Caching in Python

cachetools extends the idea of functools.lru_cache with LRU, LFU and TTL caches and decorators, all in memory and per process. diskcache stores entries in SQLite and files, so the cache survives restarts and is shared between processes on one machine without running a server. dogpile.cache puts a common API over backends such as Redis and Memcached and uses a lock so that only one process regenerates an expired value while others wait or serve the old one.

Hishel adds standards-compliant HTTP caching to HTTPX and Requests clients, which cuts repeated calls to slow or rate-limited APIs. In Django, django-redis is a full-featured Redis cache backend with more options than Django's built-in Redis support, and django-cacheops caches ORM querysets and invalidates them automatically when the underlying models change.