main.py

from TTS.api import TTS
import whisper
from tempfile import NamedTemporaryFile
import streamlit as st
# from faster_whisper import WhisperModel
from googletrans import Translator
import nltk







audio = st.file_uploader("Upload an audio file", type=["wav"])

if audio is not None:

    
    with NamedTemporaryFile(delete=False, suffix="wav") as temp_audio:
        
        temp_audio.write(audio.read())
        temp_audio_path = temp_audio.name

    

    st.write(f"Temporary audio file saved at: {temp_audio_path}")

        
    tts = TTS("tts_models/multilingual/multi-dataset/xtts_v2")

    model = whisper.load_model("base")
    result = model.transcribe(temp_audio_path)
    print(result["text"])
    fin=[]
    sentences = result["text"]
    translator = Translator()
    # for sentence in sentences:
    #     conv=translator.translate(result, dest='fr')
    #     fin+=conv.pronunciation


    
    conv=translator.translate(sentences, dest='fr')
    print(conv)
    st.write(conv.text)

    


        
    tts.tts_to_file(
            text=conv.text,
            file_path="output.wav",
            speaker_wav=temp_audio_path,
            language="fr"
        )

    st.audio("output.wav", format="audio/wav")