Update app.py

app.py

@@ -1,85 +1,107 @@
 import gradio as gr
+import pandas as pd
 import joblib
-import numpy as np
 
-# Load the trained Random Forest model
-model = joblib.load("best_random_forest_model.pkl")
+# Load model
+model = joblib.load("random_forest_model.pkl")
+
+# Define the 18 features
+input_features = [
+    "Age", "Gender", "BMI", "Physical Activity Level", "Smoking Status", "Alcohol Consumption",
+    "Diabetes", "Hypertension", "Cholesterol Level", "Family History of Alzheimer’s",
+    "Cognitive Test Score", "Depression Level", "Sleep Quality", "Dietary Habits",
+    "Genetic Risk Factor (APOE-ε4 allele)", "Social Engagement Level", "Income Level", "Stress Levels"
+]
+
+def print_debug_info(input_dict):
+    print("\n🛠️ Input received:")
+    for k, v in input_dict.items():
+        print(f"{k}: {v}")
 
-# Define the prediction function
 def predict_alzheimers(
-    age, gender, bmi, diabetes, hypertension, cholesterol,
-    physical_activity, smoking, alcohol, family_history,
-    cognitive_score, depression, sleep, dietary,
-    genetic_risk, social_engagement, income, stress
+    Age, Gender, BMI, Physical, Smoking, Alcohol, Diabetes, Hypertension, Cholesterol,
+    FamilyHistory, CognitiveScore, Depression, Sleep, Diet, GeneticRisk, Engagement,
+    Income, Stress
 ):
-    # Map categorical values to numerical ones
-    gender = 1 if gender == "Male" else 0
-    diabetes = 1 if diabetes == "Yes" else 0
-    hypertension = 1 if hypertension == "Yes" else 0
-    family_history = 1 if family_history == "Yes" else 0
-    genetic_risk = 1 if genetic_risk == "Yes" else 0
+    input_dict = {
+        "Age": Age,
+        "Gender": Gender,
+        "BMI": BMI,
+        "Physical Activity Level": Physical,
+        "Smoking Status": Smoking,
+        "Alcohol Consumption": Alcohol,
+        "Diabetes": Diabetes,
+        "Hypertension": Hypertension,
+        "Cholesterol Level": Cholesterol,
+        "Family History of Alzheimer’s": FamilyHistory,
+        "Cognitive Test Score": CognitiveScore,
+        "Depression Level": Depression,
+        "Sleep Quality": Sleep,
+        "Dietary Habits": Diet,
+        "Genetic Risk Factor (APOE-ε4 allele)": GeneticRisk,
+        "Social Engagement Level": Engagement,
+        "Income Level": Income,
+        "Stress Levels": Stress
+    }
+
+    print_debug_info(input_dict)
+    input_df = pd.DataFrame([input_dict])
+    input_encoded = pd.get_dummies(input_df)
+
+    model_features = model.feature_names_in_
+    for col in model_features:
+        if col not in input_encoded.columns:
+            input_encoded[col] = 0
+    input_encoded = input_encoded[model_features]
 
-    cholesterol = {"Low": 0, "Normal": 1, "High": 2}.get(cholesterol, 1)
-    physical_activity = {"Low": 0, "Medium": 1, "High": 2}.get(physical_activity, 1)
-    smoking = {"Never": 0, "Former": 1, "Current": 2}.get(smoking, 1)
-    alcohol = {"Never": 0, "Occasionally": 1, "Regularly": 2}.get(alcohol, 1)
-    depression = {"Low": 0, "Medium": 1, "High": 2}.get(depression, 1)
-    sleep = {"Poor": 0, "Average": 1, "Good": 2}.get(sleep, 1)
-    dietary = {"Healthy": 0, "Average": 1, "Unhealthy": 2}.get(dietary, 1)
-    social_engagement = {"Low": 0, "Medium": 1, "High": 2}.get(social_engagement, 1)
-    income = {"Low": 0, "Medium": 1, "High": 2}.get(income, 1)
-    stress = {"Low": 0, "Medium": 1, "High": 2}.get(stress, 1)
+    pred = model.predict(input_encoded)[0]
+    prob = model.predict_proba(input_encoded)[0][1] * 100
 
-    # ... your model input code continues ...
+    result = "🧠 Alzheimer's Detected" if pred == 1 else "✅ No Alzheimer's Detected"
+    return f"{result}  |  Risk Probability: {prob:.2f}%"
 
 
-    # Create input array
-    features = np.array([[
-        age, gender, bmi, diabetes, hypertension, cholesterol,
-        family_history, cognitive_score, genetic_risk,
-        physical_activity, smoking, alcohol, depression,
-        sleep, dietary, social_engagement,
-        income, stress
-    ]])
+with gr.Blocks(theme=gr.themes.Soft()) as demo:
+    gr.Markdown("## 🧠 Alzheimer's Disease Predictor")
+    gr.Markdown("Enter patient details below to predict Alzheimer's risk using a Random Forest model.")
 
