app.py

import whisper
from transformers import pipeline
import ffmpeg
import os
import streamlit as st
import torch

# Step 1: File Upload in Streamlit
uploaded_file = st.file_uploader("Upload a video file", type=["mp4", "mkv", "avi", "mov"])

if uploaded_file is not None:
    # Save the uploaded video file
    video_path = "uploaded_video.mp4"
    with open(video_path, "wb") as f:
        f.write(uploaded_file.getbuffer())
    
    # Step 2: Process the Video
    # Load Whisper Model for Transcription
    whisper_model = whisper.load_model("base")
    
    # Extract audio from video using ffmpeg
    audio_path = "audio.wav"
    ffmpeg.input(video_path).output(audio_path).run(overwrite_output=True)
    
    # Transcribe Audio
    transcription = whisper_model.transcribe(audio_path)["text"]
    st.write("Transcription:\n", transcription)
    
    # Summarize Transcription using Hugging Face Transformers
    summarizer = pipeline("summarization", model="facebook/bart-large-cnn", device=0 if torch.cuda.is_available() else -1)
    summary = summarizer(transcription, max_length=150, min_length=30, do_sample=False)[0]["summary_text"]
    st.write("\nSummary:\n", summary)

    # Cleanup
    os.remove(audio_path)
    os.remove(video_path)