Disk Cache -- Disk and file backed persistent cache.
DiskCache
_ is an Apache2 licensed disk and file backed cache library, written
in pure-Python, and compatible with Django.
The cloud-based computing of 2023 puts a premium on memory. Gigabytes of empty space is left on disks as processes vie for memory. Among these processes is Memcached (and sometimes Redis) which is used as a cache. Wouldn't it be nice to leverage empty disk space for caching?
Django is Python's most popular web framework and ships with several caching backends. Unfortunately the file-based cache in Django is essentially broken. The culling method is random and large caches repeatedly scan a cache directory which slows linearly with growth. Can you really allow it to take sixty milliseconds to store a key in a cache with a thousand items?
In Python, we can do better. And we can do it in pure-Python!
::
In [1]: import pylibmc In [2]: client = pylibmc.Client(['127.0.0.1'], binary=True) In [3]: client[b'key'] = b'value' In [4]: %timeit client[b'key']
10000 loops, best of 3: 25.4 µs per loop
In [5]: import diskcache as dc In [6]: cache = dc.Cache('tmp') In [7]: cache[b'key'] = b'value' In [8]: %timeit cache[b'key']
100000 loops, best of 3: 11.8 µs per loop
Note: Micro-benchmarks have their place but are not a substitute for real measurements. DiskCache offers cache benchmarks to defend its performance claims. Micro-optimizations are avoided but your mileage may vary.
DiskCache efficiently makes gigabytes of storage space available for caching. By leveraging rock-solid database libraries and memory-mapped files, cache performance can match and exceed industry-standard solutions. There's no need for a C compiler or running another process. Performance is a feature and testing has 100% coverage with unit tests and hours of stress.
Daren Hasenkamp
_, Founder --
"It's a useful, simple API, just like I love about Redis. It has reduced
the amount of queries hitting my Elasticsearch cluster by over 25% for a
website that gets over a million users/day (100+ hits/second)."
Mathias Petermann
_, Senior Linux System Engineer --
"I implemented it into a wrapper for our Ansible lookup modules and we were
able to speed up some Ansible runs by almost 3 times. DiskCache is saving
us a ton of time."
Does your company or website use DiskCache
? Send us a message <contact@grantjenks.com>
and let us know.
.. _Daren Hasenkamp
: https://www.linkedin.com/in/daren-hasenkamp-93006438/
.. _Mathias Petermann
: https://www.linkedin.com/in/mathias-petermann-a8aa273b/
.. image:: https://github.com/grantjenks/python-diskcache/workflows/integration/badge.svg :target: https://github.com/grantjenks/python-diskcache/actions?query=workflow%3Aintegration
.. image:: https://github.com/grantjenks/python-diskcache/workflows/release/badge.svg :target: https://github.com/grantjenks/python-diskcache/actions?query=workflow%3Arelease
Installing DiskCache
_ is simple with pip <http://www.pip-installer.org/>
_::
$ pip install diskcache
You can access documentation in the interpreter with Python's built-in help function::
import diskcache help(diskcache) # doctest: +SKIP
The core of DiskCache
_ is three data types intended for caching. Cache
_
objects manage a SQLite database and filesystem directory to store key and
value pairs. FanoutCache
_ provides a sharding layer to utilize multiple
caches and DjangoCache
_ integrates that with Django
_::
from diskcache import Cache, FanoutCache, DjangoCache help(Cache) # doctest: +SKIP help(FanoutCache) # doctest: +SKIP help(DjangoCache) # doctest: +SKIP
Built atop the caching data types, are Deque
_ and Index
_ which work as a
cross-process, persistent replacements for Python's collections.deque
and
dict
. These implement the sequence and mapping container base classes::
from diskcache import Deque, Index help(Deque) # doctest: +SKIP help(Index) # doctest: +SKIP
Finally, a number of recipes
_ for cross-process synchronization are provided
using an underlying cache. Features like memoization with cache stampede
prevention, cross-process locking, and cross-process throttling are available::
from diskcache import memoize_stampede, Lock, throttle help(memoize_stampede) # doctest: +SKIP help(Lock) # doctest: +SKIP help(throttle) # doctest: +SKIP
Python's docstrings are a quick way to get started but not intended as a
replacement for the DiskCache Tutorial
_ and DiskCache API Reference
_.
.. _Cache
: http://www.grantjenks.com/docs/diskcache/tutorial.html#cache
.. _FanoutCache
: http://www.grantjenks.com/docs/diskcache/tutorial.html#fanoutcache
.. _DjangoCache
: http://www.grantjenks.com/docs/diskcache/tutorial.html#djangocache
.. _Django
: https://www.djangoproject.com/
.. _Deque
: http://www.grantjenks.com/docs/diskcache/tutorial.html#deque
.. _Index
: http://www.grantjenks.com/docs/diskcache/tutorial.html#index
.. _recipes
: http://www.grantjenks.com/docs/diskcache/tutorial.html#recipes
For those wanting more details, this part of the documentation describes tutorial, benchmarks, API, and development.
