dcache

Distributed Cache for Humans

• Documentation: https://dcache.readthedocs.io.

Installation

pip install dcache

How to Use

• TODO

Contributing

Contributions are welcome, feel free to open an Issue or Pull Request.

Pull requests must be for the develop branch.

git clone https://github.com/HBN3tw0rk/dcache
cd dcache
git checkout develop
python -m venv .venv
pip install -r requirements_dev.txt
pre-commit install
pytest

Pitch (Portuguese)

• Part 1: https://www.loom.com/share/ee553d6915ca4dc5ba8609d48b59bd55
• Part 2: https://www.loom.com/share/50cb4ff9560943879d78864f5fa1b4e0

What is

• distributed cache for humans
• simple API like lru_cache
• multiple backends
• easy to switch the backend
• good documentation

API

dcache

fromdcacheimportdcache@dcachedefslow_function(n):
returnn**1000

dcache vs redis

importredisredis=redis.Redis(host='localhost', port=6379, db=0)
defslow_function(n):
cached=redis.get(n)
ifcached:
returncachedvalue=n**1000redis.set(n, value)
returnvaluedefslow_function2(n):
cached=redis.get(n)
ifcached:
returncachedvalue=n**1000redis.set(n, value)
returnvalue

fromdcacheimportcachefromdcache.backendsimportRedisBackendcache=dcache(RedisBackend(host='localhost', port=6379, db=0))
@cachedefslow_function(n):
returnn**1000@cachedefslow_function2(n):
returnn**1000

real example

defprocess(id, input):
cache_path=get_content_cache_path(id, input)
ifresource.file_exist(cache_path):
returnresource.get_json(cache_path)
response=slow_function(id, input)
resource.put_json(body=response, file_path=cache_path)
returnresponse

fromdcacheimportdcachefromdcache.backendsimportS3Backend@dcache(S3Backend())defprocess(id, input):
returnslow_function(id, input)

Ideas

• integration tests using containers

multiple backends

fromdcacheimportdcachefromdcache.backendsimportInMemoryBackend, RedisBackend@dcache(multiple=[InMemoryBackend(),RedisBackend(host='localhost', port=6379, db=0),])defslow_function(n):
returnn**1000

1. search on the in-memory cache;
2. if exists, return, if not, search on Redis;
3. • if exists on Redis, save in memory and return;
4. • if not, exists on Redis, run the slow_function, save on Redis, save in-memory and return;

• doesn't run if already returned

MVP

• in memory

Roadmap

• backends: Redis, Memcached, Filesystem, database, S3, etc.
• multiple backends
• plugins