Redis

Redis Skill

Bookkeep uses aiocache with Redis 7.x for distributed caching, automatically falling back to in-memory cache when Redis is unavailable. The cache layer is defined in backend/app/cache.py and provides async get/set/delete operations with pattern-based invalidation.

Quick Start

Basic Cache Operations

from app.cache import get_cached, set_cached, delete_cached, make_cache_key, CACHE_TTL

# Get with cache
cache_key = make_cache_key("book_details", book_id=123)
cached = await get_cached(cache_key)
if cached is not None:
    return cached

# Fetch and cache
result = await fetch_book(book_id)
await set_cached(cache_key, result, ttl=CACHE_TTL["book_details"])
return result

Cache Invalidation on Mutation

from app.cache import delete_cached, clear_cache_pattern, make_cache_key

# Single key invalidation
await delete_cached(make_cache_key("requests_by_hardcover", hardcover_id=book.hardcover_id))

# Pattern-based invalidation (clears all batch caches)
await clear_cache_pattern("requests_by_hardcover_batch:*")

Key Concepts

Common Patterns

Cache-Aside Pattern (Standard)

When: Every read operation that hits external API or expensive DB query

cache_key = cache.make_cache_key("trending", limit=limit, date=current_date)
cached_result = await cache.get_cached(cache_key)
if cached_result is not None:
    return cached_result

# Expensive operation
result = await hardcover_api.fetch_trending(limit)

await cache.set_cached(cache_key, result, ttl=cache.CACHE_TTL["trending"])
return result

Bypass Cache (Force Refresh)

When: User explicitly requests fresh data

cache_key = cache.make_cache_key("book_details", book_id=book_id)
cached_result = None if bypass_cache else await cache.get_cached(cache_key)

TTL Reference

See Also

• patterns
• workflows

Related Skills

• See the fastapi skill for router integration
• See the python skill for async patterns