ocr_engine.py

from PIL import Image
import pytesseract
import re

def extract_weight(img_path):
    img = Image.open(img_path).convert("L")  # Grayscale

    # OCR
    text = pytesseract.image_to_string(img, config='--psm 6')
    text = text.lower().replace('\n', ' ').strip()

    # Find weight + unit (e.g., 52.25 g, 75.8 kg)
    match = re.search(r'(\d+\.\d+|\d+)\s*(kg|g)', text)
    if match:
        number = match.group(1)
        unit = match.group(2)
        return f"{number} {unit}"
    else:
        return "Weight not detected"