app.py

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

# App title and emojis
st.title("📸 Image Processing App 🎨")
st.write("Upload an image, apply filters, and download the processed image! ✨")

# Sidebar for feature selection
st.sidebar.header("🎨 Filter Options")
filter_type = st.sidebar.selectbox(
    "Select a filter:",
    [
        "Blur",
        "Edge Detection (Canny)",
        "Grayscale",
        "Sharpen",
        "Sepia",
        "Invert Colors",
        "Brightness Adjustment",
        "Contrast Adjustment",
    ],
)

# Upload image
uploaded_file = st.file_uploader(".Upload an image 🖼️", type=["jpg", "jpeg", "png"])

if uploaded_file is not None:
    # Convert the file to an OpenCV image
    file_bytes = np.asarray(bytearray(uploaded_file.read()), dtype=np.uint8)
    img = cv2.imdecode(file_bytes, 1)
    img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)  # Convert BGR to RGB

    # Display original image
    col1, col2 = st.columns(2)
    with col1:
        st.subheader("Original Image 📷")
        st.image(img, use_container_width=True)

    # Apply filters based on user selection
    processed_img = img.copy()

    if filter_type == "Blur":
        blur_type = st.sidebar.selectbox(
            "Blur Type", ["Gaussian Blur", "Average Blur", "Median Blur", "Bilateral Filter"]
        )
        if blur_type == "Gaussian Blur":
            kernel_size = st.sidebar.slider("Kernel Size 🔍", 1, 31, 5, step=2)
            sigma_x = st.sidebar.slider("Sigma X (Standard Deviation)", 0, 10, 0)
            processed_img = cv2.GaussianBlur(processed_img, (kernel_size, kernel_size), sigma_x)

        elif blur_type == "Average Blur":
            kernel_size = st.sidebar.slider("Kernel Size 🔍", 1, 31, 5, step=2)
            processed_img = cv2.blur(processed_img, (kernel_size, kernel_size))

        elif blur_type == "Median Blur":
            kernel_size = st.sidebar.slider("Kernel Size 🔍", 1, 31, 5, step=2)
            processed_img = cv2.medianBlur(processed_img, kernel_size)

        elif blur_type == "Bilateral Filter":
            diameter = st.sidebar.slider("Diameter", 1, 31, 9, step=2)
            sigma_color = st.sidebar.slider("Sigma Color", 1, 200, 75)
            sigma_space = st.sidebar.slider("Sigma Space", 1, 200, 75)
            processed_img = cv2.bilateralFilter(processed_img, diameter, sigma_color, sigma_space)

    elif filter_type == "Edge Detection (Canny)":
        lower_threshold = st.sidebar.slider("Lower Threshold ⬇️", 0, 255, 100)
        upper_threshold = st.sidebar.slider("Upper Threshold ⬆️", 0, 255, 200)
        processed_img = cv2.Canny(processed_img, lower_threshold, upper_threshold)

    elif filter_type == "Grayscale":
        grayscale_method = st.sidebar.selectbox(
            "Grayscale Method", ["Weighted Average", "Simple Average"]
        )
        if grayscale_method == "Weighted Average":
            processed_img = cv2.cvtColor(processed_img, cv2.COLOR_RGB2GRAY)
        else:
            processed_img = np.mean(processed_img, axis=2).astype(np.uint8)

    elif filter_type == "Sharpen":
        sharpness = st.sidebar.slider("Sharpness Intensity", 1.0, 10.0, 1.0)
        kernel = np.array([[0, -1, 0], [-1, 5, -1], [0, -1, 0]]) * sharpness
        processed_img = cv2.filter2D(processed_img, -1, kernel)

    elif filter_type == "Sepia":
        sepia_intensity = st.sidebar.slider("Sepia Intensity", 0.1, 1.0, 1.0)
        sepia_filter = (
            np.array(
                [[0.272, 0.534, 0.131], [0.349, 0.686, 0.168], [0.393, 0.769, 0.189]]
            )
            * sepia_intensity
        )
        processed_img = cv2.transform(processed_img, sepia_filter)
        processed_img = np.clip(processed_img, 0, 255)

    elif filter_type == "Invert Colors":
        blend_factor = st.sidebar.slider("Inversion Blend Factor", 0.0, 1.0, 1.0)
        inverted_img = 255 - processed_img
        processed_img = cv2.addWeighted(
            processed_img, 1 - blend_factor, inverted_img, blend_factor, 0
        )

    elif filter_type == "Brightness Adjustment":
        brightness = st.sidebar.slider("Brightness ☀️", -100, 100, 0)
        processed_img = cv2.convertScaleAbs(processed_img, alpha=1, beta=brightness)

    elif filter_type == "Contrast Adjustment":
        contrast = st.sidebar.slider("Contrast ⚡", 0.1, 3.0, 1.0, step=0.1)
        processed_img = cv2.convertScaleAbs(processed_img, alpha=contrast, beta=0)

    # Display processed image
    with col2:
        st.subheader("Processed Image 🎯")
        st.image(processed_img, use_container_width=True)

    # Download button for processed image
    st.markdown("---")
    st.subheader("Download Processed Image 💾")
    buf = io.BytesIO()
    pil_img = Image.fromarray(processed_img)
    pil_img.save(buf, format="PNG")
    byte_im = buf.getvalue()
    st.download_button(
        label="📥 Download Image",
        data=byte_im,
        file_name="processed_image.png",
        mime="image/png",
    )

else:
    st.info("Please upload an image to get started. 📂")