Update app.py

app.py

@@ -208,19 +208,41 @@ def extract_feature_timeseries(video_path:str, frame_skip:int=FRAME_SKIP, debug:
 # Predictor & event logic
 # =====================================================
 def predict_scores(df):
+    """Predict illumination likelihoods using AutoGluon regression ensemble."""
     feats = ["red_ratio", "green_ratio", "red_diff", "green_diff", "z_red", "z_green"]
     X = df[feats].copy()
     ag = ag_predictor()
+
+    # Get model list (older AG versions have .model_names(), newer have .model_names)
     try:
-        proba = ag.predict_proba(X)
-        if isinstance(proba, pd.DataFrame) and (1 in proba.columns):
-            return proba[1]
-    except Exception as e:
-        print("[WARN] Predictor .predict_proba() failed, falling back:", e)
-        preds = ag.predict(X)
-        s = pd.Series(preds).astype(float)
-        rng = (s.quantile(0.95) - s.quantile(0.05)) or 1.0
-        return ((s - s.quantile(0.05)) / rng).clip(0, 1)
+        models_all = ag.model_names()
+    except Exception:
+        models_all = ag.model_names
+    print(f"[INFO] Evaluating models: {models_all}")
+
+    preds_total = []
+    for m in models_all:
+        try:
+            print(f"[INFO] → Predicting with {m}")
+            y_pred = ag._learner.predict(X, model=m)
+            preds_total.append(pd.Series(y_pred, name=m))
+        except Exception as e:
+            print(f"[WARN] Skipping model {m}: {e}")
+
+    if not preds_total:
+        print("[ERROR] No usable models, returning zeros.")
+        return pd.Series(np.zeros(len(df)))
+
+    # Average predictions from working models
+    y_mean = pd.concat(preds_total, axis=1).mean(axis=1)
+
+    # Normalize 0–1 for consistency
+    rng = (y_mean.quantile(0.95) - y_mean.quantile(0.05)) or 1.0
+    score = ((y_mean - y_mean.quantile(0.05)) / rng).clip(0, 1)
+
+    print(f"[INFO] Used {len(preds_total)} valid submodels for regression scoring.")
+    return score
+
 
 def pick_events(df,score,fps):
     z=rolling_z(score,45); strong=(z>4.0); keep=strong.rolling(3,min_periods=1).sum()>=2