"""Entry point for a Hugging Face Space with the Gradio SDK (free, unlike Docker).

A Gradio Space runs `python app.py` and expects a server on port 7860: this serves the demo API
(server.py) there, with a minimal Gradio page mounted at /gradio.

On the free tier a Gradio Space gets ZeroGPU hardware. The GPU is deliberately not used: it is
metered per visitor (a couple of minutes a day) and queued, which does not suit a score
recomputed on every keystroke, and Laya answers in a fraction of a second on CPU anyway. So
inference is pinned to the CPU.
"""
import os

# ZeroGPU refuses to start a Space that declares no @spaces.GPU function, so one is declared but
# never called. `spaces` must be imported before anything that imports torch (it patches CUDA
# initialisation); outside Hugging Face the package is absent or has no effect.
try:
    import spaces

    @spaces.GPU
    def _zerogpu_placeholder():
        """Never called: only satisfies ZeroGPU's startup check."""
except Exception as error:  # absent locally; any other failure shows in the Space logs
    print(f"spaces unavailable ({type(error).__name__}: {error})", flush=True)

# One model on CPU keeps memory and latency reasonable; override in the Space settings.
os.environ.setdefault("LAYA_BACKEND", "torch")
os.environ.setdefault("LAYA_MODELS", "typed")
os.environ.setdefault("LAYA_DEVICE", "cpu")
# About 1 s per request on the Space's CPU: wait for a pause in typing before calling the model.
os.environ.setdefault("LAYA_DEBOUNCE", "700")

import gradio as gr  # noqa: E402
import uvicorn  # noqa: E402

from server import MODELS, app  # noqa: E402

with gr.Blocks(title="Laya Demo API") as info:
    gr.Markdown(
        "## Laya Demo API\n"
        f"Modèles chargés : {', '.join(MODELS.values())}.\n\n"
        "La démo est sur `/`, l'état de l'API sur `/health`."
    )

# Spaces set GRADIO_SSR_MODE, which starts a Node server on port 7860 and leaves uvicorn unable
# to bind it ("address already in use"); this page needs no server-side rendering.
app = gr.mount_gradio_app(app, info, path="/gradio", ssr_mode=False)



def report_zerogpu_startup():
    """Send ZeroGPU the startup report that `spaces` normally sends from `gr.Blocks.launch()`.

    This app is served by uvicorn, so `launch()` never runs and ZeroGPU stops the Space
    ("No @spaces.GPU function detected during startup"). `spaces.zero.startup` only exists on
    ZeroGPU hardware; elsewhere this does nothing.
    """
    try:
        from spaces import zero
    except Exception:
        return
    if hasattr(zero, "startup"):
        zero.startup()


if __name__ == "__main__":
    report_zerogpu_startup()
    uvicorn.run(app, host="0.0.0.0", port=7860)