app.py

"""
Created on Mon Mar 28 01:04:50 2022
@author: adeep
"""
from fnmatch import translate
import cv2 as cv
import tempfile
import numpy as np
import pandas as pd
import streamlit as st 
import joblib
import os
from moviepy.editor import VideoFileClip
import speech_recognition as sr
from pydub import AudioSegment
from pydub.silence import split_on_silence
import transformers
from transformers import pipeline
import nltk
nltk.download('punkt')
nltk.download('averaged_perceptron_tagger')
import nltk
nltk.download('punkt')
nltk.download('averaged_perceptron_tagger')
from nltk.tokenize import sent_tokenize
import re
from utils import get_translation, welcome, get_large_audio_transcription

from PIL import Image

#import stanfordnlp

def main():
 
    
    st.title("Summarize Text")
    video = st.file_uploader("Choose a file", type=['mp4'])
    button = st.button("Summarize")
    
    max_c = st.sidebar.slider('Select max words', 50, 500, step=10, value=150)
    min_c = st.sidebar.slider('Select min words', 10, 450, step=10, value=50)
    gen_summ = False
    
 

    with st.spinner("Running.."):
    
        if button and video:
            tfile = tempfile.NamedTemporaryFile(delete=False)
            tfile.write(video.read())
            #st.write(tfile.name)
            v = VideoFileClip(tfile.name)
            v.audio.write_audiofile("movie.wav")
            #st.video(video, format="video/mp4", start_time=0)
            #st.audio("movie.wav")
            whole_text=get_large_audio_transcription("movie.wav")
            #st.write(whole_text)
            #summarizer = pipeline("summarization")
            #summarizer = pipeline("summarization", model="t5-base", tokenizer="t5-base", framework="pt")
            summarizer = pipeline("summarization", model="t5-large", tokenizer="t5-large", framework="pt")
            summarized = summarizer(whole_text, min_length=min_c, max_length=max_c)
            summ=summarized[0]['summary_text']
            #st.write(summ) 
            gen_summ = True
            #stf_nlp = stanfordnlp.Pipeline(processors='tokenize,mwt,pos')
            #doc = stf_nlp(summ)
            #l=[w.text.capitalize() if w.upos in ["PROPN","NNS"] else w.text for sent in doc.sentences for w in sent.words]
            #text=" ".join(l)
            #summ=truecasing_by_sentence_segmentation(summ)  
            sentences = sent_tokenize(summ, language='english')
            # capitalize the sentences
            sentences_capitalized = [s.capitalize() for s in sentences]
            # join the capitalized sentences
            summ = re.sub(" (?=[\.,'!?:;])", "", ' '.join(sentences_capitalized))
            
            if 'summary' not in st.session_state:
                st.session_state.summary=True
                st.session_state.summarization = summ
                st.session_state.gen_summ = True
                


    translate = st.sidebar.radio('Do you want to translate the text to any different language?', ('No', 'Yes'))
    if 'summary' in st.session_state:
            summarized_text = st.session_state.summarization
            st.write(summarized_text)
            gen_summ = st.session_state.gen_summ
            
    if translate == 'Yes' and gen_summ == True:
        lang_list  = ['Hindi', 'Marathi', 'Malayalam', 'Kannada', 'Telugu', 'Tamil', 'Oriya', 'Bengali', 'Gujarati', 'Urdu']
    
        s_type = st.sidebar.selectbox('Select the Language in which you want to Translate:',lang_list) 
        st.sidebar.write('You selected:', s_type)

         
        translation = get_translation(source='English', dest=s_type, text=summarized_text)

        st.sidebar.write(translation)
    elif translate == 'Yes' and gen_summ == False:
        st.error("The summary has not been generated yet. Please generate the summary first and then translate")
      
    else:
         st.write('')
             
if __name__ == '__main__':
    
    main()