app.py

import os
import re
from youtube_transcript_api import YouTubeTranscriptApi
from youtube_transcript_api.formatters import TextFormatter
from transformers import pipeline
import torch

# Enable CUDA debugging
os.environ['CUDA_LAUNCH_BLOCKING'] = '1'

# Initialize summarization pipeline
text_summary = pipeline("summarization", model="sshleifer/distilbart-cnn-12-6", torch_dtype=torch.float32, device=-1)

def preprocess_input(input_text, max_length=1024):
    tokens = len(input_text.split())
    if tokens > max_length:
        input_text = " ".join(input_text.split()[:max_length])
    return input_text

def split_text(text, max_length=1024):
    return [text[i:i + max_length] for i in range(0, len(text), max_length)]

def summary(input):
    chunks = split_text(input)
    summaries = [text_summary(chunk)[0]['summary_text'] for chunk in chunks]
    return " ".join(summaries)

def extract_video_id(url):
    regex = r"(?:youtube\.com\/(?:[^\/\n\s]+\/\S+\/|(?:v|e(?:mbed)?)\/|\S*?[?&]v=)|youtu\.be\/)([a-zA-Z0-9_-]{11})"
    match = re.search(regex, url)
    if match:
        return match.group(1)
    return None

def get_youtube_transcript(video_url):
    video_id = extract_video_id(video_url)
    if not video_id:
        return "Video ID could not be extracted."

    try:
        transcript = YouTubeTranscriptApi.get_transcript(video_id)
        formatter = TextFormatter()
        text_transcript = formatter.format_transcript(transcript)
        text_transcript = preprocess_input(text_transcript)
        return summary(text_transcript)
    except Exception as e:
        return f"An error occurred: {e}"

# Gradio interface
import gradio as gr
gr.close_all()

demo = gr.Interface(
    fn=get_youtube_transcript,
    inputs=[gr.Textbox(label="Input YouTube Url to summarize", lines=1)],
    outputs=[gr.Textbox(label="Summarized text", lines=4)],
    title="YouTube Script Summarizer",
    description="THIS APPLICATION WILL BE USED TO SUMMARIZE THE YOUTUBE VIDEO SCRIPT."
)

demo.launch()