import numpy as np
import re
import cv2
from PIL import Image
import easyocr
import os

# Initialize OCR Reader
reader = easyocr.Reader(['en'], gpu=False)

def preprocess_image(image):
    """Preprocess the image to improve OCR accuracy"""
    # Convert to grayscale
    gray = cv2.cvtColor(image, cv2.COLOR_RGB2GRAY)

    # Apply threshold to isolate digits
    _, thresh = cv2.threshold(gray, 150, 255, cv2.THRESH_BINARY_INV + cv2.THRESH_OTSU)

    return thresh

def extract_weight_from_image(pil_image):
    try:
        # Convert PIL to OpenCV format
        image = np.array(pil_image.convert("RGB"))

        # Print image shape for debugging
        print("Image shape:", image.shape)

        # Preprocess for better OCR accuracy
        processed = preprocess_image(image)

        # Save debug image
        debug_img = Image.fromarray(processed)
        debug_path = "debug_processed_image.png"
        debug_img.save(debug_path)
        print(f"✅ Processed image saved to: {debug_path}")

        # Run OCR on processed image
        result = reader.readtext(processed)

        print("✅ OCR Results:")
        for r in result:
            print(f"Text: '{r[1]}' | Confidence: {r[2] * 100:.2f}%")

        # Try to find numeric weight
        weight = None
        confidence = 0.0

        for detection in result:
            text = detection[1]
            conf = detection[2]

            # Match numbers like 53.25 or 45
            match = re.search(r"\b\d+(\.\d+)?\b", text)
            if match:
                weight = match.group()
                confidence = conf
                break

        if weight:
            return weight, round(confidence * 100, 2)
        else:
            return "No weight detected", 0.0

    except Exception as e:
        print("❌ Exception during OCR:", str(e))
        return f"Error: {str(e)}", 0.0