quick_test.py

"""
Simple test script for Singtel Bill Scanner models
This script demonstrates basic usage without heavy model downloads
"""

from transformers import pipeline
from PIL import Image
import requests
import io

def test_trocr_simple():
    """
    Test TrOCR model with a simple example
    """
    print("Testing TrOCR model...")
    
    try:
        # Create a simple test image with text
        from PIL import Image, ImageDraw, ImageFont
        
        # Create a simple image with text
        img = Image.new('RGB', (300, 100), color='white')
        draw = ImageDraw.Draw(img)
        
        # Draw some text
        try:
            # Try to use a default font
            font = ImageFont.load_default()
        except:
            font = None
            
        draw.text((10, 30), "Total: $123.45", fill='black', font=font)
        draw.text((10, 50), "Due: 2025-07-31", fill='black', font=font)
        
        # Save test image
        img.save("test_bill.png")
        print("Created test bill image: test_bill.png")
        
        # Initialize TrOCR pipeline
        print("Loading TrOCR pipeline...")
        pipe = pipeline("image-to-text", model="microsoft/trocr-base-handwritten")
        
        # Process the test image
        print("Processing test image...")
        result = pipe(img)
        
        print(f"Extracted text: {result[0]['generated_text']}")
        
        return True
        
    except Exception as e:
        print(f"Error in TrOCR test: {e}")
        return False

def quick_usage_demo():
    """
    Show quick usage examples without running heavy models
    """
    print("\n=== Quick Usage Demo ===")
    
    print("\n1. Basic TrOCR Usage:")
    print("""
from transformers import pipeline
from PIL import Image

# Initialize pipeline
pipe = pipeline("image-to-text", model="microsoft/trocr-base-handwritten")

# Process image
image = Image.open("your_bill.jpg")
result = pipe(image)
text = result[0]['generated_text']
    """)
    
    print("\n2. Bill Processing Function:")
    print("""
def process_bill(image_path):
    pipe = pipeline("image-to-text", model="microsoft/trocr-base-handwritten")
    image = Image.open(image_path)
    result = pipe(image)
    
    # Extract bill information
    text = result[0]['generated_text']
    
    # Parse specific information
    total_amount = extract_amount(text)
    due_date = extract_date(text)
    account_number = extract_account(text)
    
    return {
        'total': total_amount,
        'due_date': due_date,
        'account': account_number,
        'raw_text': text
    }
    """)

if __name__ == "__main__":
    print("=== Singtel Bill Scanner - Quick Test ===")
    
    # Show usage demo first
    quick_usage_demo()
    
    # Ask user if they want to run the actual test
    print("\n" + "="*50)
    print("Would you like to run the actual TrOCR test?")
    print("Note: This will download the model (~1.3GB) on first run")
    print("="*50)
    
    user_input = input("Run test? (y/n): ").lower().strip()
    
    if user_input == 'y' or user_input == 'yes':
        success = test_trocr_simple()
        if success:
            print("\n✅ Test completed successfully!")
            print("You can now use the models with your bill images.")
        else:
            print("\n❌ Test failed. Check the error messages above.")
    else:
        print("\nTest skipped. Refer to the usage examples above.")
        print("Run this script again with 'y' when you're ready to test.")
    
    print("\n📚 For detailed documentation, see README.md")
    print("📁 Your models will be cached in ~/.cache/huggingface/ for future use")