Upload app.py with huggingface_hub

app.py

@@ -0,0 +1,114 @@
+import streamlit as st
+import pandas as pd
+import joblib
+import os
+import sklearn  # Explicit import prevents some joblib errors
+from huggingface_hub import hf_hub_download
+
+# --- 1. PAGE CONFIG (MUST BE FIRST) ---
+st.set_page_config(page_title="Engine Failure Prediction", page_icon="🚛")
+
+# --- CONFIGURATION ---
+HF_USERNAME = os.getenv("HF_USERNAME", "iStillWaters")
+MODEL_REPO_NAME = os.getenv("MODEL_REPO_NAME", "auto_predictive_maintenance_model")
+MODEL_REPO_ID = f"{HF_USERNAME}/{MODEL_REPO_NAME}"
+
+MODEL_FILENAME = "best_engine_model.pkl"
+SCALER_FILENAME = "scaler.joblib"
+
+# CRITICAL: Must match the order in process_data.py exactly!
+EXPECTED_FEATURES = [
+    'Engine rpm', 
+    'Lub oil pressure', 
+    'Fuel pressure', 
+    'Coolant pressure', 
+    'lub oil temp', 
+    'Coolant temp'
+]
+
+# --- LOAD MODEL & SCALER ---
+@st.cache_resource
+def load_artifacts():
+    print(f"Loading artifacts from {MODEL_REPO_ID}...")
+    try:
+        # Download Model
+        model_path = hf_hub_download(repo_id=MODEL_REPO_ID, filename=MODEL_FILENAME)
+        model = joblib.load(model_path)
+        
+        # Download Scaler
+        scaler_path = hf_hub_download(repo_id=MODEL_REPO_ID, filename=SCALER_FILENAME)
+        scaler = joblib.load(scaler_path)
+        
+        return model, scaler
+    except Exception as e:
+        # We cannot use st.error here easily if it's cached, so we print to logs
+        print(f"❌ Error loading artifacts: {e}")
+        return None, None
+
+# Load them now
+model, scaler = load_artifacts()
+
+# --- UI LAYOUT ---
+st.title("🚛 Engine Failure Prediction System")
+st.markdown(f"**Model Source:** `{MODEL_REPO_ID}`")
+st.markdown("Enter real-time sensor data to predict engine health status.")
+
+# --- INPUT FORM ---
+with st.form("prediction_form"):
+    st.subheader("Sensor Telemetry")
+    col1, col2 = st.columns(2)
+    
+    with col1:
+        rpm = st.number_input("Engine RPM", min_value=0, max_value=10000, value=2000)
+        lub_oil_p = st.number_input("Lub Oil Pressure", min_value=0.0, max_value=10.0, value=4.5)
+        fuel_p = st.number_input("Fuel Pressure", min_value=0.0, max_value=20.0, value=7.0)
+    
+    with col2:
+        coolant_p = st.number_input("Coolant Pressure", min_value=0.0, max_value=10.0, value=3.0)
+        lub_oil_t = st.number_input("Lub Oil Temp (°C)", min_value=0.0, max_value=150.0, value=75.0)
+        coolant_t = st.number_input("Coolant Temp (°C)", min_value=0.0, max_value=150.0, value=80.0)
+    
+    submit_button = st.form_submit_button("Predict Engine Status")
+
+# --- PREDICTION LOGIC ---
+if submit_button:
+    if model is None or scaler is None:
+        st.error("Cannot predict: Model or Scaler not loaded. Check HF Space Logs.")
+    else:
+        # 1. Create Dataframe
+        input_data = pd.DataFrame({
+            'Engine rpm': [rpm],
+            'Lub oil pressure': [lub_oil_p],
+            'Fuel pressure': [fuel_p],
+            'Coolant pressure': [coolant_p],
+            'lub oil temp': [lub_oil_t],
+            'Coolant temp': [coolant_t]
+        })
+        
+        # 2. Reorder Columns
+        input_data = input_data[EXPECTED_FEATURES]
+        
+        try:
+            # 3. Scale
+            scaled_data = scaler.transform(input_data)
+            
+            # 4. Predict
+            prediction = model.predict(scaled_data)[0]
+            
+            try:
+                probability = model.predict_proba(scaled_data)[0][1]
+            except:
+                probability = 0.0
+            
+            # 5. Display
+            st.divider()
+            if prediction == 1:
+                st.error(f"🚨 CRITICAL WARNING: Engine Failure Predicted!")
+                st.write(f"**Confidence Level:** {probability:.2%}")
+                st.warning("Recommendation: Stop vehicle immediately and inspect cooling system.")
+            else:
+                st.success(f"✅ SYSTEM NORMAL: Engine is Healthy.")
+                st.write(f"**Failure Probability:** {probability:.2%}")
+                
+        except Exception as e:
+            st.error(f"Prediction Error: {e}")