app.py

import gradio as gr
import numpy as np
from PIL import Image
import pickle
import os

# Load the model
model_path = "numbers.pkl"
classifier = None

def load_model():
    """Load the trained model"""
    global classifier
    try:
        if os.path.exists(model_path):
            with open(model_path, 'rb') as file:
                classifier = pickle.load(file)
            print("✅ Model loaded successfully!")
        else:
            print(f"❌ Model file '{model_path}' not found.")
    except Exception as e:
        print(f"❌ Error loading model: {str(e)}")

# Load model on startup
load_model()

# Result mapping
ResultMap = {0: '0', 1: '1', 2: '2', 3: '3', 4: '4', 
             5: '5', 6: '6', 7: '7', 8: '8', 9: '9'}

def predict_digit(image):
    """
    Predict digit from uploaded image
    """
    if classifier is None:
        return "❌ Model not loaded", 0.0
    
    try:
        if image is None:
            return "❌ No image provided", 0.0
        
        # Preprocess image
        # Convert to PIL Image if it's not already
        if not isinstance(image, Image.Image):
            image = Image.fromarray(image)
        
        # Resize and convert
        image = image.convert('RGB').resize((64, 64), Image.Resampling.LANCZOS)
        
        # Convert to array and normalize
        image_array = np.asarray(image, dtype=np.float32) / 255.0
        
        # Add batch dimension
        image_array = np.expand_dims(image_array, axis=0)
        
        print(f"📸 Image processed: {image_array.shape}")
        
        # Make prediction
        result = classifier.predict(image_array)
        
        # Handle different model output formats
        if hasattr(result, 'shape') and len(result.shape) > 1:
            predicted_class = np.argmax(result, axis=1)[0] if result.shape[0] == 1 else np.argmax(result)
            confidence = float(np.max(result))
        else:
            predicted_class = int(result[0]) if hasattr(result, '__len__') else int(result)
            confidence = 1.0
        
        predicted_digit = ResultMap.get(predicted_class, "Unknown")
        
        print(f'🎯 Prediction: {predicted_digit} (Confidence: {confidence:.2%})')
        
        return f"Predicted Digit: {predicted_digit}", confidence
    
    except Exception as e:
        print(f"❌ Error in prediction: {str(e)}")
        return f"❌ Error: {str(e)}", 0.0

# Create Gradio interface
def create_gradio_interface():
    """Create the Gradio interface"""
    
    with gr.Blocks(
        theme=gr.themes.Soft(),
        title="🔢 Digit Recognition API",
        css="""
        .gradio-container {
            max-width: 700px !important;
            margin: auto !important;
        }
        .header {
            text-align: center;
            margin-bottom: 2rem;
        }
        .result-box {
            font-size: 1.5rem;
            font-weight: bold;
            text-align: center;
            padding: 1rem;
            border-radius: 10px;
        }
        """
    ) as interface:
        
        gr.HTML("""
        <div class="header">
            <h1>🔢 Digit Recognition API</h1>
            <p>Upload an image of a handwritten digit (0-9) to get AI-powered recognition</p>
        </div>
        """)
        
        with gr.Row():
            with gr.Column():
                # Input
                image_input = gr.Image(
                    label="📸 Upload Digit Image",
                    type="pil",
                    height=300
                )
                
                predict_btn = gr.Button(
                    "🚀 Predict Digit",
                    variant="primary",
                    size="lg"
                )
            
            with gr.Column():
                # Outputs
                prediction_output = gr.Textbox(
                    label="🎯 Prediction Result",
                    interactive=False,
                    elem_classes=["result-box"]
                )
                
                confidence_output = gr.Number(
                    label="📊 Confidence Score",
                    interactive=False
                )
        
        # Event handlers
        predict_btn.click(
            fn=predict_digit,
            inputs=[image_input],
            outputs=[prediction_output, confidence_output]
        )
        
        # Also predict on image upload
        image_input.change(
            fn=predict_digit,
            inputs=[image_input],
            outputs=[prediction_output, confidence_output]
        )
        
        gr.HTML("""
        <div style="text-align: center; margin-top: 2rem; color: #666;">
            <p>📋 <strong>Tips for best results:</strong></p>
            <p>• Use clear images with dark digits on light backgrounds</p>
            <p>• Ensure the digit fills most of the image</p>
            <p>• Supported formats: PNG, JPG, GIF</p>
        </div>
        """)
    
    return interface

# API endpoint for external calls
def predict_api(image):
    """API endpoint that returns JSON response"""
    prediction, confidence = predict_digit(image)
    
    # Extract just the digit from the prediction string
    digit = prediction.split(": ")[-1] if ": " in prediction else prediction
    
    return {
        "digit": digit,
        "confidence": float(confidence),
        "status": "success" if "❌" not in prediction else "error"
    }

if __name__ == "__main__":
    # Create and launch the interface
    interface = create_gradio_interface()
    
    # Launch with API enabled
    interface.launch(
        server_name="0.0.0.0",
        server_port=7860,
        share=False,
        show_api=True
    )