app.py

import gradio as gr
import os
from PO_parse import bhel  # Assuming bhel.py is in a folder named PO_parse

def process_po(customer_name, file):
    if customer_name == "BHEL":
        # Use the normalized extraction function from bhel.py
        extracted_data = bhel.extract_normalized_po_data(file.name)
        
        # Ensure output directory exists and save CSV file
        output_dir = "./output"
        os.makedirs(output_dir, exist_ok=True)  # Create the directory if it doesn't exist
        csv_file_path = os.path.join(output_dir, "extracted_po_data.csv")
        
        # Save DataFrame to CSV
        extracted_data.to_csv(csv_file_path, index=False)
        
        # Display the extracted data and provide the CSV download link
        return extracted_data, "PO data saved to CSV.", csv_file_path
    else:
        return None, "Currently unsupported customer format.", None

# Gradio Interface setup
customer_dropdown = gr.Dropdown(choices=["BHEL", "Others"], label="Customer Name")
file_input = gr.File(label="Upload PO File (PDF)")

# Define Gradio interface with function, inputs, and outputs
interface = gr.Interface(
    fn=process_po,
    inputs=[customer_dropdown, file_input],
    outputs=[
        gr.Dataframe(label="Extracted PO Details"),
        gr.Textbox(label="Summary"),
        gr.File(label="Download CSV")
    ],
    title="Purchase Order Parsing Tool",
    description="Select a customer and upload a PO file to extract details."
)

if __name__ == "__main__":
    interface.launch()