Update src/streamlit_app.py

src/streamlit_app.py

@@ -24,13 +24,29 @@ from sklearn.metrics import mean_squared_error, r2_score
 
 # SHAP
 import shap
+
+
+# --- Safe defaults for Streamlit session state ---
+defaults = {
+    "llm_result": None,
+    "automl_summary": {},
+    "shap_recommendations": [],
+    "hf_clicked": False,
+    "hf_ran_once": False,
+    "run_automl_clicked": False,
+}
+for k, v in defaults.items():
+    st.session_state.setdefault(k, v)
+
+
 if "llm_result" not in st.session_state:
     st.session_state["llm_result"] = None
 if "automl_summary" not in st.session_state:
     st.session_state["automl_summary"] = {}
 if "shap_recommendations" not in st.session_state:
     st.session_state["shap_recommendations"] = []
-
+if "hf_clicked" not in st.session_state:
+    st.session_state["hf_clicked"] = False
 
 # -------------------------
 # Config & paths
@@ -667,8 +683,13 @@ with tabs[4]:
         return {"model_obj": model, "cv_score": score, "best_params": best, "family": family_name, "study": study}
 
     # --- Run tuning across available families (user triggered) ---
-    run_btn = st.button(" Run expanded AutoML + Stacking")
-    if run_btn:
+    if "run_automl_clicked" not in st.session_state:
+        st.session_state["run_automl_clicked"] = False
+    
+    if st.button("Run expanded AutoML + Stacking"):
+        st.session_state["run_automl_clicked"] = True
+    
+    if st.session_state["run_automl_clicked"]:
         log("AutoML + Stacking initiated.")
         with st.spinner("Tuning multiple families (this may take a while depending on choices)..."):
             families_to_try = ["RandomForest", "ExtraTrees", "MLP"]
@@ -937,7 +958,7 @@ with tabs[4]:
                                     recommended_shift[name] *= 0.97  # -3%
                 
                         # Delta table
-                        st.markdown("### 🧾 Shift Adjustment Summary (vs Previous 200 Samples)")
+                        st.markdown("###  Shift Adjustment Summary (vs Previous 200 Samples)")
                         deltas = pd.DataFrame({
                             "Current Avg": prev_shift,
                             "Suggested": recommended_shift,
@@ -1052,68 +1073,79 @@ with tabs[4]:
                 else:
                     st.session_state["shap_recommendations"] = recommendations
                 
-                # --- AI Recommendation Assistant (in-memory safe for Hugging Face) ---
+                # --- AI Recommendation Assistant ---
                 st.markdown("---")
-                st.subheader("AI Recommendation Assistant (in-memory mode)")
+                st.subheader("AI Recommendation Assistant ")
                 st.caption("Generates quick local AI suggestions — no file writes required.")
                 
+                # Create or reset button states safely
+                if "hf_clicked" not in st.session_state:
+                    st.session_state["hf_clicked"] = False
                 if "llm_result" not in st.session_state:
                     st.session_state["llm_result"] = None
                 
-                if st.button("Get AI Recommendation (via HF API)", key="ai_reco"):
+                # --- Buttons ---
+                col1, col2 = st.columns(2)
+                # Click handlers with isolated session flags
+                if col1.button("Get AI Recommendation (via HF API)", key="ai_reco"):
+                    st.session_state["hf_clicked"] = True
+                    st.session_state["hf_ran_once"] = False  # reset internal control
+                
+                if col2.button("Reset Recommendation Output"):
+                    st.session_state["hf_clicked"] = False
+                    st.session_state["llm_result"] = None
+                    st.session_state["hf_ran_once"] = False
+                    st.info("Recommendation output cleared.")
+                
+                # Execute API call only once
+                if st.session_state["hf_clicked"] and not st.session_state.get("hf_ran_once", False):
                     summary = st.session_state.get("automl_summary", {})
                     if not summary:
                         st.warning("Please run AutoML first to generate context.")
-                        st.stop()
-                    try:
-                        import requests, json
-                        st.info("Contacting Hugging Face Inference API (Mixtral-8x7B-Instruct)…")
-                
-                        API_URL = "https://api-inference.huggingface.co/models/mistralai/Mixtral-8x7B-Instruct-v0.1"
-                        headers = {"Authorization": f"Bearer {st.secrets['HF_TOKEN']}"}
-                
-                        prompt = f"""
-                        You are an ML model tuning advisor.
-                        Based on this AutoML summary, suggest 3 concise, actionable steps
-                        to improve model performance if overfitting, underfitting, or data-quality issues are observed.
+                    else:
+                        try:
+                            import requests, json
+                            st.info("Contacting Hugging Face Inference API (Mixtral-8x7B-Instruct)…")
                 
