app.py

import streamlit as st
import cv2
import numpy as np
import datetime
import os
import time
import base64
import re
import glob
from camera_input_live import camera_input_live
import face_recognition

# Set wide layout
st.set_page_config(layout="wide")

# Decorator for caching images
def get_image_count():
    return {'count': 0}

# Function Definitions for Camera Feature
def save_image(image, image_count):
    timestamp = datetime.datetime.now().strftime("%Y%m%d_%H%M%S")
    filename = f"captured_image_{timestamp}_{image_count['count']}.png"
    image_count['count'] += 1
    bytes_data = image.getvalue()
    cv2_img = cv2.imdecode(np.frombuffer(bytes_data, np.uint8), cv2.IMREAD_COLOR)
    cv2.imwrite(filename, cv2_img)
    return filename

def get_image_base64(image_path):
    with open(image_path, "rb") as image_file:
        return base64.b64encode(image_file.read()).decode()

# Function Definitions for Chord Sheet Feature
def process_line(line):
    if re.search(r'\b[A-G][#b]?m?\b', line):
        line = re.sub(r'\b([A-G][#b]?m?)\b', r"<img src='\1.png' style='height:20px;'>", line)
    return line

def process_sheet(sheet):
    processed_lines = []
    for line in sheet.split('\n'):
        processed_line = process_line(line)
        processed_lines.append(processed_line)
    return '<br>'.join(processed_lines)

# Load a sample image and learn how to recognize it
known_image = face_recognition.load_image_file("known_face.jpg")
known_encoding = face_recognition.face_encodings(known_image)[0]

# Main Function
def main():
    # Layout Configuration
    col1, col2 = st.columns([2, 3])

    # Camera Section
    with col1:
        st.markdown("✨ Magic Lens: Real-Time Camera Stream 🌈")

        snapshot_interval = st.slider("Snapshot Interval (seconds)", 1, 10, 5)
        image_placeholder = st.empty()

        if 'captured_images' not in st.session_state:
            st.session_state['captured_images'] = []
        if 'last_captured' not in st.session_state:
            st.session_state['last_captured'] = time.time()

        image = camera_input_live()
        if image is not None:
            # Convert the image to RGB format for face_recognition
            rgb_image = cv2.cvtColor(np.array(image), cv2.COLOR_BGR2RGB)

            # Detect faces in the image
            face_locations = face_recognition.face_locations(rgb_image)
            face_encodings = face_recognition.face_encodings(rgb_image, face_locations)

            # Iterate over detected faces and compare with known face
            for (top, right, bottom, left), face_encoding in zip(face_locations, face_encodings):
                matches = face_recognition.compare_faces([known_encoding], face_encoding)

                if True in matches:
                    # If a match is found, draw a green rectangle and label
                    cv2.rectangle(rgb_image, (left, top), (right, bottom), (0, 255, 0), 2)
                    cv2.putText(rgb_image, "Known Face", (left, top - 10), cv2.FONT_HERSHEY_SIMPLEX, 0.9, (0, 255, 0), 2)
                else:
                    # If no match, draw a red rectangle
                    cv2.rectangle(rgb_image, (left, top), (right, bottom), (0, 0, 255), 2)

            # Convert the RGB image back to BGR format for display
            bgr_image = cv2.cvtColor(rgb_image, cv2.COLOR_RGB2BGR)
            image_placeholder.image(bgr_image, channels="BGR")

            if time.time() - st.session_state['last_captured'] > snapshot_interval:
                image_count = get_image_count()
                filename = save_image(image, image_count)
                st.session_state['captured_images'].append(filename)
                st.session_state['last_captured'] = time.time()

        sidebar_html = "<div style='display:flex;flex-direction:column;'>"
        for img_file in st.session_state['captured_images']:
            image_base64 = get_image_base64(img_file)
            sidebar_html += f"<img src='data:image/png;base64,{image_base64}' style='width:100px;'><br>"
        sidebar_html += "</div>"
        st.sidebar.markdown("## Captured Images")
        st.sidebar.markdown(sidebar_html, unsafe_allow_html=True)

        # JavaScript Timer
        st.markdown(f"<script>setInterval(function() {{ document.getElementById('timer').innerHTML = new Date().toLocaleTimeString(); }}, 1000);</script><div>Current Time: <span id='timer'></span></div>", unsafe_allow_html=True)

    # Chord Sheet Section
    with col2:
        st.markdown("## 🎬 Action! Real-Time Camera Stream Highlights 📽️")

        all_files = [f for f in glob.glob("*.png") if ' by ' in f]
        selected_file = st.selectbox("Choose a Dataset:", all_files)

        if selected_file:
            with open(selected_file, 'r', encoding='utf-8') as file:
                sheet = file.read()
            st.markdown(process_sheet(sheet), unsafe_allow_html=True)

    # Trigger a rerun only when the snapshot interval is reached
    if 'last_captured' in st.session_state and time.time() - st.session_state['last_captured'] > snapshot_interval:
        st.experimental_rerun()

if __name__ == "__main__":
    main()