app.py

import streamlit as st
import numpy as np
import pickle


try:
    with open("final_model.pkl", "rb") as f:
        model = pickle.load(f)
    st.success("✅ Model loaded successfully!")
except FileNotFoundError:
    st.error("❌ Model file not found! Please upload `final_model.pkl`.")
    model = None


st.markdown(
    """
    <style>
        body {
            background-color: #121212;
            color: #FFFFFF;
        }
        .stApp {
            background-color: #121212;
            color: #FFFFFF;
        }
        .title {
            text-align: center;
            font-size: 36px;
            font-weight: bold;
            color: #BB86FC;
        }
        .stButton > button {
            width: 100%;
            background-color: #6200EE;
            color: white;
            font-size: 18px;
            border-radius: 8px;
            padding: 10px;
            transition: 0.3s ease-in-out;
        }
        .stButton > button:hover {
            background-color: #3700B3;
            transform: scale(1.05);
        }
        .result-box {
            text-align: center;
            font-size: 20px;
            font-weight: bold;
            color: white;
            padding: 12px;
            border-radius: 8px;
            margin-top: 20px;
            box-shadow: 0px 4px 8px rgba(255, 255, 255, 0.2);
            transition: 0.3s ease-in-out;
        }
        .result-box:hover {
            transform: scale(1.05);
        }
        .high-risk {
            background-color: #D32F2F;
        }
        .low-risk {
            background-color: #388E3C;
        }
        .probability {
            background-color: #FFA726;
        }
    </style>
    """,
    unsafe_allow_html=True,
)


st.markdown("<h1 class='title'>💜 Heart Disease Risk Predictor</h1>", unsafe_allow_html=True)


with st.expander("🔹 **Personal Information**", expanded=True):
    Age = st.slider("Age", 18, 80, 40)
    Sex = st.selectbox("Sex", ["Male", "Female"])

with st.expander("🔹 **Medical Information**", expanded=True):
    RestingBP = st.number_input("Resting Blood Pressure", min_value=80, max_value=200, value=120)
    Cholesterol = st.number_input("Cholesterol Level", min_value=50, max_value=400, value=200)
    FastingBS = 0  
    MaxHR = st.number_input("Maximum Heart Rate", min_value=66, max_value=250, value=202)
    Oldpeak = st.number_input("Oldpeak (ST Depression)", min_value=-2.0, max_value=4.0, value=0.60)
    ChestPainType = st.selectbox("Chest Pain Type", ["ASY", "NAP", "ATA", "TA"])
    RestingECG = st.selectbox("Resting ECG", ["Normal", "LVH", "ST"])
    ExerciseAngina = st.selectbox("Exercise Induced Angina", ["No", "Yes"])
    ST_Slope = st.selectbox("ST Slope", ["Flat", "Up", "Down"])

if st.button("🔍 Predict Risk"):
    try:
        input_data = [[Age, Sex, ChestPainType, RestingBP, Cholesterol, FastingBS,
                       RestingECG, MaxHR, ExerciseAngina, Oldpeak, ST_Slope]]

        prediction = model.predict(input_data)[0]  
        probability = model.predict_proba(input_data)[0][1] * 100  

        # Determine Risk Category
        if prediction == 1:
            result_text = "❌⚠️ High Risk of Heart Disease!"
            result_class = "high-risk"
        else:
            result_text = "✅ Low Risk of Heart Disease!"
            result_class = "low-risk"

        # Display Result
        st.markdown(
            f"<div class='result-box {result_class}'> {result_text} </div>",
            unsafe_allow_html=True,
        )

        # Display Probability Score
        st.markdown(
            f"<div class='result-box probability'> 📊 Risk Probability: {probability:.2f}% </div>",
            unsafe_allow_html=True,
        )

    except Exception as e:
        st.error(f"⚠️ Prediction failed: {e}")