import numpy as np
import tensorflow as tf
from tensorflow.keras.models import load_model
from tensorflow.keras.preprocessing.image import load_img, img_to_array
import gradio as gr

# Define class names for the model
class_names = ["Girl", "Boy"]  # Adjusted to represent binary classification

# Load the pre-trained model
model_path = 'best_model.h5'  # Update this path
model = load_model(model_path)

# Function to preprocess the image and make predictions
def classify_image(image):
    # Load and preprocess the image
    img = load_img(image, target_size=(150, 150))  # Ensure this matches the input size of your model
    img_array = img_to_array(img) / 255.0  # Normalize pixel values
    img_array = img_array.reshape((1, 150, 150, 3))  # Reshape for the model

    # Log the shape of the image array for debugging
    print(f"Image shape for prediction: {img_array.shape}")

    # Get prediction
    prediction = model.predict(img_array)
    predicted_class = class_names[1] if prediction[0][0] > 0.5 else class_names[0]  # Interpret output
    return predicted_class

# Create a Gradio interface
iface = gr.Interface(fn=classify_image,
                     inputs=gr.Image(type="filepath"),
                     outputs="text",
                     title="Boy or Girl Classifier",
                     description="Upload an image to classify as either Boy or Girl.")

# Run the Gradio app
iface.launch(share=True)  # Set share=True if you want to share the link