Update app.py

app.py

@@ -27,10 +27,10 @@ st.markdown("Predict 30-day readmission risk with LLM explanations powered by SH
 with st.sidebar:
     st.caption(f"🧠 Model Version: `{MODEL_VERSION}`")
     model_choice = st.selectbox("Choose LLM for explanation", [
+        "meta-llama/Llama-3.2-3B-Instruct",
         "deepcogito/cogito-v1-preview-llama-3B",
         "microsoft/Phi-4-mini-instruct",
         "deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B",
-        "meta-llama/Llama-3.2-3B-Instruct",
         "TheBloke/Mistral-7B-Instruct-v0.1-GGUF",
     ])
 
@@ -75,31 +75,30 @@ st.markdown(f"### Prediction Result: {pred_label}")
 st.markdown(f"**Predicted Probability:** `{pred_proba:.2%}`")
 st.markdown(f"**LLM Model Used:** `{model_choice}`")
 
-# LLM Explanation button
-if st.button("Generate Explanation"):
-    with st.spinner(f"Generating explanation with {model_choice}..."):
-        try:
-            explanation = explain_prediction(
-                patient_id=patient["Patient_ID"],
-                patient_data=patient.to_dict(),
-                model_name=model_choice,
-                _model=model,
-                _client=llm
-            )
-            st.success("### LLM Explanation\n" + explanation)
-        except Exception as e:
-            fallback = f"Unable to generate LLM explanation due to error: {e}"
-            st.warning(f"⚠️ {fallback}")
-            explanation = fallback
-
-        st.session_state["log_df"] = log_explanation(
-            st.session_state["log_df"],
+# LLM Explanation auto-update
+with st.spinner(f"Generating explanation with {model_choice}..."):
+    try:
+        explanation = explain_prediction(
             patient_id=patient["Patient_ID"],
+            patient_data=patient.to_dict(),
             model_name=model_choice,
-            prediction=pred_proba,
-            shap_summary="SHAP summary internal only",
-            explanation=explanation
+            _model=model,
+            _client=llm
         )
+        st.success("### LLM Explanation\n" + explanation)
+    except Exception as e:
+        fallback = f"Unable to generate LLM explanation due to error: {e}"
+        st.warning(f"⚠️ {fallback}")
+        explanation = fallback
+
+    st.session_state["log_df"] = log_explanation(
+        st.session_state["log_df"],
+        patient_id=patient["Patient_ID"],
+        model_name=model_choice,
+        prediction=pred_proba,
+        shap_summary="SHAP summary internal only",
+        explanation=explanation
+    )
 
 # Log download
 if not st.session_state["log_df"].empty: