Nehal61/Procurement_automation

1 2import gradio as gr3import pandas as pd4from statsmodels.tsa.arima.model import ARIMA5import warnings6 7# Suppress specific warnings related to ARIMA8warnings.filterwarnings("ignore", category=UserWarning)9warnings.filterwarnings("ignore", category=FutureWarning)10warnings.filterwarnings("ignore", category=RuntimeWarning)11 12# Function to process the CSV and perform the forecasting13def demand_forecasting(csv_file, type_value):14    # Load the CSV file15    df = pd.read_csv(csv_file.name)  # .name is used to get the file path16 17    # Preprocess the data18    df.columns = df.columns.str.strip()19    for col in df.select_dtypes(include='object').columns:20        df[col] = df[col].str.strip()21    df['Date of Sale'] = pd.to_datetime(df['Date of Sale'], format='%d-%b-%y')22 23    # Filter the data based on 'Type'24    df_filtered = df[df['Type'] == type_value]25    df_filtered.set_index('Date of Sale', inplace=True)26 27    # Group and resample data by product name and month-end frequency28    monthly_sales_by_product = (29        df_filtered.groupby('Product Name')30                   .resample('M')['Quantity']31                   .sum()32                   .reset_index()33    )34 35    # Calculate total sales by product36    Total_sales_product = df_filtered.groupby('Product Name')['Quantity'].sum().reset_index()37 38    # Create a list to store the forecasted sums for each product39    forecast_sums = []40 41    # Get a list of unique product names42    product_names = monthly_sales_by_product['Product Name'].unique()43 44    # Loop through each product and forecast sales45    for product_name in product_names:46        # Filter data for the current product47        product_data = monthly_sales_by_product[monthly_sales_by_product['Product Name'] == product_name].copy()48 49        # Ensure 'Date of Sale' is in datetime format using .loc[]50        product_data.loc[:, 'Date of Sale'] = pd.to_datetime(product_data['Date of Sale'])51 52        # Set the 'Date of Sale' as the index and ensure the index has frequency information53        product_data.set_index('Date of Sale', inplace=True)54 55        # Check if we have enough data points and set a frequency56        if len(product_data) >= 1:  # Adjust based on ARIMA requirements57            product_data = product_data.asfreq('M')  # Set frequency to monthly58 59            try:60                # Fit the ARIMA model61                model = ARIMA(product_data['Quantity'], order=(1, 1, 1))62                model_fit = model.fit()63 64                # Forecast for the next 4 months65                forecast_steps = 466                forecast = model_fit.forecast(steps=forecast_steps)67 68 69                # Sum the forecasted values for this product70                forecast_sum = forecast.sum()71                forecast_sum = round(forecast_sum)72                forecast_sums.append({'Product Name': product_name, 'Forecasted Quantity Sum': forecast_sum})73 74            except Exception as e:75                print(f"Could not fit ARIMA model for {product_name}: {e}")76        else:77            print(f"Not enough data points for {product_name} to fit ARIMA model.")78 79    # Create a DataFrame from the forecast sums80    forecast_summary_df = pd.DataFrame(forecast_sums)81 82    # Perform a left merge to ensure all product names from Total_sales_product are included83    combined_df = pd.merge(Total_sales_product, forecast_summary_df, on='Product Name', how='left')84 85    # Rename columns for clarity86    combined_df.columns = ['Product Name', 'Total Quantity Sold', 'Forecasted Quantity Sum']87 88    # Save the combined DataFrame to a CSV file89    output_file = 'combined_df.csv'90    combined_df.to_csv(output_file, index=False)91 92    return output_file93 94# Gradio Interface95interface = gr.Interface(96    fn=demand_forecasting,97    inputs=[98        gr.File(label="Upload CSV File"),99        gr.Textbox(label="Enter 'Type' (e.g., 'EW')")100    ],101    outputs=gr.File(label="Download Combined Data CSV")102)103 104# Launch the interface105interface.launch(share=True)106