DiskCache Tutorial
_DiskCache Cache Benchmarks
_DiskCache DjangoCache Benchmarks
_Case Study: Web Crawler
_Case Study: Landing Page Caching
_Talk: All Things Cached - SF Python 2017 Meetup
_DiskCache API Reference
_DiskCache Development
_.. _DiskCache Tutorial
: http://www.grantjenks.com/docs/diskcache/tutorial.html
.. _DiskCache Cache Benchmarks
: http://www.grantjenks.com/docs/diskcache/cache-benchmarks.html
.. _DiskCache DjangoCache Benchmarks
: http://www.grantjenks.com/docs/diskcache/djangocache-benchmarks.html
.. _Talk: All Things Cached - SF Python 2017 Meetup
: http://www.grantjenks.com/docs/diskcache/sf-python-2017-meetup-talk.html
.. _Case Study: Web Crawler
: http://www.grantjenks.com/docs/diskcache/case-study-web-crawler.html
.. _Case Study: Landing Page Caching
: http://www.grantjenks.com/docs/diskcache/case-study-landing-page-caching.html
.. _DiskCache API Reference
: http://www.grantjenks.com/docs/diskcache/api.html
.. _DiskCache Development
: http://www.grantjenks.com/docs/diskcache/development.html
Comparisons to popular projects related to DiskCache
_.
Key-Value Stores ................
DiskCache
_ is mostly a simple key-value store. Feature comparisons with four
other projects are shown in the tables below.
dbm
_ is part of Python's standard library and implements a generic
interface to variants of the DBM database — dbm.gnu or dbm.ndbm. If none of
these modules is installed, the slow-but-simple dbm.dumb is used.shelve
_ is part of Python's standard library and implements a “shelf” as a
persistent, dictionary-like object. The difference with “dbm” databases is
that the values can be anything that the pickle module can handle.sqlitedict
_ is a lightweight wrapper around Python's sqlite3 database with
a simple, Pythonic dict-like interface and support for multi-thread
access. Keys are arbitrary strings, values arbitrary pickle-able objects.pickleDB
_ is a lightweight and simple key-value store. It is built upon
Python's simplejson module and was inspired by Redis. It is licensed with the
BSD three-clause license... _dbm
: https://docs.python.org/3/library/dbm.html
.. _shelve
: https://docs.python.org/3/library/shelve.html
.. _sqlitedict
: https://github.com/RaRe-Technologies/sqlitedict
.. _pickleDB
: https://pythonhosted.org/pickleDB/
Features
================ ============= ========= ========= ============ ============ Feature diskcache dbm shelve sqlitedict pickleDB ================ ============= ========= ========= ============ ============ Atomic? Always Maybe Maybe Maybe No Persistent? Yes Yes Yes Yes Yes Thread-safe? Yes No No Yes No Process-safe? Yes No No Maybe No Backend? SQLite DBM DBM SQLite File Serialization? Customizable None Pickle Customizable JSON Data Types? Mapping/Deque Mapping Mapping Mapping Mapping Ordering? Insert/Sorted None None None None Eviction? LRU/LFU/more None None None None Vacuum? Automatic Maybe Maybe Manual Automatic Transactions? Yes No No Maybe No Multiprocessing? Yes No No No No Forkable? Yes No No No No Metadata? Yes No No No No ================ ============= ========= ========= ============ ============
Quality
================ ============= ========= ========= ============ ============ Project diskcache dbm shelve sqlitedict pickleDB ================ ============= ========= ========= ============ ============ Tests? Yes Yes Yes Yes Yes Coverage? Yes Yes Yes Yes No Stress? Yes No No No No CI Tests? Linux/Windows Yes Yes Linux No Python? 2/3/PyPy All All 2/3 2/3 License? Apache2 Python Python Apache2 3-Clause BSD Docs? Extensive Summary Summary Readme Summary Benchmarks? Yes No No No No Sources? GitHub GitHub GitHub GitHub GitHub Pure-Python? Yes Yes Yes Yes Yes Server? No No No No No Integrations? Django None None None None ================ ============= ========= ========= ============ ============
Timings
These are rough measurements. See DiskCache Cache Benchmarks
_ for more
rigorous data.
