app.py

import streamlit as st
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
from sklearn.linear_model import LinearRegression
import plotly.express as px

# -----------------------------
# CONFIGURATION
# -----------------------------
st.set_page_config(page_title="Smart Energy AI", layout="wide")

st.title("⚡ Smart Energy AI Chatbot")
st.subheader("Energy Consumption Analysis, Prediction & Sustainability Recommendations")

# -----------------------------
# FILE UPLOAD SAFETY FIX
# -----------------------------
uploaded_file = st.file_uploader("Upload your energy consumption CSV", type=["csv"])

try:
    if uploaded_file is not None:
        df = pd.read_csv(uploaded_file)
    else:
        df = pd.read_csv("energy_data.csv")
except Exception as e:
    st.error("⚠️ Dataset not found or invalid file. Please upload a CSV.")
    st.stop()

# -----------------------------
# DATA PREPROCESSING
# -----------------------------
df['date'] = pd.to_datetime(df['date'], errors='coerce')
df['usage_kwh'] = pd.to_numeric(df['usage_kwh'], errors='coerce')
df.dropna(inplace=True)

if df.empty:
    st.warning("No valid data available.")
    st.stop()

# -----------------------------
# SIDEBAR FILTERS
# -----------------------------
st.sidebar.header("Filters")

appliance_list = df['appliance'].dropna().unique()

appliance_filter = st.sidebar.multiselect(
    "Select Appliance",
    options=appliance_list,
    default=appliance_list
)

filtered_df = df[df['appliance'].isin(appliance_filter)]

# -----------------------------
# DAILY ANALYSIS
# -----------------------------
daily_usage = filtered_df.groupby('date')['usage_kwh'].sum().reset_index()

total_consumption = filtered_df['usage_kwh'].sum()
avg_consumption = filtered_df['usage_kwh'].mean()

if not daily_usage.empty:
    peak_day = daily_usage.loc[daily_usage['usage_kwh'].idxmax()]
else:
    peak_day = {"date": pd.Timestamp.today(), "usage_kwh": 0}

# -----------------------------
# METRICS
# -----------------------------
col1, col2, col3 = st.columns(3)
col1.metric("Total Consumption (kWh)", f"{total_consumption:.2f}")
col2.metric("Average Usage (kWh)", f"{avg_consumption:.2f}")
col3.metric("Peak Usage Day", str(peak_day['date'].date()))

# -----------------------------
# LINE CHART (SAFE)
# -----------------------------
st.subheader("📈 Daily Energy Consumption Trend")

if not daily_usage.empty:
    fig = px.line(daily_usage, x='date', y='usage_kwh', markers=True)
    st.plotly_chart(fig, use_container_width=True)
else:
    st.info("Not enough data for chart")

# -----------------------------
# APPLIANCE CHART
# -----------------------------
st.subheader("🔌 Appliance Usage")

appliance_usage = filtered_df.groupby('appliance')['usage_kwh'].sum().reset_index()

if not appliance_usage.empty:
    fig2 = px.pie(appliance_usage, names='appliance', values='usage_kwh')
    st.plotly_chart(fig2, use_container_width=True)

# -----------------------------
# SIMPLE PREDICTION
# -----------------------------
st.subheader("🔮 Prediction")

if len(daily_usage) > 1:
    daily_usage['day'] = np.arange(len(daily_usage))

    X = daily_usage[['day']]
    y = daily_usage['usage_kwh']

    model = LinearRegression()
    model.fit(X, y)

    next_day = np.array([[len(daily_usage)]])
    prediction = model.predict(next_day)[0]

    st.success(f"Predicted next usage: {prediction:.2f} kWh")
else:
    st.info("Not enough data for prediction")

# -----------------------------
# FOOTER
# -----------------------------
st.caption("Built with Streamlit | AI + Sustainability + Data Analytics")