Text-to-speech generation script

text_to_speech.py

import argparse
import asyncio
import os
from pathlib import Path

from openai import AsyncOpenAI

openai = AsyncOpenAI()

async def text_to_speech(text_file: str, instructions_file: str = 'instructions.txt', voice: str = 'onyx') -> None:
    # Read the text file
    with open(text_file, 'r', encoding='utf-8') as f:
        input_text = f.read()

    # Read instructions if file exists, otherwise use empty string
    instructions = ''
    if os.path.exists(instructions_file):
        with open(instructions_file, 'r', encoding='utf-8') as f:
            instructions = f.read()

    # Generate output filename by replacing extension with .mp3
    output_file = str(Path(text_file).with_suffix('.mp3'))

    # Generate speech
    async with openai.audio.speech.with_streaming_response.create(
        model="gpt-4o-mini-tts",
        voice=voice,
        input=input_text,
        instructions=instructions,
        response_format="mp3",
    ) as response:
        with open(output_file, "wb") as f:
            async for chunk in response.iter_bytes():
                f.write(chunk)

    print(f"Created {output_file}")

def main():
    parser = argparse.ArgumentParser(description='Convert text file to speech MP3')
    parser.add_argument('text_file', help='Input text file to convert')
    parser.add_argument('--instructions', '-i',
                      default='instructions.txt',
                      help='File containing voice instructions (default: instructions.txt)')
    parser.add_argument('--voice', '-v',
                      default='onyx',
                      choices=['alloy', 'echo', 'fable', 'onyx', 'nova', 'shimmer'],
                      help='Voice to use for the speech (default: onyx)')

    args = parser.parse_args()

    asyncio.run(text_to_speech(args.text_file, args.instructions, args.voice))

if __name__ == "__main__":
    main()