src/streamlit_app.py

import streamlit as st
import numpy as np
import pandas as pd
import joblib

# Set page config
st.set_page_config(page_title="Stress Detection using One-Class SVM", layout="centered")

# Custom CSS for background and styles
st.markdown(
    """
    <style>
    .stApp {
        background-image: url("https://i.postimg.cc/vZb3ymYT/360-F-1375669005-ebg3mldxps5-ZYr-QFl-Y6-EX3e-CINw-VDeo-F.jpg");
        background-size: cover;
        background-repeat: no-repeat;
        background-attachment: fixed;
    }

    .block-container {
        background-color: rgba(0, 0, 0, 0.6);
        color: white;
        padding: 2rem;
        border-radius: 15px;
        max-width: 800px;
        margin: 2rem auto;
        box-shadow: 0 8px 32px rgba(0, 0, 0, 0.5);
        backdrop-filter: blur(8px);
        -webkit-backdrop-filter: blur(8px);
        border: 1px solid rgba(255, 255, 255, 0.1);
    }

    .sensor-row {
        display: flex;
        justify-content: space-around;
        font-size: 1.1rem;
        margin-top: 1.2rem;
        margin-bottom: 1rem;
    }

    .sensor-row > div {
        padding: 0.5rem 1rem;
        background-color: rgba(255, 255, 255, 0.1);
        border-radius: 8px;
    }

    .scroll-box {
        max-height: 400px;
        overflow-y: auto;
        border: 1px solid #ccc;
        padding: 1rem;
        background-color: rgba(255,255,255,0.05);
        border-radius: 10px;
    }

    h1, h2, h3, p, div {
        color: white !important;
    }

    section.main > div:first-child {
        padding-top: 0rem;
    }

    header[data-testid="stHeader"] {
        height: 0rem;
        visibility: hidden;
    }
    </style>
    """,
    unsafe_allow_html=True
)

st.title("Stress Detection")
st.markdown("Select a mode to detect stress from sensor readings:")

# Load model and scaler
try:
    model = joblib.load("one_class_svm_stress_model.pkl")
    scaler = joblib.load("scaler.pkl")
except Exception as e:
    st.error(f"Error loading model or scaler: {e}")
    st.stop()

# Load or create hardcoded dataset
try:
    df = pd.read_csv("src/simulated_stress_data.csv")
    df = df[['HR', 'HRV', 'EDA']].head(100)
except:
    df = pd.DataFrame({
        "HR": np.random.randint(60, 120, 100),
        "HRV": np.random.uniform(20, 80, 100),
        "EDA": np.random.uniform(0.1, 5.0, 100)
    })

# Radio button selection
mode = st.radio("Choose input mode:", ["Manual Readings", "Generate Readings", "Test Dataset"], horizontal=True)

# Manual Input
if mode == "Manual Readings":
    hr = st.number_input("Heart Rate (HR)", min_value=60, max_value=120, value=80)
    hrv = st.number_input("Heart Rate Variability (HRV)", min_value=20.0, max_value=80.0, value=50.0)
    eda = st.number_input("Electrodermal Activity (EDA)", min_value=0.1, max_value=5.0, value=2.0)

    if st.button("Predict"):
        sample = np.array([[hr, hrv, eda]])
        scaled = scaler.transform(sample)
        pred = model.predict(scaled)
        label = "Stress" if pred[0] == -1 else "No Stress"

        st.markdown(
            f"""
            <div class="sensor-row">
                <div><strong>HR:</strong> {hr} bpm</div>
                <div><strong>HRV:</strong> {hrv:.2f} ms</div>
                <div><strong>EDA:</strong> {eda:.2f} µS</div>
            </div>
            """, unsafe_allow_html=True
        )

        st.subheader("Prediction")
        if label == "No Stress":
            st.success(label)
        else:
            st.error(label)

