Update app.py

app.py

@@ -3,68 +3,103 @@ import pandas as pd
 import joblib
 import numpy as np
 import plotly.graph_objects as go
+import os
 
-# Set Page Config
-st.set_page_config(page_title="AI Predictive Maintenance", layout="wide")
+# --- PAGE CONFIGURATION ---
+st.set_page_config(
+    page_title="Jet Engine AI Predictor",
+    page_icon="✈️",
+    layout="wide"
+)
 
-# Load Model
+# --- MODEL LOADING WITH SAFETY CHECK ---
 @st.cache_resource
 def load_model():
-    return joblib.load('engine_model.pkl')
+    model_path = 'engine_model.pkl'
+    if not os.path.exists(model_path):
+        return None
+    return joblib.load(model_path)
 
 model = load_model()
 
-st.title("✈️ Smart Maintenance: Jet Engine RUL Predictor")
-st.markdown("---")
+# --- UI HEADER ---
+st.title("✈️ Jet Engine Predictive Maintenance System")
+st.markdown("""
+This AI model predicts the **Remaining Useful Life (RUL)** of a turbofan engine based on sensor readings. 
+It helps engineers decide when to perform maintenance *before* a failure occurs.
+""")
+
+# Check if model is loaded, if not, show instructions
+if model is None:
+    st.error("⚠️ **Model file 'engine_model.pkl' not found!**")
+    st.info("Please upload the `engine_model.pkl` file you generated locally to the 'Files and versions' tab.")
+    st.stop()
+
+# --- SIDEBAR INPUTS ---
+st.sidebar.header("🛠️ Input Sensor Data")
+st.sidebar.markdown("Adjust the sliders based on engine telemetry:")
+
+# Feature list: ['cycles', 's2', 's3', 's4', 's7', 's8', 's11', 's12', 's13', 's15', 's17', 's20', 's21']
+cycle = st.sidebar.slider("Current Operational Cycle", 1, 350, 100)
+s2 = st.sidebar.slider("Sensor 2 (LPC Outlet Temp)", 640.0, 650.0, 642.5)
+s3 = st.sidebar.slider("Sensor 3 (HPC Outlet Temp)", 1580.0, 1600.0, 1589.0)
+s4 = st.sidebar.slider("Sensor 4 (LPT Outlet Temp)", 1400.0, 1430.0, 1408.0)
+s7 = st.sidebar.slider("Sensor 7 (HPC Outlet Press)", 550.0, 560.0, 553.5)
+s11 = st.sidebar.slider("Sensor 11 (HPC Speed)", 47.0, 48.5, 47.5)
 
-# Layout: 2 Columns
-col1, col2 = st.columns([1, 2])
+# Hidden features (filled with mean values to keep UI clean)
+other_features = [550, 2388, 521, 8.4, 392, 39, 23] 
 
-with col1:
-    st.header("📥 Sensor Inputs")
-    cycle = st.slider("Current Flight Cycles", 1, 350, 100)
-    s2 = st.number_input("Sensor 2 (LPC Outlet Temp)", value=642.0)
-    s3 = st.number_input("Sensor 3 (HPC Outlet Temp)", value=1585.0)
-    s4 = st.number_input("Sensor 4 (LPT Outlet Temp)", value=1405.0)
-    s7 = st.number_input("Sensor 7 (HPC Outlet Press)", value=553.0)
-    s11 = st.number_input("Sensor 11 (HPC Speed)", value=47.5)
+# --- PREDICTION LOGIC ---
+st.markdown("### 🔍 Diagnostic Analysis")
+
+if st.button("Run AI Prediction", type="primary"):
+    # Prepare input array (Must match training features exactly)
+    input_data = np.array([[cycle, s2, s3, s4, s7, 1300, s11, 550, 2388, 521, 8.4, 392, 39, 23, 1]])
     
-    # Static values for remaining features to simplify UI
-    other_features = [550, 2388, 521, 8.4, 392, 39, 23] 
+    # Take only the first 15 features as defined in training
+    prediction = model.predict(input_data[:, :15])
+    rul = max(0, int(prediction[0]))
 
-    if st.button("Analyze Engine Health", type="primary"):
-        inputs = np.array([[cycle, s2, s3, s4, s7, s11] + other_features])
-        prediction = model.predict(inputs)[0]
-        st.session_state['prediction'] = max(0, int(prediction))
+    # --- RESULTS DISPLAY ---
+    col1, col2 = st.columns([1, 1])
 
-with col2:
-    st.header("📊 Diagnostic Results")
-    if 'prediction' in st.session_state:
-        rul = st.session_state['prediction']
-        
-        # 1. Visual Gauge Chart
+    with col1:
+        # Gauge Chart
         fig = go.Figure(go.Indicator(
             mode = "gauge+number",
             value = rul,
-            title = {'text': "Remaining Useful Life (Cycles)"},
+            domain = {'x': [0, 1], 'y': [0, 1]},
+            title = {'text': "Estimated Cycles Remaining", 'font': {'size': 24}},
             gauge = {
-                'axis': {'range': [0, 200]},
-                'bar': {'color': "black"},
-                'steps' : [
+                'axis': {'range': [0, 200], 'tickwidth': 1},
+                'bar': {'color': "darkblue"},
+                'steps': [
                     {'range': [0, 30], 'color': "red"},
-                    {'range': [30, 70], 'color': "yellow"},
-                    {'range': [70, 200], 'color': "green"}],
+                    {'range': [30, 75], 'color': "orange"},
+                    {'range': [75, 200], 'color': "green"}],
+                'threshold': {
+                    'line': {'color': "black", 'width': 4},
+                    'thickness': 0.75,
+                    'value': rul}
             }
         ))
         st.plotly_chart(fig)
 
-        # 2. Status Logic
-        if rul < 30:
-            st.error(f"CRITICAL: Engine failure likely within {rul} cycles. Ground the aircraft immediately!")
-        elif rul < 70:
-            st.warning(f"CAUTION: Maintenance due soon. Estimated life: {rul} cycles.")
+    with col2:
+        st.write("### Engine Health Status")
+        if rul <= 30:
+            st.error(f"🚨 **CRITICAL STATE**\n\nEngine failure predicted within **{rul} cycles**. Maintenance required immediately.")
+        elif rul <= 75:
+            st.warning(f"⚠️ **CAUTION**\n\nEngine showing signs of wear. Estimated life: **{rul} cycles**. Schedule inspection soon.")
         else:
-            st.success(f"HEALTHY: Engine is operating within safe parameters ({rul} cycles remaining).")
+            st.success(f"✅ **HEALTHY**\n\nEngine operating normally. Estimated life: **{rul} cycles**.")
+        
+        st.info("**Note:** RUL (Remaining Useful Life) is an estimate based on simulation data patterns.")
+
+else:
+    st.write("Click the button on the left to analyze the current sensor inputs.")
 
+# --- FOOTER ---
 st.markdown("---")
-st.info("B.Tech AI&DS Special Project: Industrial Time-Series Forecasting")
+st.caption("B.Tech AI & Data Science Special Project | Developed for Industrial Predictive Maintenance")