# -*- coding: utf-8 -*-
"""Fuel_CompAPP.ipynb

Automatically generated by Colab.

Original file is located at
    https://colab.research.google.com/drive/1OtjGIkl4y8uY9b3pvalX4LzTIUDaGpYT
"""



import pandas as pd
import seaborn as sns
import matplotlib.pyplot as plt
import plotly.express as px
import gradio as gr



try:
    df = pd.read_csv("https://raw.githubusercontent.com/baren-kumar/File-/refs/heads/main/data-fuel-consumption.csv")
except FileNotFoundError:
    raise FileNotFoundError("https://raw.githubusercontent.com/baren-kumar/File-/refs/heads/main/data-fuel-consumption.csv' was not found.")

def create_dashboard_components(num_cylinders, vehicle_class):
    """
    Generates all dashboard components based on the selected number of cylinders and vehicle class.
    """
    # Filter data for both selected criteria
    filtered_df = df[(df['CYLINDERS'] == num_cylinders) & (df['VEHICLECLASS'] == vehicle_class)]

    if filtered_df.empty:
        # Return empty components and a message if no data is found
        return None, None, None, f"No data available for {num_cylinders} cylinders and {vehicle_class}.", None

    # Group by MODEL and calculate total CO2 emissions
    emissions_by_model = filtered_df.groupby('MODEL')['CO2EMISSIONS'].sum().reset_index()

    # Create the bar chart using Plotly
    bar_chart = px.bar(
        emissions_by_model,
        x='MODEL',
        y='CO2EMISSIONS',
        title=f'Total CO2 Emissions by Model for {num_cylinders} Cylinders and {vehicle_class}'
    )
    bar_chart.update_layout(
        xaxis_title="Vehicle Model",
        yaxis_title="Total CO2 Emissions",
        xaxis={'categoryorder': 'total descending'}
    )

    # Create the pie chart using Plotly
    pie_chart = px.pie(
        emissions_by_model,
        values='CO2EMISSIONS',
        names='MODEL',
        title=f'Distribution of CO2 Emissions for {num_cylinders} Cylinders and {vehicle_class}'
    )

    # Create a summary string
    total_emissions = filtered_df['CO2EMISSIONS'].sum()
    average_emissions = filtered_df['CO2EMISSIONS'].mean()
    summary = (
        f"Summary for vehicles with {num_cylinders} cylinders and vehicle class '{vehicle_class}':\n"
        f"Total CO2 Emissions: {total_emissions:,.2f} g/km\n"
        f"Average CO2 Emissions: {average_emissions:,.2f} g/km\n"
        f"Number of Records: {len(filtered_df)}"
    )

    return bar_chart, pie_chart, filtered_df, summary, None

# Get a unique sorted list of cylinder counts for the dropdown
unique_cylinders = sorted(df['CYLINDERS'].unique().tolist())
unique_vehicle_classes = sorted(df['VEHICLECLASS'].unique().tolist())

# Create the Gradio interface
with gr.Blocks(title="CO2 Emissions Dashboard") as dashboard:
    gr.Markdown("# 📊 CO2 Emissions Dashboard")
    gr.Markdown("Use the filters below to explore CO2 emissions data.")

    with gr.Row():
        cylinder_selector = gr.Dropdown(
            choices=unique_cylinders,
            label="Select Number of Cylinders",
            value=unique_cylinders[0]
        )
        vehicle_class_selector = gr.Dropdown(
            choices=unique_vehicle_classes,
            label="Select Vehicle Class",
            value=unique_vehicle_classes[0]
        )

    with gr.Row():
        with gr.Column(scale=2):
            chart_bar = gr.Plot(label="Bar Chart")
        with gr.Column(scale=1):
            chart_pie = gr.Plot(label="Pie Chart")

    with gr.Row():
        summary_output = gr.Textbox(label="Summary", interactive=False)

    with gr.Row():
        table_output = gr.Dataframe(label="Data Table")

    # Define event to update all components when either dropdown value changes
    # A cleaner approach is to use a single function that takes both inputs
    def update_all(cylinders, vehicle_class):
        return create_dashboard_components(cylinders, vehicle_class)

    cylinder_selector.change(
        fn=update_all,
        inputs=[cylinder_selector, vehicle_class_selector],
        outputs=[chart_bar, chart_pie, table_output, summary_output]
    )

    vehicle_class_selector.change(
        fn=update_all,
        inputs=[cylinder_selector, vehicle_class_selector],
        outputs=[chart_bar, chart_pie, table_output, summary_output]
    )

    # Load the initial components when the app starts
    dashboard.load(
        fn=lambda: create_dashboard_components(unique_cylinders[0], unique_vehicle_classes[0]),
        outputs=[chart_bar, chart_pie, table_output, summary_output]
    )
    gr.Markdown("---")
    gr.Markdown("👨‍💻 Developed by **Baren Kumar Baidya **  \n📧 [eng.barenbscme@gmail.com](mailto:eng.barenbscme@gmail.com)")

# Launch the Gradio app
dashboard.launch()