,34 @@ def ask_question(question, temperature, top_p, repetition_penalty, web_search):
 
             Current Question: {question}
 
+            If the web search results don't contain relevant information, state that the information is not available in the search results.
             Provide a concise and direct answer to the question:
             """
+            prompt_val = ChatPromptTemplate.from_template(prompt_template)
+            formatted_prompt = prompt_val.format(context=context_str, question=question)
         else:
+            embed = get_embeddings()
             database = FAISS.load_local("faiss_database", embed, allow_dangerous_deserialization=True)
+            history_str = "\n".join([f"Q: {item['question']}\nA: {item['answer']}" for item in conversation_history])
+            
             if is_related_to_history(question, conversation_history):
                 context_str = "No additional context needed. Please refer to the conversation history."
             else:
                 retriever = database.as_retriever()
                 relevant_docs = retriever.get_relevant_documents(question)
                 context_str = "\n".join([doc.page_content for doc in relevant_docs])
-            prompt_template = prompt  # Use the original prompt for document-based answers
-        
-        prompt_val = ChatPromptTemplate.from_template(prompt_template)
-        formatted_prompt = prompt_val.format(history=history_str, context=context_str, question=question)
+            
+            prompt_val = ChatPromptTemplate.from_template(prompt)
+            formatted_prompt = prompt_val.format(history=history_str, context=context_str, question=question)
         
         answer = generate_chunked_response(model, formatted_prompt)
         answer = re.split(r'Question:|Current Question:', answer)[-1].strip()
         
-        memory_database[question] = answer
+        if not web_search:
+            memory_database[question] = answer
     
-    conversation_history = manage_conversation_history(question, answer, conversation_history)
+    if not web_search:
+        conversation_history = manage_conversation_history(question, answer, conversation_history)
     
     return answer