import gradio as gr
import pandas as pd
import random
from transformers import AutoTokenizer, AutoModelForCausalLM
import plotly.graph_objects as go

# Attempt to import PyTorch
try:
    import torch
except ImportError:
    torch = None

# Load GPT-2 model if PyTorch is available
if torch:
    tokenizer = AutoTokenizer.from_pretrained("gpt2")
    model = AutoModelForCausalLM.from_pretrained("gpt2")
else:
    tokenizer, model = None, None
    print("PyTorch not available. GPT-2 model will not be loaded.")

# Analyze energy data and provide consumption details, recommendations, and weather tips
def analyze_energy_data(energy_data, location, language):
    appliances = {}
    total_kwh = 0
    peak_hours_rate = 20
    off_peak_hours_rate = 12
    benchmarks = {"AC": 400, "Refrigerator": 200, "Lighting": 150, "Fan": 100}
    alerts = []

    try:
        for line in energy_data.strip().split("\n"):
            appliance, kwh = line.split(":")
            kwh_value = float(kwh.strip().split(" ")[0])
            appliances[appliance.strip()] = kwh_value
            total_kwh += kwh_value

            if appliance.strip() in benchmarks and kwh_value > benchmarks[appliance.strip()]:
                alert_message = (
                    f"Your {appliance.strip()} usage exceeds the limit by "
                    f"{kwh_value - benchmarks[appliance.strip()]:.2f} kWh."
                )
                alerts.append(alert_message)

    except Exception:
        return (
            "Error: Enter data in the correct format (e.g., AC: 500 kWh).",
            "",
            "",
            "",
            "",
            "",
            0.0
        )

    total_bill = total_kwh * peak_hours_rate
    optimized_bill = sum(
        appliances[app] * (off_peak_hours_rate if app in ["AC", "Refrigerator"] else peak_hours_rate)
        for app in appliances
    )
    savings = total_bill - optimized_bill
    carbon_emissions = total_kwh * 0.707  # Approx kg of CO2 per kWh
    weather_tips = (
        f"Considering high temperatures in {location}, keep windows closed during peak heat hours to optimize cooling."
        if "Lahore" in location
        else "Check local weather to optimize energy usage."
    )

    return (
        f"Your current bill is PKR {total_bill:.2f}, potentially saving PKR {savings:.2f}.",
        "\n".join([f"{appliance}: {random.choice(['Use during off-peak hours.', 'Turn off when not in use.'])}" for appliance in appliances]),
        weather_tips,
        "\n".join(alerts),
        f"Your carbon footprint: {carbon_emissions:.2f} kg of CO2. Consider using renewable energy.",
        f"AI Recommendation: Optimize usage of AC and lighting based on peak hours to reduce costs and emissions.",
        savings
    )

# Build the Gradio UI
def build_ui():
    with gr.Blocks() as demo:
        gr.Markdown("## Energy Consumption Analyzer")
        
        with gr.Row():
            energy_data = gr.Textbox(
                label="Enter Energy Data",
                placeholder="e.g., AC: 500 kWh\nRefrigerator: 300 kWh",
                lines=5
            )
            location = gr.Textbox(label="Enter Location", placeholder="e.g., Lahore")
            language = gr.Dropdown(
                label="Select Language",
                choices=["English", "Urdu"],
                value="English"
            )

        with gr.Row():
            analyze_button = gr.Button("Analyze")

        with gr.Row():
            output_bill = gr.Text(label="Estimated Bill")
            output_recommendations = gr.Text(label="Recommendations")

        with gr.Row():
            output_weather_tips = gr.Text(label="Weather Tips")
            output_alerts = gr.Text(label="Alerts")
            output_carbon_footprint = gr.Text(label="Carbon Footprint")
            output_ai_recommendation = gr.Text(label="AI Recommendations")

        analyze_button.click(
            analyze_energy_data,
            inputs=[energy_data, location, language],
            outputs=[
                output_bill,
                output_recommendations,
                output_weather_tips,
                output_alerts,
                output_carbon_footprint,
                output_ai_recommendation
            ]
        )

    return demo

# Launch the Gradio app
demo = build_ui()
demo.launch()