ocr_engine.py

import cv2
import pytesseract
import numpy as np
from PIL import Image

def extract_weight_from_image(pil_img):
    try:
        # Convert PIL image to OpenCV
        img = pil_img.convert("RGB")
        img = np.array(img)
        img = cv2.cvtColor(img, cv2.COLOR_RGB2BGR)

        # Preprocess
        gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
        blur = cv2.GaussianBlur(gray, (3, 3), 0)

        # OCR
        text = pytesseract.image_to_string(blur, config='--psm 7 digits')
        weight = ''.join(filter(lambda c: c in '0123456789.', text))

        confidence = 95  # Replace with real confidence logic if needed
        return weight.strip(), confidence
    except Exception as e:
        print(f"OCR error: {e}")
        return "", 0