src/components/text_to_speech_gtts.py

"""
Created By: ishwor subedi
Date: 2024-07-31
"""
import base64
from io import BytesIO

from gtts import gTTS
from gtts.tokenizer import pre_processors


class TextToSpeech:
    def __init__(self):
        self.preprocessing = [pre_processors.tone_marks, pre_processors.end_of_line, pre_processors.word_sub,
                              pre_processors.abbreviations]

    def conversion(self, text: str, lang: str, tld: str) -> str:
        """
        Convert text to speech and return the Base64-encoded MP3 data.
        :param text: The text to convert to speech.
        :param lang: The language in which to convert the text.
        :return: Base64-encoded MP3 data as a string.
        """
        tts = gTTS(text=text, lang=lang, slow=False, tld=tld, pre_processor_funcs=self.preprocessing)
        mp3_fp = BytesIO()
        tts.write_to_fp(mp3_fp)
        mp3_fp.seek(0)

        mp3_binary = mp3_fp.getvalue()

        base64_mp3 = base64.b64encode(mp3_binary).decode("utf-8")

        return base64_mp3