try catch

app.py

@@ -21,19 +21,25 @@ def remove_citation(text: str) -> str:
     pattern = r"【\d+†\w+】"
     return re.sub(pattern, "📚", text)
 
-# Initialize session state for messages and thread_id
+# Initialize session state
 if "messages" not in st.session_state:
     st.session_state["messages"] = []
 if "thread_id" not in st.session_state:
     st.session_state["thread_id"] = None
+# A flag to indicate if a request is in progress
+if "is_in_request" not in st.session_state:
+    st.session_state["is_in_request"] = False
 
 st.title("Solution Specifier A")
 
 def predict(user_input: str) -> str:
     """
     This function calls our OpenAIAssistantRunnable to get a response.
-    We either create a new thread (no thread_id yet) or continue the existing thread.
-    If the server complains that a run is still active, we reset the thread_id and retry once.
+    If st.session_state["thread_id"] is None, we start a new thread.
+    Otherwise, we continue the existing thread.
+
+    If a concurrency error occurs ("Can't add messages to thread..."), we reset
+    the thread_id and try again once on a fresh thread.
     """
     try:
         if st.session_state["thread_id"] is None:
@@ -41,19 +47,19 @@ def predict(user_input: str) -> str:
             response = extractor_llm.invoke({"content": user_input})
             st.session_state["thread_id"] = response.thread_id
         else:
-            # Continue the existing thread
+            # Continue existing thread
             response = extractor_llm.invoke(
                 {"content": user_input, "thread_id": st.session_state["thread_id"]}
             )
+
         output = response.return_values["output"]
         return remove_citation(output)
 
     except openai.error.BadRequestError as e:
-        # If the error says a run is still active, reset to a new thread and re-invoke once
-        err_msg = str(e)
-        if "while a run" in err_msg:
+        # If we get the specific concurrency error, reset thread and try once more
+        if "while a run" in str(e):
             st.session_state["thread_id"] = None
-            # Re-invoke once to get a fresh thread
+            # Now create a new thread for the same user input
             try:
                 response = extractor_llm.invoke({"content": user_input})
                 st.session_state["thread_id"] = response.thread_id
@@ -63,11 +69,10 @@ def predict(user_input: str) -> str:
                 st.error(f"Error after resetting thread: {e2}")
                 return ""
         else:
-            # Some other bad request
-            st.error(err_msg)
+            # Some other 400 error
+            st.error(str(e))
             return ""
     except Exception as e:
-        # Catch-all for any other error
         st.error(str(e))
         return ""
 
@@ -80,11 +85,16 @@ for msg in st.session_state["messages"]:
         with st.chat_message("assistant"):
             st.write(msg["content"])
 
-# Create the chat input widget at the bottom of the page
+# Chat input at the bottom of the page
 user_input = st.chat_input("Type your message here...")
 
-# When the user hits ENTER on st.chat_input
-if user_input:
+# Process the user input only if:
+# 1) There is some text, and
+# 2) We are not already handling a request (is_in_request == False)
+if user_input and not st.session_state["is_in_request"]:
+    # Lock to prevent duplicate requests
+    st.session_state["is_in_request"] = True
+
     # Add the user message to session state
     st.session_state["messages"].append({"role": "user", "content": user_input})
 
@@ -92,12 +102,15 @@ if user_input:
     with st.chat_message("user"):
         st.write(user_input)
 
-    # Get the assistant's response
+    # Get assistant response
     response_text = predict(user_input)
 
-    # Add the assistant response to session state
+    # Add assistant response to session state
     st.session_state["messages"].append({"role": "assistant", "content": response_text})
 
-    # Display the assistant's reply
+    # Display assistant response
     with st.chat_message("assistant"):
-        st.write(response_text)
+        st.write(response_text)
+
+    # Release the lock
+    st.session_state["is_in_request"] = False