Check streaming

Last good version is the chatgpt

app.py

@@ -122,7 +122,7 @@ def app():
     }
     conversational_qa_chain = _inputs | _context | ANSWER_PROMPT | llm
 
-    st.header("Hello Doctor!")
+    st.header("Ask Away!")
     for message in st.session_state.messages:
         with st.chat_message(message["role"]):
             st.write(message["content"])
@@ -140,19 +140,29 @@ def app():
             with st.spinner("Thinking..."):
                 for _ in range(3):  # Retry up to 3 times
                     try:
-                        response = conversational_qa_chain.invoke(
+                        response_generator = stream_conversational_qa_chain(
                             {
                                 "question": prompts2,
                                 "chat_history": chistory,
                             }
                         )
-                        st.write(response)
-                        message = {"role": "assistant", "content": response}
-                        st.session_state.messages.append(message)
+                        message_content = ""
+                        for response_part in response_generator:
+                            message_content += response_part
+                            st.chat_message("assistant", message_content)
+                        st.session_state.messages.append({"role": "assistant", "content": message_content})
                         break
                     except Exception as e:
                         st.error(f"An error occurred: {e}")
                         time.sleep(2)  # Wait 2 seconds before retrying
 
+def stream_conversational_qa_chain(inputs):
+    try:
+        response = conversational_qa_chain.invoke(inputs)
+        for part in response:
+            yield part
+    except Exception as e:
+        raise e
+
 if __name__ == '__main__':
     app()