app.py

import gradio as gr
from pytube import YouTube
import subprocess
from huggingsound import SpeechRecognitionModel
import torch
from transformers import pipeline

def process_video(video_url):
    response = {
        'status': 'Success',
        'message': '',
        'data': ''
    }

    try:
        yt = YouTube(video_url)
        audio_file = yt.streams.filter(only_audio=True, file_extension='mp4').first().download(filename='ytaudio.mp4')
        subprocess.run(['ffmpeg', '-i', 'ytaudio.mp4', '-acodec', 'pcm_s16le', '-ar', '16000', 'ytaudio.wav'], check=True)
    except Exception as e:
        response['status'] = 'Error'
        response['message'] = f'Failed to download and convert video: {str(e)}'
        return response

    try:
        device = "cuda" if torch.cuda.is_available() else "cpu"
        model = SpeechRecognitionModel("jonatasgrosman/wav2vec2-large-xlsr-53-english", device=device)
        transcription = model.transcribe(['ytaudio.wav'])[0]['transcription']
    except Exception as e:
        response['status'] = 'Error'
        response['message'] = f'Failed during speech recognition: {str(e)}'
        return response

    try:
        summarization = pipeline('summarization')
        summarized_text = summarization(transcription, max_length=130, min_length=30, do_sample=False)
        response['data'] = summarized_text[0]['summary_text']
    except Exception as e:
        response['status'] = 'Error'
        response['message'] = f'Failed during summarization: {str(e)}'
        return response

    return response

iface = gr.Interface(
    fn=process_video,
    inputs=gr.inputs.Textbox(lines=2, placeholder="Enter YouTube Video URL Here..."),
    outputs=[
        gr.outputs.Textbox(label="Status"),
        gr.outputs.Textbox(label="Message"),
        gr.outputs.Textbox(label="Summary")
    ],
    title="YouTube Video Summarizer",
    description="This tool extracts audio from a YouTube video, transcribes it, and provides a summary.",
    enable_queue=True  # Enable request queuing
)

iface.launch()