Cache

Cache application results that can be rebuilt, such as catalog summaries, lookup data, or expensive calculations. Keep the permanent data in your database; decide how long a cached result may be stale and when to invalidate it.

Use the configured cache in your application

Start with the file store included in a normal installation. Set these values in your application’s .env before bootstrap:

CACHE_STORE=file
CACHE_PREFIX=my-app

The default directory is storage/framework/cache/data and must be writable by the application process. You do not need Redis or a separate cache package for this store. config/cache.py defines the configured stores and their options.

Call the Cache facade from an async controller, command, job, or service after the application has booted:

from orionis.support.facades import Cache


async def remember_preferences(user_id: int, preferences: dict) -> dict:
    key = f"user:{user_id}:preferences:v1"
    await Cache.set(key, preferences, ttl=300)
    return await Cache.get(key)


async def invalidate_preferences(user_id: int) -> None:
    await Cache.delete(f"user:{user_id}:preferences:v1")

The first function stores a JSON-compatible dictionary for 300 seconds and immediately returns it. After expiry or invalidation, get() returns None. Authorize the user and persist preference changes before caching them; this example only demonstrates cache operations.

For cache-aside, use await Cache.remember(key, 300, resolver): the zero-argument resolver runs only on a miss and may be async. Concurrent misses can still invoke it more than once. Use the regeneration guidance when duplicate computation matters.

Choose the next step

1. Store and regenerate values: keys, TTL in seconds, cache-aside, batches, and counters.
2. Configure stores and locks: choose memory, file, Redis, Memcached, or database according to sharing requirements.
3. Cache source-derived artifacts: the separate synchronous FileBasedCache API for dictionaries compiled from source files.

FileBasedCache is for compiled artifacts; use the Cache facade for everyday application values. The two APIs have different serialization and invalidation rules.