routes/audio.py

"""Speech-to-Text & Text-to-Speech API Endpoints"""

from fastapi import APIRouter, UploadFile, File, HTTPException
from fastapi.responses import StreamingResponse
import logging

from models.audio import (
    STTResponse,
    TTSRequest,
    TTSResponse
)

from services.stt_service import transcribe_audio
from services.tts_service import text_to_speech

router = APIRouter(prefix="/audio", tags=["Audio"])
logger = logging.getLogger(__name__)


# ============================================================
# SPEECH TO TEXT (Whisper)
# ============================================================

@router.post("/speech-to-text", response_model=STTResponse)
async def speech_to_text_endpoint(file: UploadFile = File(...)):
    """
    Convert speech to text using openai/whisper-large-v3.

    - Upload an audio file (wav, mp3, m4a…)
    - Returns transcribed English text
    """
    try:
        audio_bytes = await file.read()

        result = transcribe_audio(audio_bytes)

        response_data = STTResponse(
            text=result,
            model_name="openai/whisper-large-v3",
            language="en",
            duration_seconds=None  # optional filler
        )

        logger.info(f"STT completed: {response_data.text[:40]}...")
        return response_data

    except Exception as e:
        logger.error(f"STT error: {str(e)}")
        raise HTTPException(status_code=500, detail=f"Speech-to-text failed: {str(e)}")


# ============================================================
# TEXT TO SPEECH (Bark)
# ============================================================

@router.post("/text-to-speech", response_model=TTSResponse)
async def text_to_speech_endpoint(request: TTSRequest):
    """
    Convert text to synthesized speech using Bark.
    Returns streamed audio.
    """
    try:
        audio_bytes = text_to_speech(request.text)

        metadata = TTSResponse(
            message="Audio generated successfully",
            audio_format="wav",
            length_seconds=None,
            model_name="suno/bark"
        )

        logger.info(f"TTS generated for text: {request.text[:40]}...")

        return StreamingResponse(
            iter([audio_bytes]),
            media_type="audio/wav",
            headers={
                "X-Audio-Metadata": metadata.model_dump_json()
            }
        )

    except Exception as e:
        logger.error(f"TTS error: {str(e)}")
        raise HTTPException(status_code=500, detail=f"Text-to-speech failed: {str(e)}")