-                        Use case: {summary.get('use_case')}
-                        Target: {summary.get('target')}
-                        Final R²: {summary.get('final_r2')}
-                        Final RMSE: {summary.get('final_rmse')}
-                        Leaderboard: {summary.get('leaderboard')}
-                        """
+                            API_URL = "https://api-inference.huggingface.co/models/mistralai/Mixtral-8x7B-Instruct-v0.1"
+                            headers = {"Authorization": f"Bearer {st.secrets['HF_TOKEN']}"}
+                            prompt = f"""
+                            You are an ML model tuning advisor.
+                            Based on this AutoML summary, suggest 3 concise, actionable steps
+                            to improve model performance if overfitting, underfitting, or data-quality issues are observed.
                 
-                        payload = {
-                            "inputs": prompt,
-                            "parameters": {"max_new_tokens": 200, "temperature": 0.7}
-                        }
+                            Use case: {summary.get('use_case')}
+                            Target: {summary.get('target')}
+                            Final R²: {summary.get('final_r2')}
+                            Final RMSE: {summary.get('final_rmse')}
+                            Leaderboard: {summary.get('leaderboard')}
+                            """
                 
-                        response = requests.post(API_URL, headers=headers, json=payload, timeout=60)
-                        response.raise_for_status()
-                        result = response.json()
+                            payload = {"inputs": prompt, "parameters": {"max_new_tokens": 200, "temperature": 0.7}}
+                            response = requests.post(API_URL, headers=headers, json=payload, timeout=60)
+                            response.raise_for_status()
+                            result = response.json()
                 
-                        if isinstance(result, list) and "generated_text" in result[0]:
-                            text = result[0]["generated_text"]
-                        elif isinstance(result, dict) and "generated_text" in result:
-                            text = result["generated_text"]
-                        else:
-                            text = json.dumps(result, indent=2)
+                            if isinstance(result, list) and "generated_text" in result[0]:
+                                text = result[0]["generated_text"]
+                            elif isinstance(result, dict) and "generated_text" in result:
+                                text = result["generated_text"]
+                            else:
+                                text = json.dumps(result, indent=2)
                 
-                        st.session_state["llm_result"] = text.strip()
-                        log("HF API recommendation generated successfully.")
-                        st.success("AI Recommendation (Mixtral-8x7B-Instruct):")
-                        st.markdown(st.session_state["llm_result"])
+                            st.session_state["llm_result"] = text.strip()
+                            st.session_state["hf_ran_once"] = True
+                            st.success("✅ AI Recommendation (Mixtral-8x7B-Instruct):")
+                            st.markdown(st.session_state["llm_result"])
                 
-                    except Exception as e:
-                        err_msg = f"HF Inference API call failed: {e}"
-                        st.error(err_msg)
-                        log(err_msg)
+                        except Exception as e:
+                            st.error(f"HF Inference API call failed: {e}")
 
                 
-                # Persist output even after rerun
+                
+                # --- Always display cached result, even on rerun ---
                 if st.session_state["llm_result"]:
-                    st.success("AI Recommendation (cached):")
+                    st.markdown("### Cached AI Recommendation:")
                     st.markdown(st.session_state["llm_result"])