app/cache.py

"""缓存层：模型列表等热数据 TTL 缓存。"""
from __future__ import annotations

import asyncio
import time
from typing import Any, Optional

from cachetools import TTLCache

# 模型列表缓存：默认 30s TTL，最多 32 个厂商
_models_cache: "TTLCache[str, list[Any]]" = TTLCache(maxsize=32, ttl=30)


def get_cached_models(provider_id: str) -> Optional[list[Any]]:
    return _models_cache.get(provider_id)


def set_cached_models(provider_id: str, models: list[Any]) -> None:
    _models_cache[provider_id] = models


def invalidate_models_cache(provider_id: Optional[str] = None) -> None:
    if provider_id:
        _models_cache.pop(provider_id, None)
    else:
        _models_cache.clear()