services/tts_service.py

import logging
import io
from gtts import gTTS
import numpy as np
import wave

logger = logging.getLogger(__name__)

async def generate_tts(text: str) -> bytes:
    """
    Convert text to speech using Google Text-to-Speech (gTTS).
    Returns MP3 audio bytes.
    """
    try:
        if not text or len(text) > 500:
            raise ValueError("Text must be between 1-500 characters")
        
        logger.info(f"TTS: Generating audio for '{text[:50]}...'")
        
        # Generate audio using gTTS
        tts = gTTS(text=text, lang='en', slow=False)
        
        # Save to bytes buffer
        audio_buffer = io.BytesIO()
        tts.write_to_fp(audio_buffer)
        audio_bytes = audio_buffer.getvalue()
        
        logger.info(f"✓ TTS Success: {len(audio_bytes)} bytes generated")
        return audio_bytes
        
    except Exception as e:
        logger.error(f"✗ TTS Error: {str(e)}")
        # Return fallback silent audio
        return generate_silent_wav()

def generate_silent_wav() -> bytes:
    """Generate 1-second silent WAV file as fallback"""
    sample_rate = 22050
    duration = 1.0
    silence = np.zeros(int(sample_rate * duration), dtype=np.int16)
    
    buffer = io.BytesIO()
    with wave.open(buffer, 'wb') as wav:
        wav.setnchannels(1)
        wav.setsampwidth(2)
        wav.setframerate(sample_rate)
        wav.writeframes(silence.tobytes())
    
    return buffer.getvalue()