jowoojun / ring

Shift cache paradigm to code and forget about storages. With built-in memcache & redis + asyncio support.

Ring

Let's concentrate on code, not about storages.

Ring shows a way to control cache in point of view of code - not about storages. Ring's decorator is convenient but also keeping fluency for general scenarios.

:mod:`asyncio` support!

Take an explicit but fully automated cache. Ring decorators convert your functions to cached version of them, with extra control methods.

Function cache

import ring
import memcache
import requests

mc = memcache.Client(['127.0.0.1:11211'])

# working for mc, expire in 60sec
@ring.func.memcache(mc, time=60)
def get_url(url):
    return requests.get(url).content

# normal way - it is cached
data = get_url('http://example.com')

It is a normal smart cache flow.

But ring is different when you want to explicitly control it.

# delete the cache
get_url.delete('http://example.com')
# get cached data or None
data_or_none = get_url.get('http://example.com')

# get internal cache key
key = get_url.key('http://example.com')
# and access directly to the backend
direct_data = mc.get(key)

Method cache

import ring
import redis

rc = redis.StrictRedis()

class User(dict):
    def __ring_key__(self):
        return self['id']

    # working for rc, no expiration
    # using json coder for non-bytes cache data
    @ring.func.redis(rc, coder='json')
    def data(self):
        return self.copy()

    # parameters are also ok!
    @ring.func.redis(rc, coder='json')
    def child(self, child_id):
        return {'user_id': self['id'], 'child_id': child_id}

user = User(id=42, name='Ring')

# create and get cache
user_data = user.data()  # cached
user['name'] = 'Ding'
# still cached
cached_data = user.data()
assert user_data == cached_data
# refresh
updated_data = user.data.update()
assert user_data != updated_data

# id is the cache key so...
user2 = User(id=42)
# still hitting the same cache
assert updated_data == user2.data()

Installation

PyPI is the recommended way.

$ pip install ring

To browse versions and tarballs, visit:

https://pypi.python.org/pypi/ring/

To use memcache or redis, don't forget to install related libraries. For example: python-memcached, python3-memcached, pylibmc, redis-py etc

It may require to install and run related services too: memcached and redis.

Contribution

First, install ring in editable mode. Use 'tests' extra to install tests requirements.

$ pip install -e '.[tests]'

Run pytest to check the test set is ready.

$ pytest -vv

Note: Can't you install it because of compile errors? Don't forget to install and run memcached and redis locally.

For macOS:

$ brew install libmemcached memcached redis
$ brew services start memcached
$ brew services start redis

For debian/ubuntu:

$ apt install libmemcached-dev memcached redis-server
$ service memcached start
$ service redis start