Upload folder using huggingface_hub

app.py

@@ -1,53 +1,3 @@
-import streamlit as st
-import pandas as pd
-import joblib
-import os
-from huggingface_hub import hf_hub_download
-
-# ==========================================
-# 1. Page Configuration & Model Loading
-# ==========================================
-st.set_page_config(page_title="Engine Predictive Maintenance", layout="centered")
-
-# Configuration from previous training steps
-REPO_ID = "dpanchali/predictive_maintenance_model"
-FILENAME = "predictive_maintenance_model.joblib"
-
-@st.cache_resource
-def load_model():
-    """Download and load the model from Hugging Face Hub."""
-    try:
-        # Download the model file from the repository
-        model_path = hf_hub_download(repo_id=REPO_ID, filename=FILENAME)
-        model = joblib.load(model_path)
-        return model
-    except Exception as e:
-        st.error(f"Error loading model from Hugging Face: {e}")
-        return None
-
-# Load the model
-model = load_model()
-
-# ==========================================
-# 2. UI Layout
-# ==========================================
-st.title("Engine Condition Predictor")
-st.markdown("""
-This application uses a trained **XGBoost** model to predict the health of an engine
-based on real-time sensor data.
-""")
-
-st.header("Input Engine Sensor Data")
-st.info("Enter the current readings from the engine sensors below:")
-
-# Creating columns for a cleaner layout
-col1, col2 = st.columns(2)
-
-with col1:
-    engine_rpm = st.number_input("Engine RPM", min_value=0, max_value=10000, value=700, step=10)
-    lub_oil_pressure = st.number_input("Lub Oil Pressure (bar)", min_value=0.0, max_value=20.0, value=2.5, format="%.4f")
-    fuel_pressure = st.number_input("Fuel Pressure (bar)", min_value=0.0, max_value=50.0, value=11.8, format="%.4f")
-
 %%writefile predictive_maintenance_project/deployment/app.py
 import streamlit as st
 import pandas as pd
@@ -60,7 +10,6 @@ from huggingface_hub import hf_hub_download
 # ==========================================
 st.set_page_config(page_title="Engine Predictive Maintenance", layout="centered")
 
-# Configuration from previous training steps
 REPO_ID = "dpanchali/predictive_maintenance_model"
 FILENAME = "predictive_maintenance_model.joblib"
 
@@ -68,120 +17,89 @@ FILENAME = "predictive_maintenance_model.joblib"
 def load_model():
     """Download and load the model from Hugging Face Hub."""
     try:
-        # Download the model file from the repository
         model_path = hf_hub_download(repo_id=REPO_ID, filename=FILENAME)
         model = joblib.load(model_path)
         return model
     except Exception as e:
-        st.error(f"Error loading model from Hugging Face: {e}")
+        st.error(f"Error loading model: {e}")
         return None
 
-# Load the model
 model = load_model()
 
 # ==========================================
 # 2. UI Layout
 # ==========================================
-st.title("Engine Condition Predictor")
+st.title("🚢 Engine Condition Predictor")
 st.markdown("""
-This application uses a trained **XGBoost** model to predict the health of an engine
-based on real-time sensor data.
+This application uses a trained **XGBoost** model to predict engine health 
+based on sensor patterns.
 """)
 
 st.header("Input Engine Sensor Data")
-st.info("Enter the current readings from the engine sensors below:")
 
-# Creating columns for a cleaner layout
 col1, col2 = st.columns(2)
 
 with col1:
-    engine_rpm = st.number_input("Engine RPM", min_value=0, max_value=10000, value=700, step=10)
-    lub_oil_pressure = st.number_input("Lub Oil Pressure (bar)", min_value=0.0, max_value=20.0, value=2.5, format="%.4f")
-    fuel_pressure = st.number_input("Fuel Pressure (bar)", min_value=0.0, max_value=50.0, value=11.8, format="%.4f")
+    engine_rpm = st.number_input("Engine RPM", min_value=0, max_value=10000, value=700)
+    lub_oil_pressure = st.number_input("Lub Oil Pressure (bar)", min_value=0.0, value=2.5, format="%.4f")
+    fuel_pressure = st.number_input("Fuel Pressure (bar)", min_value=0.0, value=11.8, format="%.4f")
 
 with col2:
-    coolant_pressure = st.number_input("Coolant Pressure (bar)", min_value=0.0, max_value=20.0, value=3.2, format="%.4f")
-    lub_oil_temp = st.number_input("Lub Oil Temp (°C)", min_value=0.0, max_value=200.0, value=84.1, format="%.4f")
-    coolant_temp = st.number_input("Coolant Temp (°C)", min_value=0.0, max_value=200.0, value=81.6, format="%.4f")
-
-    load_index = engine_rpm * fuel_pressure / 100 
-    thermal_stress = coolant_temp - lub_oil_temp
+    coolant_pressure = st.number_input("Coolant Pressure (bar)", min_value=0.0, value=3.2, format="%.4f")
+    lub_oil_temp = st.number_input("Lub Oil Temp (°C)", min_value=0.0, value=84.1, format="%.4f")
+    coolant_temp = st.number_input("Coolant Temp (°C)", min_value=0.0, value=81.6, format="%.4f")
 
