import streamlit as st
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
from sklearn.metrics import mean_squared_error
from datetime import date

# Set page configuration
st.set_page_config(page_title="Air Quality Monitoring Dashboard", layout="wide")

# Sidebar for navigation
st.sidebar.title("Air Quality Monitoring")
dashboard_choice = st.sidebar.radio("Select Dashboard:", ("User Dashboard", "Admin Dashboard"))

# WHO Guidelines for display in both dashboards
who_guidelines = {
    "Pollutant": ["NO2", "O3"],
    "Max Safe Concentration (µg/m³)": [25, 100]
}

# Data Simulation for Current Day and Predictions
today_data = {
    "NO2 (µg/m³)": np.random.uniform(20, 60),  # Simulated data
    "O3 (µg/m³)": np.random.uniform(50, 120)   # Simulated data
}

# Simulated predictions for the next three days
next_three_days = pd.DataFrame({
    "Day": ["Day 1", "Day 2", "Day 3"],
    "NO2 (µg/m³)": np.random.uniform(20, 60, 3),
    "O3 (µg/m³)": np.random.uniform(50, 120, 3)
})

# Simulated actual and predicted data for Admin Dashboard
y_test = pd.DataFrame({
    "NO2": np.random.uniform(20, 60, 100),
    "O3": np.random.uniform(50, 120, 100)
})
y_test_pred = pd.DataFrame({
    "NO2": np.random.uniform(20, 60, 100),
    "O3": np.random.uniform(50, 120, 100)
})

### User Dashboard ###
if dashboard_choice == "User Dashboard":
    st.title("🌍 User Dashboard: Air Quality Monitoring")

    # Current Day Pollutant Concentrations
    st.sidebar.markdown(f"Today's Date: **{date.today().strftime('%B %d, %Y')}**")
    st.sidebar.markdown("### Current Pollutant Concentrations")
    st.sidebar.write(pd.DataFrame(today_data, index=["Concentration"]))

    st.sidebar.markdown("### WHO Guidelines")
    st.sidebar.write(pd.DataFrame(who_guidelines))

    # WHO guideline alerts
    if today_data["NO2 (µg/m³)"] > 25:
        st.sidebar.error("⚠️ NO2 levels are above WHO guidelines!")
    else:
        st.sidebar.success("✅ NO2 levels are within safe limits.")

    if today_data["O3 (µg/m³)"] > 100:
        st.sidebar.error("⚠️ O3 levels are above WHO guidelines!")
    else:
        st.sidebar.success("✅ O3 levels are within safe limits.")

    # Predictions for the next three days
    st.write("### Predictions for the Next 3 Days")
    st.dataframe(next_three_days)

    # Visualization: Prediction vs WHO Guidelines
    st.write("### Visualization: Predictions vs WHO Guidelines")
    fig, ax = plt.subplots(figsize=(8, 4))
    ax.plot(next_three_days["Day"], next_three_days["NO2 (µg/m³)"], label="NO2", marker='o', color="red")
    ax.axhline(y=25, color="red", linestyle='--', label="NO2 WHO Limit (25 µg/m³)")
    ax.plot(next_three_days["Day"], next_three_days["O3 (µg/m³)"], label="O3", marker='x', color="blue")
    ax.axhline(y=100, color="blue", linestyle='--', label="O3 WHO Limit (100 µg/m³)")

    ax.set_title("Predicted Pollutant Levels vs WHO Guidelines")
    ax.set_xlabel("Days")
    ax.set_ylabel("Concentration (µg/m³)")
    ax.legend()
    st.pyplot(fig)

### Admin Dashboard ###
elif dashboard_choice == "Admin Dashboard":
    st.title("🔧 Admin Dashboard: Detailed Monitoring & Model Performance")

    # Model selection for monitoring (simulated)
    model_name = st.sidebar.selectbox(
        "Select model for monitoring:",
        ("Decision Tree", "Random Forest", "XGBoost")
    )

    # Displaying prediction and actual comparison
    st.write("### Predicted vs Actual Values")

    col1, col2 = st.columns(2)
    with col1:
        st.write("**Actual Values**")
        st.dataframe(y_test)

    with col2:
        st.write("**Predicted Values**")
        st.dataframe(y_test_pred)

    # Performance Metrics
    mse_no2 = mean_squared_error(y_test["NO2"], y_test_pred["NO2"])
    mse_o3 = mean_squared_error(y_test["O3"], y_test_pred["O3"])
    rmse_no2 = np.sqrt(mse_no2)
    rmse_o3 = np.sqrt(mse_o3)

    st.markdown("### Model Performance Metrics")
    col3, col4 = st.columns(2)
    col3.metric("NO2 - Mean Squared Error (MSE)", f"{mse_no2:.2f}")
    col4.metric("O3 - Mean Squared Error (MSE)", f"{mse_o3:.2f}")
    col3.metric("NO2 - Root Mean Squared Error (RMSE)", f"{rmse_no2:.2f}")
    col4.metric("O3 - Root Mean Squared Error (RMSE)", f"{rmse_o3:.2f}")

    # Visualization of Actual vs Predicted Values
    st.markdown("### Visualization: Actual vs Predicted Values")
    fig, ax = plt.subplots(figsize=(10, 5))
    ax.plot(y_test["NO2"], label="Actual NO2", color="blue", marker='o')
    ax.plot(y_test_pred["NO2"], label="Predicted NO2", color="red", linestyle='--')
    ax.set_title(f"NO2 Actual vs Predicted - {model_name}")
    ax.set_xlabel("Samples")
    ax.set_ylabel("Concentration (µg/m³)")
    ax.legend()

    st.pyplot(fig)

    # Feature Importance (simulated example)
    st.markdown("### Feature Importance")
    feature_importances = pd.Series([0.3, 0.2, 0.5], index=["Feature 1", "Feature 2", "Feature 3"])

    fig, ax = plt.subplots(figsize=(8, 4))
    feature_importances.plot(kind="barh", ax=ax)
    ax.set_title(f"Feature Importance - {model_name}")
    st.pyplot(fig)

# say bye
st.sidebar.write("### Thank you for using our app!")