app.py

import streamlit as st
import os
import time
import logging
from moviepy.video.io.VideoFileClip import VideoFileClip
import whisper
from transformers import pipeline
import cv2
import numpy as np

# Set up logging configuration
logging.basicConfig(level=logging.INFO, format='%(asctime)s - %(levelname)s - %(message)s')
logger = logging.getLogger()

# Set up directories
UPLOAD_FOLDER = './uploads'
PROCESSED_FOLDER = './processed'
os.makedirs(UPLOAD_FOLDER, exist_ok=True)
os.makedirs(PROCESSED_FOLDER, exist_ok=True)

# Initialize the summarizer model from Hugging Face
summarizer = pipeline("summarization", model="meta-llama/Llama-3.2-1B")

# Function to extract audio from video using MoviePy
def extract_audio_from_video(video_file_path, audio_file_path):
    try:
        logger.info(f"Extracting audio from video: {video_file_path}")
        video_clip = VideoFileClip(video_file_path)
        audio_clip = video_clip.audio
        audio_clip.write_audiofile(audio_file_path)
        logger.info(f"Audio extraction completed and saved to: {audio_file_path}")
    except Exception as e:
        logger.error(f"Error extracting audio from video: {e}")

# Function to transcribe audio using Whisper
def transcribe_audio(audio_file_path):
    try:
        logger.info(f"Transcribing audio: {audio_file_path}")
        model = whisper.load_model("base")  # Load the Whisper model
        result = model.transcribe(audio_file_path)
        logger.info("Audio transcription completed.")
        return result['text']
    except Exception as e:
        logger.error(f"Error transcribing audio: {e}")
        return ""

# Function to summarize the transcription text using Hugging Face Transformers
def split_text_into_chunks(text, chunk_size=1000):
    chunks = []
    for i in range(0, len(text), chunk_size):
        chunks.append(text[i:i+chunk_size])
    return chunks

def summarize_text(text):
    try:
        chunks = split_text_into_chunks(text)
        summaries = []
        for chunk in chunks:
            summary = summarizer(chunk, max_new_tokens=500, min_length=50, do_sample=False)
            summaries.append(summary[0]['summary_text'])
        return " ".join(summaries)
    except Exception as e:
        st.error(f"Error summarizing text: {e}")
        return "Error occurred during summarization."


# Function to detect scene changes in the video using OpenCV (optional)
def detect_scene_changes(video_file_path):
    try:
        logger.info(f"Detecting scene changes in video: {video_file_path}")
        cap = cv2.VideoCapture(video_file_path)
        ret, prev_frame = cap.read()
        prev_frame_gray = cv2.cvtColor(prev_frame, cv2.COLOR_BGR2GRAY)
        
        scene_changes = []
        frame_count = 0
        
        while ret:
            ret, frame = cap.read()
            if not ret:
                break
            frame_gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
            diff = cv2.absdiff(frame_gray, prev_frame_gray)
            non_zero_count = np.count_nonzero(diff)
            
            if non_zero_count > 1000000:  # Threshold for detecting scene change
                scene_changes.append(frame_count)
            
            prev_frame_gray = frame_gray
            frame_count += 1

        cap.release()
        logger.info(f"Scene changes detected: {scene_changes}")
        return scene_changes
    except Exception as e:
        logger.error(f"Error detecting scene changes: {e}")
        return []

# Streamlit App
st.title("Automated Video Summarization")

# File upload widget
uploaded_file = st.file_uploader("Choose a video file", type=["mp4", "avi", "mov"])

if uploaded_file is not None:
    try:
        # Save the uploaded file to the upload folder with a timestamp
        timestamp = int(time.time())
        filename = f"{timestamp}_{uploaded_file.name}"
        video_path = os.path.join(UPLOAD_FOLDER, filename)

        # Save the uploaded file
        with open(video_path, "wb") as f:
            f.write(uploaded_file.getbuffer())
        
        logger.info(f"File uploaded successfully: {filename}")
        st.write(f"File uploaded successfully: {filename}")

        # Extract audio from the video
        audio_file_path = os.path.join(PROCESSED_FOLDER, f"{filename}.wav")
        extract_audio_from_video(video_path, audio_file_path)
        st.write("Audio extracted successfully.")

        # Transcribe the audio
        transcription = transcribe_audio(audio_file_path)
        st.write("Transcription completed.")

        # Add a spinner while the summarization is happening
        with st.spinner('Generating summary, please wait...'):
            summary = summarize_text(transcription)
        st.subheader("Summary")
        st.write(summary)

        # Optional: Detect scene changes in the video
        scene_changes = detect_scene_changes(video_path)
        st.subheader("Scene Changes Detected at Frames")
        st.write(scene_changes)

    except Exception as e:
        logger.error(f"An error occurred: {e}")
        st.error(f"An error occurred: {e}")