app.py

!pip install pytubefix moviepy transformers gradio torch

import os
import torch
from pytubefix import YouTube
from moviepy.editor import VideoFileClip
from transformers import pipeline

# ---- STEP 1: Download YouTube Video ----
url = "https://www.youtube.com/watch?v=VgxnyKnB3qc&ab"
yt = YouTube(url)
title = yt.title
print(f"Downloading: {title}")

video_stream = yt.streams.get_highest_resolution()
video_path = f"/content/{title}.mp4"
video_stream.download(filename=video_path)

print(f"Video saved as: {video_path}")

# ---- STEP 2: Extract Audio from Video ----
output_audio = f"/content/{title}.wav"

video = VideoFileClip(video_path)
video.audio.write_audiofile(output_audio)

print(f"Audio extracted: {output_audio}")

# ---- STEP 3: Transcribe Audio ----
asr = pipeline(task="automatic-speech-recognition", model="distil-whisper/distil-small.en")

def transcribe_audio(audio_file):
    print("Transcribing audio...")
    transcription_result = asr(audio_file)
    transcribed_text = transcription_result["text"]
    return transcribed_text

transcribed_text = transcribe_audio(output_audio)
print("Transcription Complete:\n", transcribed_text[:500])  # Preview first 500 characters

# ---- STEP 4: Summarize Transcription ----
summarizer = pipeline("summarization", model="facebook/bart-large-cnn")

def summarize_text(text):
    if len(text.split()) < 50:
        return "Text too short to summarize."
    
    print("Summarizing text...")
    summary_result = summarizer(text, max_length=100, min_length=30, do_sample=False)
    return summary_result[0]['summary_text']

summarized_text = summarize_text(transcribed_text)
print("\nSummary:\n", summarized_text)

# ---- OPTIONAL: Save Results to File ----
with open(f"/content/{title}_transcription.txt", "w") as f:
    f.write(transcribed_text)

with open(f"/content/{title}_summary.txt", "w") as f:
    f.write(summarized_text)

print("Transcription & Summary saved!")