Update app.py

app.py

@@ -314,28 +314,114 @@ chain_neo4j = (
 
 
 
+# def generate_answer(message, choice, retrieval_mode, selected_model):
+#     logging.debug(f"generate_answer called with choice: {choice} and retrieval_mode: {retrieval_mode}")
+
+#     try:
+#         # Handle hotel-related queries
+#         if "hotel" in message.lower() or "hotels" in message.lower() and "birmingham" in message.lower():
+#             response = fetch_google_hotels()
+#             return response, extract_addresses(response)
+        
+#         # Handle restaurant-related queries
+#         if "restaurant" in message.lower() or "restaurants" in message.lower() and "birmingham" in message.lower():
+#             response = fetch_yelp_restaurants()
+#             return response, extract_addresses(response)
+
+#         # Handle flight-related queries
+#         if "flight" in message.lower() or "flights" in message.lower() and "birmingham" in message.lower():
+#             response = fetch_google_flights()
+#             return response, extract_addresses(response)
+
+#         if retrieval_mode == "VDB":
+#             if selected_model == chat_model:
+#                 # Use GPT-4o with its vector store and template
+#                 retriever = gpt_retriever
+#                 prompt_template = QA_CHAIN_PROMPT_1 if choice == "Details" else QA_CHAIN_PROMPT_2
+#                 context = retriever.get_relevant_documents(message)
+#                 prompt = prompt_template.format(context=context, question=message)
+
+#                 qa_chain = RetrievalQA.from_chain_type(
+#                     llm=chat_model,
+#                     chain_type="stuff",
+#                     retriever=retriever,
+#                     chain_type_kwargs={"prompt": prompt_template}
+#                 )
+#                 response = qa_chain({"query": message})
+#                 return response['result'], extract_addresses(response['result'])
+
+#             elif selected_model == phi_pipe:
+#                 # Use Phi-3.5 with its vector store and a simplified prompt
+#                 retriever = phi_retriever
+#                 context = retriever.get_relevant_documents(message)
+#                 prompt = f"""
+#                 Here is the information based on the documents provided:
+#                 {context}
+
+#                 {message}
+#                 """
+
+#                 logging.debug(f"Phi-3.5 Prompt: {prompt}")
+
+#                 response = selected_model(prompt, **{
+#                     "max_new_tokens": 300,
+#                     "return_full_text": False,
+#                     "temperature": 0.5,
+#                     "do_sample": False,
+#                 })
+
+#                 if response:
+#                     generated_text = response[0]['generated_text']
+#                     logging.debug(f"Phi-3.5 Response: {generated_text}")
+#                     return generated_text, extract_addresses(generated_text)
+#                 else:
+#                     logging.error("Phi-3.5 did not return any response.")
+#                     return "No response generated.", []
+
+#         elif retrieval_mode == "KGF":
+#             response = chain_neo4j.invoke({"question": message})
+#             return response, extract_addresses(response)
+#         else:
+#             return "Invalid retrieval mode selected.", []
+
+#     except Exception as e:
+#         logging.error(f"Error in generate_answer: {e}")
+#         return "Sorry, I encountered an error while processing your request.", []
+
+
+# Short Prompt Template for Phi-3.5 Proprietary Model
+
+phi_short_template = f"""
+As an expert on Birmingham, Alabama, I will provide concise, accurate, and informative responses to your queries. Given the sunny weather today, {current_date}, feel free to ask me anything you need to know about the city.
+
+- Brief details and facts
+- No unnecessary elaboration
+- Focus on relevance
+
+{{context}}
+Question: {{question}}
+Answer:
+"""
+
+# Incorporate this template into the Phi-3.5 response generation section
 def generate_answer(message, choice, retrieval_mode, selected_model):
     logging.debug(f"generate_answer called with choice: {choice} and retrieval_mode: {retrieval_mode}")
 
     try:
-        # Handle hotel-related queries
         if "hotel" in message.lower() or "hotels" in message.lower() and "birmingham" in message.lower():
             response = fetch_google_hotels()
             return response, extract_addresses(response)
         
-        # Handle restaurant-related queries
         if "restaurant" in message.lower() or "restaurants" in message.lower() and "birmingham" in message.lower():
             response = fetch_yelp_restaurants()
             return response, extract_addresses(response)
 
-        # Handle flight-related queries
         if "flight" in message.lower() or "flights" in message.lower() and "birmingham" in message.lower():
             response = fetch_google_flights()
             return response, extract_addresses(response)
 
         if retrieval_mode == "VDB":
             if selected_model == chat_model:
-                # Use GPT-4o with its vector store and template
                 retriever = gpt_retriever
                 prompt_template = QA_CHAIN_PROMPT_1 if choice == "Details" else QA_CHAIN_PROMPT_2
                 context = retriever.get_relevant_documents(message)
@@ -351,15 +437,9 @@ def generate_answer(message, choice, retrieval_mode, selected_model):
                 return response['result'], extract_addresses(response['result'])
 
             elif selected_model == phi_pipe:
-                # Use Phi-3.5 with its vector store and a simplified prompt
                 retriever = phi_retriever
                 context = retriever.get_relevant_documents(message)
-                prompt = f"""
-                Here is the information based on the documents provided:
-                {context}
-
-                {message}
-                """
+                prompt = phi_short_template.format(context=context, question=message)
 
                 logging.debug(f"Phi-3.5 Prompt: {prompt}")
 
@@ -392,7 +472,6 @@ def generate_answer(message, choice, retrieval_mode, selected_model):
 
 
 
-
 def bot(history, choice, tts_choice, retrieval_mode, model_choice):
     if not history:
         return history