app.py

import os
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import gradio as gr
from groq import Groq

# Set up the API key (replace with your actual Groq API key)
os.environ["GROQ_API_KEY"] = "My_API_key"

# Groq client setup
client = Groq(api_key=os.environ.get("My_API_key"))

# Function to generate traffic optimization strategies from Groq
def generate_traffic_optimization(data: str):
    response = client.chat.completions.create(
        messages=[
            {
                "role": "user",
                "content": f"Generate a detailed traffic flow optimization strategy for the following data. Include peak hours, vehicle type distributions, and actionable suggestions to improve flow: {data}",
            }
        ],
        model="llama3-8b-8192",  # You can choose a different model if necessary
    )
    return response.choices[0].message.content

# Visualization function to generate the traffic flow chart
def generate_traffic_chart(df_filtered):
    # Summarize traffic data by aggregating vehicle counts for each record
    df_filtered['Total'] = df_filtered['CarCount'] + df_filtered['BikeCount'] + df_filtered['BusCount'] + df_filtered['TruckCount']

    # Select relevant columns for the optimization request (you can adjust this as needed)
    traffic_data_summary = df_filtered[['Time', 'Total', 'Traffic Situation', 'CarCount', 'BikeCount', 'BusCount', 'TruckCount']]

    # Calculate some basic statistics
    avg_car_count = df_filtered['CarCount'].mean()
    avg_bike_count = df_filtered['BikeCount'].mean()
    avg_bus_count = df_filtered['BusCount'].mean()
    avg_truck_count = df_filtered['TruckCount'].mean()
    peak_traffic_time = df_filtered.loc[df_filtered['Total'].idxmax()]['Time']

    # Generate summary for Groq API input
    summary_str = traffic_data_summary.head(10).to_string(index=False)

    # Get the optimization strategy from Groq
    optimization_strategy = generate_traffic_optimization(summary_str)

    # Visualization of traffic flow data
    time_labels = df_filtered['Time'].head(10)
    car_counts = df_filtered['CarCount'].head(10)
    bike_counts = df_filtered['BikeCount'].head(10)
    bus_counts = df_filtered['BusCount'].head(10)
    truck_counts = df_filtered['TruckCount'].head(10)

    # Create the stacked bar chart for vehicle counts
    fig, ax = plt.subplots(figsize=(10, 6))
    ax.bar(time_labels, car_counts, label='Cars', color='blue')
    ax.bar(time_labels, bike_counts, bottom=car_counts, label='Bikes', color='green')
    ax.bar(time_labels, bus_counts, bottom=np.array(car_counts) + np.array(bike_counts), label='Buses', color='red')
    ax.bar(time_labels, truck_counts, bottom=np.array(car_counts) + np.array(bike_counts) + np.array(bus_counts), label='Trucks', color='yellow')

    # Customize chart
    ax.set_xlabel('Time')
    ax.set_ylabel('Vehicle Count')
    ax.set_title('Traffic Flow by Vehicle Type')
    ax.legend()

    # Save the plot as a file
    plt.xticks(rotation=45)
    plt.tight_layout()
    chart_path = "/tmp/traffic_chart.png"
    plt.savefig(chart_path)
    plt.close()

    # Return additional insights
    insights = f"""
    Average Car Count: {avg_car_count:.2f}
    Average Bike Count: {avg_bike_count:.2f}
    Average Bus Count: {avg_bus_count:.2f}
    Average Truck Count: {avg_truck_count:.2f}
    Peak Traffic Time: {peak_traffic_time}
    """
    
    return optimization_strategy + "\n\n" + insights, chart_path

# Function to process the uploaded file and run traffic optimization
def process_traffic_file(file):
    # Load the dataset
    df = pd.read_csv(file.name)
    
    # Optionally, you may filter data for specific days or time intervals
    # For example, let's filter the data for a specific day:
    df_filtered = df[df['Day of the week'] == 'Monday']

    # Generate traffic chart and optimization strategy
    optimization_strategy, chart_path = generate_traffic_chart(df_filtered)

    return optimization_strategy, chart_path

# Gradio interface
iface = gr.Interface(
    fn=process_traffic_file,
    inputs=gr.File(label="Upload CSV with Traffic Data"),
    outputs=[gr.Textbox(label="Optimization Strategy and Insights"), gr.Image(label="Traffic Flow Chart")],
    title="Traffic Flow Optimization",
    description="Upload a CSV file with traffic data, and the app will generate traffic optimization strategies, provide insights, and visualize the traffic flow."
)

# Launch the app
iface.launch()