speak.py

#!/usr/bin/env python3

import os
import sys
import subprocess
from datetime import datetime
from urllib.request import Request, urlopen
from urllib.error import HTTPError
import json

TTS_API_BASE_URL = os.getenv('TTS_API_BASE_URL', 'https://api.openai.com/v1')
TTS_API_KEY = os.getenv('TTS_API_KEY') or os.getenv('OPENAI_API_KEY')
TTS_MODEL = os.getenv('TTS_MODEL')
TTS_VOICE = os.getenv('TTS_VOICE')
TTS_LANG = os.getenv('TTS_LANG')
TTS_AUDIO_FORMAT = os.getenv('TTS_AUDIO_FORMAT', 'mp3')
TTS_DEBUG = os.getenv('TTS_DEBUG')

def speak(input: str) -> bytes:
    model = TTS_MODEL or 'kokoro'
    voice = TTS_VOICE or 'af_bella'
    format = TTS_AUDIO_FORMAT
    lang = TTS_LANG
    response_format = format
    speed = 1.0

    body: dict[str, str | float] = {
        'input': input,
        'model': model,
        'voice': voice,
        'format': format,
        'response_format': response_format,
        'speed': speed,
        'lang': lang
    }

    url = f"{TTS_API_BASE_URL}/audio/speech"
    headers = {'Content-Type': 'application/json'}
    if TTS_API_KEY:
        headers['Authorization'] = f"Bearer {TTS_API_KEY}"

    try:
        timestamp = datetime.now().timestamp() * 1000
        if TTS_DEBUG:
            print(f"Generating audio from text: {len(input)} characters...")
        
        req = Request(url, data=json.dumps(body).encode('utf-8'), headers=headers, method='POST')
        with urlopen(req) as response:
            if response.status != 200:
                raise Exception(f"HTTP error with the status: {response.status}")
            buffer = response.read()
        
        elapsed = datetime.now().timestamp() * 1000 - timestamp
        if TTS_DEBUG:
            print(f"Audio generation took {elapsed:.0f} ms.")
        return buffer
    except HTTPError as e:
        if TTS_DEBUG:
            print(f"Error: {e}", file=sys.stderr)
        raise
    except Exception as e:
        if TTS_DEBUG:
            print(f"Error: {e}", file=sys.stderr)
        raise

def main():
    args = sys.argv[1:]
    input = ' '.join(args)
    
    if not input.strip():
        if not sys.stdin.isatty():
            input = sys.stdin.read().strip()
    
    if not input.strip():
        print('Usage:')
        print()
        print('./speak.py "How are you?"')
        print('echo "Good morning" | ./speak.py')
        sys.exit(1)
    
    filename = datetime.now().isoformat()[:16].replace('-', '').replace(':', '').replace('T', '') + f'.{TTS_AUDIO_FORMAT}'
    buffer = speak(input)
    
    with open(filename, 'wb') as f:
        f.write(buffer)
    
    if TTS_DEBUG:
        print(f"Saved {len(buffer)} bytes to {filename}")
    
    try:
        speaker = subprocess.Popen(['play', filename])
        speaker.wait()
    except FileNotFoundError:
        print(f"Failed to play audio: 'play' command not found", file=sys.stderr)
    except Exception as err:
        print(f"Failed to play audio: {err}", file=sys.stderr)

if __name__ == '__main__':
    main()