parse_toshiba.py

import re
from typing import Dict

def parse_toshiba(file_text: str) -> Dict:
    parsed_data = {
        "Purchase Order No": "",
        "Date": "",
        "Supplier": "Toshiba Transmission and Distribution Systems (India) Pvt Ltd",
        "Invoice Address": "",
        "Delivery Address": "",
        "Currency": "INR",
        "Payment Terms": "",
        "Items": []
    }

    try:
        parsed_data["Purchase Order No"] = re.search(r"Purchase Order : (\w+)", file_text).group(1)
        parsed_data["Date"] = re.search(r"Order Date : (\d{4}-\d{2}-\d{2})", file_text).group(1)

        item_pattern = re.compile(r"Pos\.\s+(\d+)\s+([\w\s]+)\s+(\d+)\s+([\d.]+)\s+INR\s+([\d.]+)")
        for match in item_pattern.finditer(file_text):
            parsed_data["Items"].append({
                "Position": match.group(1),
                "Description": match.group(2).strip(),
                "Quantity": int(match.group(3)),
                "Unit Price": float(match.group(4)),
                "Total Price": float(match.group(5)),
            })
    except Exception as e:
        print(f"Error parsing Toshiba PO: {e}")
    
    return parsed_data