fix(stapp): fall back to st.rerun() when fragment scope rerun is unavailable

Issue #564 reports a StreamlitAPIException on startup when the LLM selectbox
in the sidebar detects a stale selected_idx (e.g. from a prior session_state).
The handler calls st.rerun(scope="fragment"), which is only legal during a
fragment rerun — on the initial script run it raises:

  StreamlitAPIException: scope="fragment" can only be specified from
  @st.fragment-decorated functions during fragment reruns.

Wrap the call so that on the initial run (or any context where fragment
rerun is unavailable) we fall back to a full st.rerun() instead of crashing.

Author: Kailigithub

frontends/stapp.py

@@ -105,7 +105,11 @@ def render_sidebar():
     st.caption(f"LLM Core: {llm_labels.get(current_idx, str(current_idx))}")
     selected_idx = st.selectbox("LLM", [idx for idx, _, _ in llm_options], index=next((i for i, (idx, _, _) in enumerate(llm_options) if idx == current_idx), 0), format_func=llm_labels.get, label_visibility="collapsed", key="sidebar_llm_select")
     if selected_idx != current_idx:
-        agent.next_llm(selected_idx); st.rerun(scope="fragment")
+        agent.next_llm(selected_idx)
+        try:
+            st.rerun(scope="fragment")
+        except st.StreamlitAPIException:
+            st.rerun()
     if st.button(T('force_stop')):
         agent.abort(); st.toast("Stop signal sended"); st.rerun()
     if st.button(T('desktop_pet')):