DiskCache: Disk-Backed Cache for Python Without Redis or Memcached

💾 Need persistent cache in Python without spinning up Redis or Memcached? DiskCache does it in one line.

diskcache is an open-source Python library (Apache 2) that uses SQLite + memory-mapped files to deliver disk-based caching that outperforms classic in-memory solutions.

⚡ Faster than Memcached?

# Memcached
%timeit client[b'key']
# → 25.4 µs per loop

# DiskCache  
%timeit cache[b'key']
# → 11.8 µs per loop

Yes, DiskCache is faster than Memcached for local reads.

🔧 Basic usage:

import diskcache as dc

cache = dc.Cache('tmp')
cache['key'] = 'value'
print(cache['key'])  # 'value'

# With stampede-prevention memoization
from diskcache import memoize_stampede

@memoize_stampede(cache, expire=60)
def expensive(param):
    return compute_something(param)

✨ Features:

• Thread-safe and process-safe (great for multiprocessing)
• Eviction policies: LRU, LFU, and more
• Compatible with Django as a cache backend
• Persistent Deque and Dict across processes
• Cross-process Lock and throttle

💡 Explanation in a nutshell

#

DiskCache stores values in SQLite (for small keys) or on the filesystem (for large values). Since it uses memory-mapped files, reads don’t require copying data between processes. This makes it surprisingly fast for a disk cache, and fully persistent across restarts — no separate server required.

More information at the link 👇

DiskCache: Disk Backed Cache — DiskCache 5.6.1 documentation

grantjenks.com ↗