"""Speech-to-text: faster-whisper on the tower-of-joy GPU.

Import of faster_whisper is deferred so the gateway can run (health checks,
protocol tests) without the heavy speech extras installed.
"""

from .config import settings

_model = None


def transcribe(pcm: bytes, sample_rate: int | None = None) -> str:
    """Transcribe raw s16le mono PCM to text."""
    global _model
    if _model is None:
        from faster_whisper import WhisperModel

        _model = WhisperModel(settings.stt_model, device=settings.stt_device)

    import numpy as np

    audio = np.frombuffer(pcm, dtype=np.int16).astype(np.float32) / 32768.0
    segments, _info = _model.transcribe(audio, language="en")
    return " ".join(segment.text.strip() for segment in segments).strip()