Update app.py

app.py

@@ -59,15 +59,44 @@ for message in st.session_state.messages:
         st.markdown(message["content"])
 
 # Function to generate responses
-def generate_response(messages):
-    # Convert the messages to a format compatible with Hugging Face Inference API
-    messages_str = "\n".join([f"{msg['role']}: {msg['content']}" for msg in messages if msg['role'] != 'system'])
+# def generate_response(messages):
+#     # Convert the messages to a format compatible with Hugging Face Inference API
+#     messages_str = "\n".join([f"{msg['role']}: {msg['content']}" for msg in messages if msg['role'] != 'system'])
     
-    # Make API call to generate response
-    output = client.chat_completion(inputs=messages_str, max_tokens, temperature, top_p)
+#     # Make API call to generate response
+#     output = client(inputs=messages_str, max_tokens, temperature, top_p)
     
-    return output['generated_text']
+#     return output['generated_text']
 
+def respond(message, history, max_tokens, temperature, top_p):
+    # Prepare the list of messages for the chat completion
+    messages = [{"role": "system", "content": st.session_state.messages[0]["content"]}]
+
+    for val in history:
+        if val["role"] == "user":
+            messages.append({"role": "user", "content": val["content"]})
+        elif val["role"] == "assistant":
+            messages.append({"role": "assistant", "content": val["content"]})
+
+    messages.append({"role": "user", "content": message})
+
+    # Generate response
+    response = ""
+    response_container = st.empty()  # Placeholder to update the response text dynamically
+
+    for message in client.chat_completion(
+        messages,
+        max_tokens=max_tokens,
+        stream=True,
+        temperature=temperature,
+        top_p=top_p,
+    ):
+        token = message.choices[0].delta.content
+        response += token
+        response_container.text(response)  # Stream the response
+
+    return response
+    
 # User input
 if user_input := st.chat_input("Ask a health question..."):
     # Display user message
@@ -76,7 +105,7 @@ if user_input := st.chat_input("Ask a health question..."):
     st.session_state.messages.append({"role": "user", "content": user_input})
 
     # Generate and display assistant response
-    response = generate_response(st.session_state.messages)
+    response = respond(user_input, st.session_state.messages, max_tokens, temperature, top_p)
     with st.chat_message("assistant"):
         st.markdown(response)
     st.session_state.messages.append({"role": "assistant", "content": response})