application/utils/text_to_speech.py

# application/utils/text_to_speech.py - NEW FILE
import asyncio
import edge_tts
from aiohttp import web  # Import aiohttp

async def generate_tts(text: str) -> bytes:
    """
    Generates speech from text using edge-tts and returns the audio as a byte stream.
    """
    try:
        communicate = edge_tts.Communicate(text, "en-US-GuyNeural")  # Or any other preferred voice
        return await communicate.stream()

    except Exception as e:
        print(f"Error in TTS generation: {e}")
        return b""  # Return empty bytes on error


async def tts_handler(request: web.Request) -> web.Response:
    """
    aiohttp handler for the /tts endpoint.  This is the correct way to handle
    streaming with aiohttp (and thus edge-tts).
    """
    text = request.query.get('text')
    if not text:
        return web.Response(status=400, text="No text provided")

    try:
        response = web.StreamResponse()
        response.content_type = 'audio/mpeg'
        await response.prepare(request)

        async for chunk in generate_tts(text):
            await response.write(chunk)

        await response.write_eof()
        return response
    except Exception as e:
        print(f"Error in TTS handler: {e}")
        return web.Response(status=500, text=str(e))