import streamlit as st
import cv2
import numpy as np
from PIL import Image

st.title("Image Processing\n(Comparison View)")

# Upload image in the sidebar
uploaded_file = st.sidebar.file_uploader("Upload an Image", type=["png", "jpg", "jpeg"])

if uploaded_file is not None:
    image = Image.open(uploaded_file)
    img_array = np.array(image)

    # Select an operation using a dropdown
    option = st.selectbox("Choose an comparison:", [
        "None", "Convert to Grayscale", "Rotate Image", "Blur Image", "Convert to Color Space", "Edge Detection"
    ])

    # Convert to Grayscale
    if option == "Convert to Grayscale":
        gray_image = cv2.cvtColor(img_array, cv2.COLOR_RGB2GRAY)

        col1, col2 = st.columns(2)
        with col1:
            st.image(image, caption="Original Image", use_container_width=True)
        with col2:
            st.image(gray_image, caption="Grayscale Image", use_container_width=True)

    # Rotate Image
    elif option == "Rotate Image":
        angle = st.slider("Select Rotation Angle", -180, 180, 0)
        (h, w) = img_array.shape[:2]
        center = (w // 2, h // 2)
        matrix = cv2.getRotationMatrix2D(center, angle, 1.0)
        rotated_image = cv2.warpAffine(img_array, matrix, (w, h))

        col1, col2 = st.columns(2)
        with col1:
            st.image(image, caption="Original Image", use_container_width=True)
        with col2:
            st.image(rotated_image, caption=f"Rotated by {angle}°", use_container_width=True)

    # Blur Image
    elif option == "Blur Image":
        blur_level = st.slider("Select Blur Level", 1, 20, 5)
        kernel_size = (blur_level * 2 + 1, blur_level * 2 + 1)  # Ensure it's always an odd number
        blurred_image = cv2.GaussianBlur(img_array, kernel_size, 0)

        col1, col2 = st.columns(2)
        with col1:
            st.image(image, caption="Original Image", use_container_width=True)
        with col2:
            st.image(blurred_image, caption=f"Blurred (Level {blur_level})", use_container_width=True)

    # Convert to Color Space (Fixed BGR to RGB)
    elif option == "Convert to Color Space":
        color_space = st.selectbox("Choose a color space:", ["RGB", "BGR2RGB", "Grayscale"])

        converted_image = None  # Initialize to avoid `NoneType` errors

        if color_space == "RGB":
            converted_image = img_array  # Already in RGB format
        elif color_space == "BGR2RGB":
            converted_image = cv2.cvtColor(img_array, cv2.COLOR_BGR2RGB)  # Fixed BGR to RGB conversion
        elif color_space == "Grayscale":
            converted_image = cv2.cvtColor(img_array, cv2.COLOR_RGB2GRAY)

        if converted_image is not None:
            col1, col2 = st.columns(2)
            with col1:
                st.image(image, caption="Original Image", use_container_width=True)
            with col2:
                st.image(converted_image, caption=f"{color_space} Image", use_container_width=True)

    # Edge Detection (Canny)
    elif option == "Edge Detection":
        low_threshold = st.slider("Lower Threshold", 0, 255, 50)
        high_threshold = st.slider("Upper Threshold", 0, 255, 150)

        gray_image = cv2.cvtColor(img_array, cv2.COLOR_RGB2GRAY)  # Convert to grayscale first
        edges = cv2.Canny(gray_image, low_threshold, high_threshold)  # Apply Canny Edge Detection

        col1, col2 = st.columns(2)
        with col1:
            st.image(image, caption="Original Image", use_container_width=True)
        with col2:
            st.image(edges, caption="Edge Detection (Canny)", use_container_width=True)