-    # Predict
-    prediction = model.predict(features)[0]
-    probability = model.predict_proba(features)[0][1] * 100
+    with gr.Row():
+        with gr.Column():
+            age = gr.Slider(40, 100, step=1, label="Age")
+            gender = gr.Radio(["Male", "Female"], label="Gender")
+            bmi = gr.Slider(15, 40, step=0.1, label="BMI")
+            physical = gr.Radio(["Low", "Medium", "High"], label="Physical Activity Level")
+            smoking = gr.Radio(["Never", "Former", "Current"], label="Smoking Status")
+            alcohol = gr.Radio(["Never", "Occasionally", "Regularly"], label="Alcohol Consumption")
+            diabetes = gr.Radio(["No", "Yes"], label="Diabetes")
+            hypertension = gr.Radio(["No", "Yes"], label="Hypertension")
+            cholesterol = gr.Radio(["Normal", "High", "Low"], label="Cholesterol Level")
 
-    if prediction == 1:
-        result = f"❌ Alzheimer's Detected | Risk Probability: {probability:.2f}%"
-    else:
-        result = f"✅ No Alzheimer's Detected | Risk Probability: {probability:.2f}%"
+        with gr.Column():
+            family = gr.Radio(["No", "Yes"], label="Family History of Alzheimer’s")
+            cognitive = gr.Slider(0, 100, step=1, label="Cognitive Test Score")
+            depression = gr.Radio(["Low", "Medium", "High"], label="Depression Level")
+            sleep = gr.Radio(["Poor", "Average", "Good"], label="Sleep Quality")
+            diet = gr.Radio(["Healthy", "Average", "Unhealthy"], label="Dietary Habits")
+            genetic = gr.Radio(["No", "Yes"], label="Genetic Risk Factor (APOE-ε4 allele)")
+            engagement = gr.Radio(["Low", "Medium", "High"], label="Social Engagement Level")
+            income = gr.Radio(["Low", "Medium", "High"], label="Income Level")
+            stress = gr.Radio(["Low", "Medium", "High"], label="Stress Levels")
 
-    return result
+    with gr.Row():
+        submit_btn = gr.Button("🔍 Predict Alzheimer's Risk", scale=2)
+    
+    output = gr.Textbox(label="Prediction")
 
-# Create the Gradio interface
-app = gr.Interface(
-    fn=predict_alzheimers,
-    inputs=[
-        gr.Slider(40, 100, value=60, label="Age"),
-        gr.Radio(["Male", "Female"], label="Gender"),
-        gr.Slider(15, 40, value=25, label="BMI"),
-        gr.Radio(["Low", "Medium", "High"], label="Physical Activity Level"),
-        gr.Radio(["Never", "Former", "Current"], label="Smoking Status"),
-        gr.Radio(["Never", "Occasionally", "Regularly"], label="Alcohol Consumption"),
-        gr.Radio(["No", "Yes"], label="Diabetes"),
-        gr.Radio(["No", "Yes"], label="Hypertension"),
-        gr.Radio(["Low", "Normal", "High"], label="Cholesterol Level"),
-        gr.Radio(["No", "Yes"], label="Family History of Alzheimer's"),
-        gr.Slider(0, 100, value=50, label="Cognitive Test Score"),
-        gr.Radio(["Low", "Medium", "High"], label="Depression Level"),
-        gr.Radio(["Poor", "Average", "Good"], label="Sleep Quality"),
-        gr.Radio(["Healthy", "Average", "Unhealthy"], label="Dietary Habits"),
-        gr.Radio(["No", "Yes"], label="Genetic Risk Factor (APOE-ε4 allele)"),
-        gr.Radio(["Low", "Medium", "High"], label="Social Engagement Level"),
-        gr.Radio(["Low", "Medium", "High"], label="Income Level"),
-        gr.Radio(["Low", "Medium", "High"], label="Stress Levels"),
-    ],
-    outputs=gr.Textbox(label="Prediction"),
-    title="🔍 Predict Alzheimer's Risk",
-    theme=gr.themes.Soft()
-)
+    submit_btn.click(
+        fn=predict_alzheimers,
+        inputs=[
+            age, gender, bmi, physical, smoking, alcohol, diabetes,
+            hypertension, cholesterol, family, cognitive, depression,
+            sleep, diet, genetic, engagement, income, stress
+        ],
+        outputs=output
+    )
 
-# Launch the app
-app.launch()
+if __name__ == "__main__":
+    demo.launch()