# Generate Random Input
elif mode == "Generate Readings":
    if st.button("Generate and Predict"):
        hr = np.random.randint(60, 120)
        hrv = np.random.uniform(20, 80)
        eda = np.random.uniform(0.1, 5.0)

        sample = np.array([[hr, hrv, eda]])
        scaled = scaler.transform(sample)
        pred = model.predict(scaled)
        label = "Stress" if pred[0] == -1 else "No Stress"

        st.markdown(
            f"""
            <div class="sensor-row">
                <div><strong>HR:</strong> {hr} bpm</div>
                <div><strong>HRV:</strong> {hrv:.2f} ms</div>
                <div><strong>EDA:</strong> {eda:.2f} µS</div>
            </div>
            """, unsafe_allow_html=True
        )

        st.subheader("Prediction")
        if label == "No Stress":
            st.success(label)
        else:
            st.error(label)


# Test Dataset (Scrollable)

elif mode == "Test Dataset":
    st.markdown("### Select a row from test dataset for prediction:")

    if "page" not in st.session_state:
        st.session_state.page = 0
    if "last_prediction" not in st.session_state:
        st.session_state.last_prediction = None
        st.session_state.last_row = None

    rows_per_page = 5
    total_pages = max(1, (len(df) - 1) // rows_per_page + 1)

    st.markdown("""
        <style>
        .scrollable-table {
            max-height: 350px;
            overflow-y: auto;
            padding: 10px;
            background-color: rgba(255,255,255,0.05);
            border-radius: 10px;
            border: 1px solid #ccc;
        }
        .row-box {
            border-radius: 8px;
            padding: 6px;
            margin-bottom: 4px;
            background-color: rgba(0, 128, 255, 0.15);
            height: 40px;
            display: flex;
            align-items: center;
            transition: background-color 0.3s ease;
        }
        .row-box:hover {
            background-color: rgba(0, 128, 255, 0.3);
        }
        </style>
    """, unsafe_allow_html=True)

    with st.container():
        col1, col2, col3, col4 = st.columns([3, 3, 3, 2])
        col1.markdown("**HR**")
        col2.markdown("**HRV**")
        col3.markdown("**EDA**")
        col4.markdown("**Predict**")

        page_data = df.iloc[
            st.session_state.page * rows_per_page : (st.session_state.page + 1) * rows_per_page
        ]

        for idx, row in page_data.iterrows():
            col1, col2, col3, col4 = st.columns([3, 3, 3, 2])
            with col1:
                st.markdown(f'<div class="row-box">{row["HR"]}</div>', unsafe_allow_html=True)
            with col2:
                st.markdown(f'<div class="row-box">{row["HRV"]:.2f}</div>', unsafe_allow_html=True)
            with col3:
                st.markdown(f'<div class="row-box">{row["EDA"]:.2f}</div>', unsafe_allow_html=True)
            with col4:
                if st.button("Select", key=f"select_{idx}"):
                    sample = np.array([[row['HR'], row['HRV'], row['EDA']]])
                    sample_scaled = scaler.transform(sample)
                    pred = model.predict(sample_scaled)
                    label = "Stress" if pred[0] == -1 else "No Stress"

                    st.session_state.last_prediction = label
                    st.session_state.last_row = row

    col1, col2, col3 = st.columns([1.2, 1.2, 3])
    with col1:
        if st.button("⬅️ Previous"):
            if st.session_state.page > 0:
                st.session_state.page -= 1

    with col2:
        if st.button("Next ➡️"):
            if st.session_state.page < total_pages - 1:
                st.session_state.page += 1

    with col3:
        st.markdown(
            f"<div style='text-align:left; margin-top: 0.5rem;'>Page {st.session_state.page + 1} of {total_pages}</div>",
            unsafe_allow_html=True
    )

    if st.session_state.last_prediction and st.session_state.last_row is not None:
        row = st.session_state.last_row
        label = st.session_state.last_prediction

        st.markdown("---")
        st.markdown("### Prediction")
        st.markdown(
            f"""
            <div class="sensor-row">
                <div><strong>HR:</strong> {row['HR']} bpm</div>
                <div><strong>HRV:</strong> {row['HRV']:.2f} ms</div>
                <div><strong>EDA:</strong> {row['EDA']:.2f} µS</div>
            </div>
            """, unsafe_allow_html=True
        )

        if label == "No Stress":
            st.success(label)
        else:
            st.error(label)