Update app.py

app.py

@@ -2,9 +2,22 @@ import pandas as pd
 import joblib
 import gradio as gr
 
-# Load model artifact once
-artifact = joblib.load("pue_artifact.joblib")
+# -------------------------------------------------
+# Load model artifact ONCE (must exist in Space repo)
+# -------------------------------------------------
+ARTIFACT_PATH = "pue_artifact.joblib"
+artifact = joblib.load(ARTIFACT_PATH)
 
+FEATURES = artifact["features"]
+ride_model = artifact["ride_model"]
+discount_model = artifact["discount_model"]
+route_model = artifact["route_model"]
+ride_encoder = artifact["ride_encoder"]
+route_encoder = artifact["route_encoder"]
+
+# -------------------------------------------------
+# Prediction Logic
+# -------------------------------------------------
 def predict_pue(
     zone_id, hour, day_of_week, is_weekend,
     weather_factor, event_factor, traffic_index,
@@ -14,65 +27,71 @@ def predict_pue(
 ):
 
     input_data = {
-        "zone_id": int(zone_id),
-        "hour": int(hour),
-        "day_of_week": int(day_of_week),
-        "is_weekend": int(is_weekend),
-        "weather_factor": float(weather_factor),
-        "event_factor": float(event_factor),
-        "traffic_index": float(traffic_index),
-        "distance_km": float(distance_km),
-        "duration_min": float(duration_min),
-        "base_fare": float(base_fare),
-        "avg_distance": float(avg_distance),
-        "avg_duration": float(avg_duration),
-        "avg_fare": float(avg_fare),
-        "discount_usage_rate": float(discount_usage_rate),
-        "total_rides": int(total_rides)
+        "zone_id": zone_id,
+        "hour": hour,
+        "day_of_week": day_of_week,
+        "is_weekend": is_weekend,
+        "weather_factor": weather_factor,
+        "event_factor": event_factor,
+        "traffic_index": traffic_index,
+        "distance_km": distance_km,
+        "duration_min": duration_min,
+        "base_fare": base_fare,
+        "avg_distance": avg_distance,
+        "avg_duration": avg_duration,
+        "avg_fare": avg_fare,
+        "discount_usage_rate": discount_usage_rate,
+        "total_rides": total_rides
     }
 
-    X = pd.DataFrame([input_data])
-    X = X.reindex(columns=artifact["features"], fill_value=0)
+    # Enforce feature order (CRITICAL for ONNX & Java)
+    row = {f: float(input_data.get(f, 0.0)) for f in FEATURES}
+    X = pd.DataFrame([[row[f] for f in FEATURES]], columns=FEATURES)
 
-    ride = artifact["ride_encoder"].inverse_transform(
-        artifact["ride_model"].predict(X)
-    )[0]
+    # Predictions
+    ride_pred = ride_model.predict(X)
+    ride = ride_encoder.inverse_transform(ride_pred)[0]
 
-    discount_prob = artifact["discount_model"].predict_proba(X)[0][1]
+    discount_prob = float(discount_model.predict_proba(X)[0][1])
 
-    route = artifact["route_encoder"].inverse_transform(
-        artifact["route_model"].predict(X)
-    )[0]
+    route_pred = route_model.predict(X)
+    route = route_encoder.inverse_transform(route_pred)[0]
 
     return {
         "Recommended Ride Type": ride,
-        "Discount Probability": round(float(discount_prob), 2),
+        "Discount Probability": round(discount_prob, 2),
         "Preferred Route": route
     }
 
-# Gradio UI
+# -------------------------------------------------
+# Gradio UI (Hugging Face Space)
+# -------------------------------------------------
 app = gr.Interface(
     fn=predict_pue,
     inputs=[
-        gr.Number(label="Zone ID"),
-        gr.Slider(0,23,step=1,label="Hour"),
-        gr.Slider(0,6,step=1,label="Day of Week"),
-        gr.Radio([0,1],label="Weekend"),
-        gr.Slider(1,1.5,label="Weather Factor"),
-        gr.Slider(1,1.6,label="Event Factor"),
-        gr.Slider(0.5,2,label="Traffic Index"),
-        gr.Number(label="Distance (km)"),
-        gr.Number(label="Duration (min)"),
-        gr.Number(label="Base Fare"),
-        gr.Number(label="Avg Distance"),
-        gr.Number(label="Avg Duration"),
-        gr.Number(label="Avg Fare"),
-        gr.Slider(0,1,label="Discount Usage Rate"),
-        gr.Number(label="Total Rides")
+        gr.Number(label="Zone ID", value=50),
+        gr.Slider(0, 23, step=1, label="Hour", value=15),
+        gr.Slider(0, 6, step=1, label="Day of Week", value=5),
+        gr.Radio([0, 1], label="Weekend", value=0),
+        gr.Slider(1.0, 1.5, label="Weather Factor", value=1.1),
+        gr.Slider(1.0, 1.6, label="Event Factor", value=1.2),
+        gr.Slider(0.5, 2.0, label="Traffic Index", value=1.3),
+        gr.Number(label="Distance (km)", value=5),
+        gr.Number(label="Duration (min)", value=10),
+        gr.Number(label="Base Fare", value=50),
+        gr.Number(label="Avg Distance", value=5),
+        gr.Number(label="Avg Duration", value=10),
+        gr.Number(label="Avg Fare", value=120),
+        gr.Slider(0, 1, label="Discount Usage Rate", value=0.3),
+        gr.Number(label="Total Rides", value=15),
     ],
-    outputs="json",
-    title="Personalized User Experience – Real-Time ML",
-    description="Real-time personalization for ride-hailing apps"
+    outputs=gr.JSON(label="Prediction"),
+    title="Personalized User Experience – Real-Time ML (CPU)",
+    description=(
+        "Inference-only Personalized User Experience model. "
+        "Model trained on Hugging Face, artifact-based, "
+        "and ready for ONNX + Java CPU deployment."
+    )
 )
 
 app.launch()