app.py

from transformers import pipeline
from youtube_transcript_api import YouTubeTranscriptApi
from keybert import KeyBERT
import gradio as gr
from keyphrase_vectorizers import KeyphraseCountVectorizer
import requests
from bs4 import BeautifulSoup

def summarize_transcript(url):
    video_id = url.split("=")[1]
        
    transcript = YouTubeTranscriptApi.get_transcript(video_id)

    result = ""
    for i in transcript:
        result += ' ' + i['text']

    summarizer = pipeline('summarization')
    
    num_iters = int(len(result)/1000)
    summarized_text = []
    for i in range(0, num_iters + 1):
      start = 0
      start = i * 1000
      end = (i + 1) * 1000
      out = summarizer(result[start:end])
      out = out[0]
      out = out['summary_text']
      summarized_text.append(out)
      summ = str(summarized_text)
      print(summ)

      #keywords
      words = []
      kw_model = KeyBERT()
      keywords = kw_model.extract_keywords(summ)
      w = kw_model.extract_keywords(summ, vectorizer=KeyphraseCountVectorizer())
      for s in w:
        words.append(s[0])

      #tags
      request = requests.get(url)
      html = BeautifulSoup(request.content,"html.parser")
      tags = html.find_all("meta",property = "og:video:tag")
      lst = []
      for tag in tags:
        lst.append(tag['content'])
    return (summ,words,lst)
    
gradio_ui = gr.Interface(fn = summarize_transcript,
                  inputs = [gr.inputs.Textbox(label = "Enter the YouTube URL below:")],
                  outputs = [gr.outputs.Textbox(label = "Transcript Summary"),gr.outputs.Textbox(label = "Keywords"),gr.outputs.Textbox(label = "Hash Tags")],
                  title = "YouTube Transcript Summarizer",
                  theme = "grass",
                  description = "Here You can see the SUMMARY,KEYWORDS and HASHTAGS of the YouTube video you want to watch")
               
gradio_ui.launch(inline = False)