import gradio as gr
import numpy as np
import joblib
from tensorflow.keras.utils import load_img, img_to_array

# Load the pre-trained model
model = joblib.load("flower1.pkl")

# Define the class names (must match the training dataset order)
class_names = ["Rose", "Tulip", "Daisy", "Sunflower", "Daffodil","Lily","Orchid","Lotus","Jasmine","Marigold","Hibiscus","Tube rose","Gladiolus","Frangipani","Iris","Dahlia"]

# Function to preprocess the image and make predictions
def predict_flower(image):
    # Resize and preprocess the image
    img = image.resize((150, 150))  # Resize image to match the model input
    img_array = np.array(img) / 255.0  # Normalize pixel values to [0, 1]
    img_array = img_array.reshape((1, 150, 150, 3))  # Add batch dimension
    
    # Make prediction
    predictions = model.predict(img_array)
    predicted_class = class_names[np.argmax(predictions)]
    return f"The predicted flower is: {predicted_class}"

# Create the Gradio interface
title = "Flower Classification"
description = "Upload an image of a flower, and the model will predict the type of flower (Daisy, Rose, Sunflower, Tulip etc.)."

gr_interface = gr.Interface(
    fn=predict_flower,  # Function to process predictions
    inputs=gr.Image(type="pil"),  # Input: Image (PIL format)
    outputs=gr.Textbox(),  # Output: Textbox for predicted class
    title=title,
    description=description
)

# Launch the Gradio app
if __name__ == "__main__":
    gr_interface.launch()