import streamlit as st
import numpy as np
from PIL import Image
from tensorflow.keras.models import load_model
import os
import time
import requests
from io import BytesIO

# Function to load the model (cached for efficiency)
@st.cache_resource
def load_vgg_model():
    return load_model("brain_tumor_vgg16_model.keras")

# Function to run the prediction and show progress
def run_prediction(image_path, model, img_size):
    # Create a progress bar and status text
    progress_bar = st.progress(0)
    status_text = st.empty()

    for i in range(100):
        progress_bar.progress(i + 1)
        status_text.text(f"Processing... {i+1}%")
        time.sleep(0.01)

    # Process and predict
    try:
        # Check if the input is a file-like object or a string path
        if isinstance(image_path, str):
            image = Image.open(image_path).convert("RGB")
        else:
            image = Image.open(image_path).convert("RGB")

        img_array = np.array(image.resize((img_size, img_size)))
        img_array = np.expand_dims(img_array, axis=0)
        img_array = img_array / 255.0

        prediction = model.predict(img_array)
        class_predicted = (prediction > 0.5).astype("int32")[0][0]

        # Display the result
        if class_predicted == 1:
            st.error(f"Prediction: Tumor Detected (Probability: {100*prediction[0][0]:.2f}%)")
        else:
            st.success(f"Prediction: No Tumor Detected (Probability: {100*(1 - prediction[0][0]):.2f}%)")
    except Exception as e:
        st.error(f"An error occurred during classification: {e}")

    # Clear the progress bar
    progress_bar.empty()
    status_text.empty()

# Streamlit UI
st.title("Brain Tumor MRI Classification App")
st.write("Upload your own image to get a prediction.")

# Load the model
model = load_vgg_model()

# Define image size (must match the model's input size)
img_size = 150

# --- UI for user image upload ---
st.subheader("Upload Your Own Image")
uploaded_file = st.file_uploader("Choose an image...", type=["jpg", "jpeg", "png"])

if uploaded_file is not None:
    st.image(uploaded_file, caption='Uploaded MRI Scan', use_container_width=True)

    # Use a button to trigger the classification explicitly
    if st.button("Check for Brain Tumor"):
        run_prediction(uploaded_file, model, img_size)