Update app.py

app.py

@@ -1,28 +1,67 @@
 import gradio as gr
-import joblib
 import numpy as np
+import joblib
+import os
+import pandas as pd
+
+from evidently.report import Report
+from evidently.metrics import DataDriftPreset
 
 # Load trained model
-model = joblib.load("model/house_price_model.pkl")
+model_path = "model/house_price_model.pkl"
+if not os.path.exists(model_path):
+    raise FileNotFoundError("Model file not found. Make sure 'model/house_price_model.pkl' exists.")
+model = joblib.load(model_path)
+
+# Sample reference and current data (fake for demo, replace with real if available)
+reference_data = pd.DataFrame({
+    "RM": np.random.normal(6, 0.5, 100),
+    "LSTAT": np.random.normal(12, 5, 100),
+    "PTRATIO": np.random.normal(18, 1.5, 100),
+})
+current_data = pd.DataFrame({
+    "RM": np.random.normal(6.2, 0.6, 100),
+    "LSTAT": np.random.normal(14, 5, 100),
+    "PTRATIO": np.random.normal(19, 2, 100),
+})
+
+# Generate drift report and save it
+def generate_drift_report():
+    report = Report(metrics=[DataDriftPreset()])
+    report.run(reference_data=reference_data, current_data=current_data)
+    report.save_html("evidently_report.html")
+    return "evidently_report.html"
 
-# Define prediction function
-def predict_price(rm, lstat, ptratio):
+# Predict function
+def predict(rm, lstat, ptratio):
     input_data = np.array([[rm, lstat, ptratio]])
     prediction = model.predict(input_data)[0]
     return round(prediction, 2)
 
-# Define Gradio interface
-iface = gr.Interface(
-    fn=predict_price,
-    inputs=[
-        gr.Slider(3.0, 9.0, step=0.1, label="Average Rooms (RM)"),
-        gr.Slider(1.0, 40.0, step=0.1, label="Lower Status Population (%) (LSTAT)"),
-        gr.Slider(12.0, 22.0, step=0.1, label="Pupil-Teacher Ratio (PTRATIO)")
-    ],
-    outputs=gr.Number(label="Predicted House Price (in $1000s)"),
-    title="🏠 Boston Housing Price Predictor",
-    description="Enter the house features to predict the price using a linear regression model.",
-)
-
-# Launch app
-iface.launch()
+# Wrapper to return the HTML drift report
+def show_report():
+    if not os.path.exists("evidently_report.html"):
+        generate_drift_report()
+    with open("evidently_report.html", "r", encoding="utf-8") as f:
+        html_content = f.read()
+    return gr.HTML(html_content)
+
+# Gradio UI setup
+with gr.Blocks() as demo:
+    gr.Markdown("# 🏡 Boston Housing Price Predictor + Drift Report")
+
+    with gr.Tab("🔮 Predict Price"):
+        rm = gr.Slider(3.0, 9.0, step=0.1, label="Average Rooms (RM)")
+        lstat = gr.Slider(1.0, 40.0, step=0.1, label="Lower Status Population (%) (LSTAT)")
+        ptratio = gr.Slider(12.0, 22.0, step=0.1, label="Pupil-Teacher Ratio (PTRATIO)")
+        predict_btn = gr.Button("Predict")
+        result = gr.Number(label="🏠 Predicted Price ($1000s)")
+        predict_btn.click(predict, inputs=[rm, lstat, ptratio], outputs=result)
+
+    with gr.Tab("📊 View Data Drift Report"):
+        report_output = gr.HTML()
+        generate_btn = gr.Button("Generate Report")
+        generate_btn.click(fn=show_report, outputs=report_output)
+
+if __name__ == "__main__":
+    demo.launch()