Update app.py

app.py

@@ -3,11 +3,9 @@ import streamlit as st
 import chatbot as demo_chat
 from transformers import AutoModelForCausalLM, AutoTokenizer
 
-st.title("Hi, I am Chatbot Philio :mermaid:")
-st.write("I am your hotel booking assistant for today.")
+st.title("Hi, I am Chatbot Philio :woman:")
+st.write("I am your hotel booking assistant. Feel free to start chatting with me.")
 
-#tokenizer, model = demo_chat.load_model()
-pipeline = demo_chat.load_pipeline()
 scrollable_div_style = """
 <style>
 .scrollable-div {
@@ -27,19 +25,32 @@ def render_chat_history(chat_history):
             with st.chat_message(message["role"]):
                 st.markdown(message["content"])
 
-def generate_response(chat_history):
-    tokenized_chat = tokenizer.apply_chat_template(chat_history, tokenize=True, add_generation_prompt=True, return_tensors="pt")
-    outputs = model.generate(tokenized_chat, do_sample =True, max_new_tokens=50, temperature = 0.3, top_p = 0.85)
-    answer = tokenizer.decode(outputs[0][tokenized_chat.shape[1]:],skip_special_tokens=True)
-    final_answer = answer.split("<")[0]
-    return final_answer
+# def generate_response(chat_history):
+#     tokenized_chat = tokenizer.apply_chat_template(chat_history, tokenize=True, add_generation_prompt=True, return_tensors="pt")
+#     outputs = model.generate(tokenized_chat, do_sample =True, max_new_tokens=50, temperature = 0.3, top_p = 0.85)
+#     answer = tokenizer.decode(outputs[0][tokenized_chat.shape[1]:],skip_special_tokens=True)
+#     final_answer = answer.split("<")[0]
+#     return final_answer
 
 #Application 
 #Langchain memory in session cache 
 if 'memory' not in st.session_state:
-    st.session_state.memory = demo_chat.demo_miny_memory(model)
+    st.session_state.memory = demo_chat.demo_miny_memory()
 
-system_content = "You are a friendly chatbot who always helps the user book a hotel room based on his/her needs.Based on the current social norms you wait for the user's response to your proposals."
+system_content = """
+You are an AI having conversation with a human. Below is an instruction that describes a task. 
+Write a response that appropriately completes the request.
+Reply with the most helpful and logic answer. During the conversation you need to ask the user 
+the following questions to complete the hotel booking task.
+
+1) Where would you like to stay and when?
+2) How many people are staying in the room?
+3) Do you prefer any ammenities like breakfast included or gym?
+4) What is your name, your email address and phone number? 
+
+Make sure you receive a logical answer from the user from every question to complete the hotel 
+booking process.
+"""
 #Check if chat history exists in this session
 if 'chat_history' not in st.session_state:
     st.session_state.chat_history = [
@@ -50,8 +61,8 @@ if 'chat_history' not in st.session_state:
         {"role": "assistant", "content": "Hello, how can I help you today?"},
     ] #Initialize chat history
     
-if 'model' not in st.session_state:
-    st.session_state.model = model
+# if 'model' not in st.session_state:
+#     st.session_state.model = model
     
 st.markdown('<div class="scrollable-div">', unsafe_allow_html=True) #add css style to container
 render_chat_history(st.session_state.chat_history)
@@ -64,9 +75,9 @@ if input_text := st.chat_input(placeholder="Here you can chat with our hotel boo
     st.session_state.chat_history.append({"role" : "user", "content" : input_text}) #append message to chat history
 
     with st.spinner("Generating response..."):
-        first_answer = demo_chat.generate_from_pipeline(st.session_state.chat_history, pipeline)
+        first_answer = demo_chat.demo_chain(input_text, st.session_state.memory)
         
         with st.chat_message("assistant"):
             st.markdown(first_answer)
         st.session_state.chat_history.append({"role": "assistant", "content": first_answer})    
-st.markdown('</div>', unsafe_allow_html=True)
+st.markdown('</div>', unsafe_allow_html=True)