 # ==========================================
 # 3. Prediction Logic
 # ==========================================
 if st.button("Predict Engine Condition", type="primary"):
     if model is not None:
-        # Prepare input data in the same format as training
-        input_data = pd.DataFrame([{
-            'Engine rpm': engine_rpm,
-            'Lub oil pressure': lub_oil_pressure,
-            'Fuel pressure': fuel_pressure,
-            'Coolant pressure': coolant_pressure,
-            'lub oil temp': lub_oil_temp,
-            'Coolant temp': coolant_temp,
-            'load_index' : load_index,
-            'thermal_stress'  : thermal_stress
-        }])
-
-        # Perform prediction
-        prediction = model.predict(input_data)[0]
-        prediction_proba = model.predict_proba(input_data)[0]
-
-        # Display results
-        st.divider()
-        st.subheader("Results")
-
-        if prediction == 1:
-            st.success("**Prediction: Engine is in Good Condition**")
-        else:
-            st.error("**Prediction: Maintenance Required (Potential Fault Detected)**")
-
-        # Display confidence scores
-        st.write(f"**Confidence Score:** {max(prediction_proba)*100:.2f}%")
-
-        # Optional: Display gauge or progress bar for health
-        health_score = prediction_proba[1] # Probability of class 1 (Good)
-        st.progress(health_score)
-    else:
-        st.warning("Model is not loaded. Please check your Hugging Face credentials and Repo ID.")
-
-# ==========================================
-# 4. Footer
-# ==========================================
-st.sidebar.markdown("### Model Information")
-st.sidebar.text(f"Repo: {REPO_ID}")
-st.sidebar.text(f"File: {FILENAME}")
-st.sidebar.markdown("---")
-st.sidebar.write("This tool is intended for predictive maintenance scheduling based on sensor patterns.")
+        # Calculate engineered features right before prediction
+        load_index = float(engine_rpm * fuel_pressure / 100)
+        thermal_stress = float(coolant_temp - lub_oil_temp)
+
+        # Create DataFrame with EXACT column names and order from your error message
+        input_data = pd.DataFrame([[
+            engine_rpm, 
+            lub_oil_pressure, 
+            fuel_pressure, 
+            coolant_pressure, 
+            lub_oil_temp, 
+            coolant_temp, 
+            load_index, 
+            thermal_stress
+        ]], columns=[
+            'Engine rpm', 'Lub oil pressure', 'Fuel pressure', 'Coolant pressure', 
+            'lub oil temp', 'Coolant temp', 'load_index', 'thermal_stress'
+        ])
 
         # Perform prediction
         prediction = model.predict(input_data)[0]
         prediction_proba = model.predict_proba(input_data)[0]
 
-        # Display results
         st.divider()
         st.subheader("Results")
 
-        if prediction == 1:
-            st.success("**Prediction: Engine is in Good Condition**")
+        if prediction == 0: # Assuming 0 is Good based on typical maintenance datasets
+             st.success("**Status: Engine is in Good Condition**")
         else:
-            st.error("**Prediction: Maintenance Required (Potential Fault Detected)**")
+             st.error("**Status: Maintenance Required (Potential Fault)**")
 
-        # Display confidence scores
+        # Confidence Visual
+        health_score = prediction_proba[0] # Probability of "Healthy"
         st.write(f"**Confidence Score:** {max(prediction_proba)*100:.2f}%")
-
-        # Optional: Display gauge or progress bar for health
-        health_score = prediction_proba[1] # Probability of class 1 (Good)
         st.progress(health_score)
+        
     else:
-        st.warning("Model is not loaded. Please check your Hugging Face credentials and Repo ID.")
+        st.error("Model could not be loaded. Check your Hugging Face Repo ID.")
 
 # ==========================================
-# 4. Footer
+# 4. Sidebar Information
 # ==========================================
-st.sidebar.markdown("### Model Information")
-st.sidebar.text(f"Repo: {REPO_ID}")
-st.sidebar.text(f"File: {FILENAME}")
-st.sidebar.markdown("---")
-st.sidebar.write("This tool is intended for predictive maintenance scheduling based on sensor patterns.")
+with st.sidebar:
+    st.markdown("### Model Details")
+    st.text(f"Repo: {REPO_ID}")
+    st.markdown("---")
+    st.write("Calculated Features:")
+    st.caption(f"Load Index: {engine_rpm * fuel_pressure / 100:.2f}")
+    st.caption(f"Thermal Stress: {coolant_temp - lub_oil_temp:.2f}")