import gradio as gr
import tensorflow as tf
import numpy as np
from PIL import Image
from huggingface_hub import hf_hub_download
import os
import pandas as pd
import logging

# Setup logging
logging.basicConfig(level=logging.INFO)
logger = logging.getLogger(__name__)

# Configuration
MODEL_REPO = "Ahmedhassan54/Image-Classification"
MODEL_FILE = "best_model.h5"

# Initialize model to None
model = None

def load_model():
    global model
    try:
        logger.info("⏳ Downloading model...")
        model_path = hf_hub_download(
            repo_id=MODEL_REPO,
            filename=MODEL_FILE,
            cache_dir=".",
            force_download=True
        )
        logger.info(f"📁 Model path: {model_path}")
        
        # Verify file exists
        if not os.path.exists(model_path):
            raise FileNotFoundError(f"Model file not found at {model_path}")
            
        logger.info("🔄 Loading TensorFlow model...")
        model = tf.keras.models.load_model(model_path)
        logger.info("✅ Model loaded successfully!")
        
    except Exception as e:
        logger.error(f"❌ Model loading failed: {str(e)}")
        model = None
        raise gr.Error(f"Model loading failed. Check logs for details.")

# Load model when app starts
load_model()

def classify_image(image):
    try:
        if image is None:
            raise gr.Error("Please upload an image first")
            
        logger.info("🖼️ Processing image...")
        
        # Convert to PIL Image if numpy array
        if isinstance(image, np.ndarray):
            image = Image.fromarray(image)
            
        # Resize and normalize
        image = image.resize((150, 150))
        img_array = np.array(image) / 255.0
        if len(img_array.shape) == 3:
            img_array = np.expand_dims(img_array, axis=0)
            
        logger.info(f"📊 Input shape: {img_array.shape}")
        
        if model is None:
            raise gr.Error("Model not loaded - using demo mode")
            return {"Cat": 0.5, "Dog": 0.5}, pd.DataFrame({'Class': ['Cat', 'Dog'], 'Confidence': [0.5, 0.5]})
            
        pred = model.predict(img_array, verbose=0)
        confidence = float(pred[0][0])
        logger.info(f"🔮 Prediction confidence: {confidence}")
        
        results = {
            "Cat": round(1 - confidence, 4),
            "Dog": round(confidence, 4)
        }
        
        plot_data = pd.DataFrame({
            'Class': ['Cat', 'Dog'],
            'Confidence': [1 - confidence, confidence]
        })
        
        return results, plot_data
        
    except Exception as e:
        logger.error(f"💥 Classification error: {str(e)}")
        raise gr.Error(f"Error processing image: {str(e)}")

css = """
.gradio-container { max-width: 900px; margin: auto; }
footer { visibility: hidden; }
.progress-bar { color: #ff4d4d !important; }
"""

with gr.Blocks(css=css, theme=gr.themes.Soft()) as demo:
    gr.Markdown("""
    # 🐾 Cat vs Dog Classifier 🦮
    Upload an image to classify whether it's a cat or dog
    """)
    
    with gr.Row():
        with gr.Column():
            image_input = gr.Image(label="Upload Image", type="pil")
            with gr.Row():
                submit_btn = gr.Button("Classify 🚀", variant="primary")
                clear_btn = gr.Button("Clear 🗑️")
        
        with gr.Column():
            label_output = gr.Label(label="Predictions")
            confidence_bar = gr.BarPlot(
                pd.DataFrame({'Class': ['Cat', 'Dog'], 'Confidence': [0.5, 0.5]}),
                x="Class", y="Confidence", y_lim=[0,1],
                title="Confidence Scores", width=400, height=300
            )
    
    # Examples
    gr.Examples(
        examples=[
            ["https://upload.wikimedia.org/wikipedia/commons/1/15/Cat_August_2010-4.jpg"],
            ["https://upload.wikimedia.org/wikipedia/commons/d/d9/Collage_of_Nine_Dogs.jpg"]
        ],
        inputs=image_input,
        outputs=[label_output, confidence_bar],
        fn=classify_image,
        cache_examples=True
    )
    
    # Button actions
    submit_btn.click(
        fn=classify_image,
        inputs=image_input,
        outputs=[label_output, confidence_bar],
        api_name="predict"
    )
    
    clear_btn.click(
        fn=lambda: [None, pd.DataFrame({'Class': ['Cat', 'Dog'], 'Confidence': [0.5, 0.5]})],
        inputs=None,
        outputs=[image_input, confidence_bar]
    )

if __name__ == "__main__":
    demo.launch(debug=True)