Use as cache backend
anystore can be used as a simple cache backend and provides some useful helpers for caching parts of an application. For performance reasons, the storage backend should be a redis instance or a sql database, but of course any other (e.g. file-like) backend could be configured.
All high level functions described in the basic usage of course work here as well.
Set up
Assuming a redis server is running at localhost, anystore could be configured via environment vars (see configuration) or store initialization at runtime.
export ANYSTORE_URI=redis://localhost
# takes any valid redis connection uri string:
export ANYSTORE_URI=redis://user:password@host:port
A useful example configuration for a 1 hour redis cache (assuming all data is json serializable) could look like this:
from anystore import get_store
store = get_store(
uri="redis://localhost",
serialization_mode="json",
raise_on_nonexist=False, # just return `None` for missing data
default_ttl=3600,
backend_config={
"redis_prefix": "my-cache-001"
}
)
All these options could be configured with these environment variables:
ANYSTORE_URI=redis://localhost
ANYSTORE_SERIALIZATION_MODE=json
ANYSTORE_RAISE_ON_NONEXIST=0
ANYSTORE_DEFAULT_TTL=3600
ANYSTORE_BACKEND_CONFIG__REDIS_PREFIX=my-cache-001
Use the cache
Simply set up the store and store and retrieve data from it:
from anystore import get_store
# if configured via env var
store = get_store()
# or configure during runtime
store = get_store("redis://localhost")
# store data valid for 1 hour
store.put("cache_key", data, ttl=3600)
# retrieve and remove from cache
res = store.pop("cache_key")
anycache decorator
Decorate a function with it to retrieve the cached result on next call. By default, the cache key is computed from the input arguments (see below for configuration).
from anystore import anycache
@anycache()
def calculate(data, ttl=3600):
# a very time consuming task
return result
# 1. time: slow
res = calculate(100)
# 2. time: fast, as now cached
res = calculate(100)
decorator options
Most notably the store to use for the decorator and a key_func to compute the cache key based on the input.
from anystore import get_store, anycache
store = get_store("s3://slow-cache")
# transform the input `name` to uppercase for the cache key
@anycache(store=store, key_func: lambda x: x.upper())
def get_result(name):
# do something
return result
For all the parameters, see reference