app.py

import pdfplumber
import pandas as pd
import gradio as gr
import re

# Define function to extract data
def extract_data(pdf_file):
    data = []
    columns = ["Purchase Order No", "Date", "SI No", "Material Description", "Unit", "Quantity", "Dely Qty", "Dely Date", "Unit Rate", "Value"]

    purchase_order_no = None
    purchase_order_date = None

    with pdfplumber.open(pdf_file) as pdf:
        for page in pdf.pages:
            text = page.extract_text().splitlines()
            
            # Extract Purchase Order No and Date dynamically from the first page
            if not purchase_order_no or not purchase_order_date:
                for line in text:
                    # Search for Purchase Order No
                    po_match = re.search(r'Purchase Order No[:\s]+(\d+)', line)
                    if po_match:
                        purchase_order_no = po_match.group(1)
                    
                    # Search for Date
                    date_match = re.search(r'Date[:\s]+(\d{2}\.\d{2}\.\d{4})', line)
                    if date_match:
                        purchase_order_date = date_match.group(1)
                    
                    # Stop searching if both fields are found
                    if purchase_order_no and purchase_order_date:
                        break

            # Process each line to extract relevant data rows
            for line in text:
                # Using regex or keywords to identify each row
                try:
                    # Example row pattern match for SI No (Assuming starts with numbers in multiples of 10)
                    if re.match(r'^\d+\s', line):  
                        parts = line.split()
                        si_no = parts[0]  # Extract SI No
                        material_desc = "BPS 017507\nMaterial Number: {}\nHSN Code: 8310\nIGST: 18%".format(parts[2])  # Example Material Description
                        unit = "NO"
                        quantity = parts[3]
                        dely_qty = parts[4]
                        dely_date = parts[5]
                        unit_rate = parts[6]
                        value = parts[7]

                        # Append data as a row in the correct order
                        data.append([
                            purchase_order_no,
                            purchase_order_date,
                            si_no,
                            material_desc,
                            unit,
                            quantity,
                            dely_qty,
                            dely_date,
                            unit_rate,
                            value
                        ])
                except (ValueError, IndexError):
                    continue  # Skip lines that do not match the expected pattern

    # Convert data to a DataFrame and save it as Excel
    df = pd.DataFrame(data, columns=columns)
    excel_path = "/tmp/Extracted_Purchase_Order_Data.xlsx"
    df.to_excel(excel_path, index=False)
    return excel_path

# Set up Gradio interface
iface = gr.Interface(
    fn=extract_data,
    inputs=gr.File(label="Upload PDF"),
    outputs=gr.File(label="Download Excel"),
    title="PDF Data Extractor"
)

# Launch the app
iface.launch()