import logging
import signal
import sys
import time
from contextlib import asynccontextmanager
from typing import Optional

import torch
from fastapi import FastAPI, Request, Response
from fastapi.middleware.cors import CORSMiddleware
from fastapi.responses import JSONResponse
from prometheus_client import generate_latest, CONTENT_TYPE_LATEST
from slowapi import Limiter

from .api import router as tts_router, SPEAKERS
from .config import Settings
from .engine import load_models, get_model
from .metrics import limiter, _synthesis_latency, _audio_size_bytes, _errors_total, _request_counter

logger = logging.getLogger(__name__)

# ─── Глобальный state ─────────────────────────────────────────────

config: Optional[Settings] = None


@asynccontextmanager
async def lifespan(app: FastAPI):
    """Управление жизненным циклом приложения"""
    global config
    cfg = app.state.config

    logger.info("=" * 60)
    logger.info("🚀 Запуск Silero TTS API")
    logger.info("=" * 60)

    # Показываем конфигурацию
    logger.info("Конфигурация:")
    logger.info(f"  Модель:       {cfg.model}")
    logger.info(f"  Устройство:   {cfg.device}")
    logger.info(f"  Sample rate:  {cfg.sample_rate} Hz")
    logger.info(f"  Голоса:       {', '.join(SPEAKERS.keys())}")
    logger.info(f"  CORS:         {cfg.cors_origins}")
    logger.info(f"  Логирование:  {cfg.log_level}")
    logger.info("-" * 60)

    # Загрузка моделей
    load_models(cfg)

    logger.info("✅ ТTS API готов к работе")
    logger.info("=" * 60)

    yield  # ← сервер работает

    # Graceful shutdown
    logger.info("🛑 Остановка сервиса...")
    # Освобождаем GPU память
    if cfg.device.startswith("cuda"):
        torch.cuda.empty_cache()
    logger.info("👋 ТTS API остановлен")


# ─── Создание приложения ──────────────────────────────────────────


def create_app() -> FastAPI:
    global config

    cfg = Settings()
    config = cfg

    app = FastAPI(
        title="Silero TTS API",
        version="0.2.0",
        description="REST API синтеза русской речи на Silero v5.5",
        lifespan=lifespan,
    )

    # ── Настройка логгера ──
    logging.basicConfig(
        level=getattr(logging, cfg.log_level.upper(), logging.INFO),
        format="%(asctime)s | %(levelname)-8s | %(name)s | %(message)s",
        datefmt="%H:%M:%S",
        handlers=[logging.StreamHandler(sys.stdout)],
    )

    # ── Конфиг в app.state ──
    app.state.config = cfg

    # ── SlowAPI лимитер ──
    app.state.limiter = limiter

    # ── SlowAPI middleware ──
    from slowapi.errors import RateLimitExceeded
    from starlette.responses import JSONResponse

    @app.exception_handler(RateLimitExceeded)
    async def rate_limit_handler(request: Request, exc: RateLimitExceeded):
        return JSONResponse(
            status_code=429,
            content={
                "detail": "Превышен лимит запросов. Попробуйте позже."
            },
        )

    # ── CORS middleware ──
    app.add_middleware(
        CORSMiddleware,
        allow_origins=cfg.cors_origins,
        allow_credentials=True,
        allow_methods=["*"],
        allow_headers=["*"],
    )

    # ── Request logging middleware ──
    @app.middleware("http")
    async def log_requests(request: Request, call_next):
        start = time.perf_counter()
        response = await call_next(request)
        duration_ms = (time.perf_counter() - start) * 1000
        logger.info(
            "%s %s → %s (%.1f ms)",
            request.method,
            request.url.path,
            response.status_code,
            duration_ms,
        )

        # ── Prometheus метрика ──
        if request.url.path.startswith("/tts"):
            _request_counter.labels(
                endpoint=request.url.path,
                status=str(response.status_code),
            ).inc()

        return response

    # ── Prometheus /metrics endpoint ──
    @app.get("/metrics", tags=["monitoring"])
    async def metrics(request: Request):
        return Response(
            content=generate_latest(),
            media_type=CONTENT_TYPE_LATEST,
        )

    # ── Health check (до загрузки моделей) ──
    @app.get("/health", tags=["metadata"])
    async def health():
        try:
            get_model()
            return {"status": "healthy", "model_loaded": True}
        except Exception as e:
            return JSONResponse(
                status_code=503,
                content={"status": "unhealthy", "error": str(e)},
            )

    # ── Роуты TTS ──
    app.include_router(tts_router)

    return app


# ─── Точка входа ──────────────────────────────────────────────────


def main():
    """Запуск через uvicorn"""
    import uvicorn

    app = create_app()
    uvicorn.run(
        app,
        host="0.0.0.0",
        port=8001,
        log_level="info",
    )


if __name__ == "__main__":
    main()