Replaces a static SPA that shipped an Airtable PAT in the JS bundle.
The new architecture holds all secrets server-side, fronts the app
behind Apache on optical-dev with the shared-vhost split-build pattern,
and is designed for a later Azure AD/MSAL swap-in.
- backend/ FastAPI + uvicorn, local auth (Azure AD stub), Airtable
proxy with TTL cache, Zoho .xlsx/.csv parser, merge
service for utilisation summaries. 28 pytest tests.
- frontend/ React + Vite + TS + Tailwind + Recharts SPA. Login entry
chunk 12.83 KB gzipped; Recharts lazy-loaded. No tokens
or Airtable URLs in the built bundle.
- deploy/ Idempotent deploy.sh (port auto-pick 8200-8299,
.env-persisted) + split-build Apache include template.
- docker-compose.yml pins name: utilisation-dept and binds 127.0.0.1.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
58 lines
1.7 KiB
Python
58 lines
1.7 KiB
Python
"""TTL cache with per-key async locks.
|
|
|
|
Decisions:
|
|
- cachetools.TTLCache handles the time-based eviction; we wrap it with
|
|
per-key asyncio.Lock instances so that the first concurrent caller does
|
|
the load and the rest wait. Without this we'd stampede Airtable on
|
|
cache expiry.
|
|
- A single asyncio.Lock guards lock-dict mutations.
|
|
"""
|
|
|
|
from __future__ import annotations
|
|
|
|
import asyncio
|
|
from typing import Awaitable, Callable, Hashable, TypeVar
|
|
|
|
from cachetools import TTLCache
|
|
|
|
|
|
T = TypeVar("T")
|
|
|
|
|
|
class TTLAsyncCache:
|
|
def __init__(self, ttl: int, maxsize: int = 32) -> None:
|
|
self._cache: TTLCache = TTLCache(maxsize=maxsize, ttl=ttl)
|
|
self._locks: dict[Hashable, asyncio.Lock] = {}
|
|
self._locks_guard = asyncio.Lock()
|
|
|
|
async def _get_lock(self, key: Hashable) -> asyncio.Lock:
|
|
async with self._locks_guard:
|
|
lock = self._locks.get(key)
|
|
if lock is None:
|
|
lock = asyncio.Lock()
|
|
self._locks[key] = lock
|
|
return lock
|
|
|
|
def peek(self, key: Hashable) -> T | None:
|
|
return self._cache.get(key)
|
|
|
|
def invalidate(self, key: Hashable) -> None:
|
|
self._cache.pop(key, None)
|
|
|
|
async def get_or_set(
|
|
self,
|
|
key: Hashable,
|
|
loader: Callable[[], Awaitable[T]],
|
|
) -> T:
|
|
# Fast path — no lock needed if value present.
|
|
if key in self._cache:
|
|
return self._cache[key]
|
|
|
|
lock = await self._get_lock(key)
|
|
async with lock:
|
|
# Re-check inside the lock; another coroutine may have loaded it.
|
|
if key in self._cache:
|
|
return self._cache[key]
|
|
value = await loader()
|
|
self._cache[key] = value
|
|
return value
|