================ ============= ========= ========= ============ ============ Project diskcache dbm shelve sqlitedict pickleDB ================ ============= ========= ========= ============ ============ get 25 µs 36 µs 41 µs 513 µs 92 µs set 198 µs 900 µs 928 µs 697 µs 1,020 µs delete 248 µs 740 µs 702 µs 1,717 µs 1,020 µs ================ ============= ========= ========= ============ ============
Caching Libraries .................
joblib.Memory
_ provides caching functions and works by explicitly saving
the inputs and outputs to files. It is designed to work with non-hashable and
potentially large input and output data types such as numpy arrays.klepto
_ extends Python’s lru_cache
to utilize different keymaps and
alternate caching algorithms, such as lfu_cache
and mru_cache
. Klepto
uses a simple dictionary-sytle interface for all caches and archives... _klepto
: https://pypi.org/project/klepto/
.. _joblib.Memory
: https://joblib.readthedocs.io/en/latest/memory.html
Data Structures ...............
dict
_ is a mapping object that maps hashable keys to arbitrary
values. Mappings are mutable objects. There is currently only one standard
Python mapping type, the dictionary.pandas
_ is a Python package providing fast, flexible, and expressive data
structures designed to make working with “relational” or “labeled” data both
easy and intuitive.Sorted Containers
_ is an Apache2 licensed sorted collections library,
written in pure-Python, and fast as C-extensions. Sorted Containers
implements sorted list, sorted dictionary, and sorted set data types... _dict
: https://docs.python.org/3/library/stdtypes.html#typesmapping
.. _pandas
: https://pandas.pydata.org/
.. _Sorted Containers
: http://www.grantjenks.com/docs/sortedcontainers/
Pure-Python Databases .....................
ZODB
_ supports an isomorphic interface for database operations which means
there's little impact on your code to make objects persistent and there's no
database mapper that partially hides the datbase.CodernityDB
_ is an open source, pure-Python, multi-platform, schema-less,
NoSQL database and includes an HTTP server version, and a Python client
library that aims to be 100% compatible with the embedded version.TinyDB
_ is a tiny, document oriented database optimized for your
happiness. If you need a simple database with a clean API that just works
without lots of configuration, TinyDB might be the right choice for you... _ZODB
: http://www.zodb.org/
.. _CodernityDB
: https://pypi.org/project/CodernityDB/
.. _TinyDB
: https://tinydb.readthedocs.io/
Object Relational Mappings (ORM) ................................
Django ORM
_ provides models that are the single, definitive source of
information about data and contains the essential fields and behaviors of the
stored data. Generally, each model maps to a single SQL database table.SQLAlchemy
_ is the Python SQL toolkit and Object Relational Mapper that
gives application developers the full power and flexibility of SQL. It
provides a full suite of well known enterprise-level persistence patterns.Peewee
_ is a simple and small ORM. It has few (but expressive) concepts,
making it easy to learn and intuitive to use. Peewee supports Sqlite, MySQL,
and PostgreSQL with tons of extensions.SQLObject
_ is a popular Object Relational Manager for providing an object
interface to your database, with tables as classes, rows as instances, and
columns as attributes.Pony ORM
_ is a Python ORM with beautiful query syntax. Use Python syntax
for interacting with the database. Pony translates such queries into SQL and
executes them in the database in the most efficient way... _Django ORM
: https://docs.djangoproject.com/en/dev/topics/db/
.. _SQLAlchemy
: https://www.sqlalchemy.org/
.. _Peewee
: http://docs.peewee-orm.com/
.. _SQLObject
: http://sqlobject.org/
.. _Pony ORM
: https://ponyorm.com/
SQL Databases .............
SQLite
_ is part of Python's standard library and provides a lightweight
disk-based database that doesn’t require a separate server process and allows
accessing the database using a nonstandard variant of the SQL query language.MySQL
_ is one of the world’s most popular open source databases and has
become a leading database choice for web-based applications. MySQL includes a
standardized database driver for Python platforms and development.PostgreSQL
_ is a powerful, open source object-relational database system
with over 30 years of active development. Psycopg is the most popular
PostgreSQL adapter for the Python programming language.Oracle DB
_ is a relational database management system (RDBMS) from the
Oracle Corporation. Originally developed in 1977, Oracle DB is one of the
most trusted and widely used enterprise relational database engines.Microsoft SQL Server
_ is a relational database management system developed
by Microsoft. As a database server, it stores and retrieves data as requested
by other software applications... _SQLite
: https://docs.python.org/3/library/sqlite3.html
.. _MySQL
: https://dev.mysql.com/downloads/connector/python/
.. _PostgreSQL
: http://initd.org/psycopg/
.. _Oracle DB
: https://pypi.org/project/cx_Oracle/
.. _Microsoft SQL Server
: https://pypi.org/project/pyodbc/
Other Databases ...............
Memcached
_ is free and open source, high-performance, distributed memory
object caching system, generic in nature, but intended for use in speeding up
dynamic web applications by alleviating database load.Redis
_ is an open source, in-memory data structure store, used as a
database, cache and message broker. It supports data structures such as
strings, hashes, lists, sets, sorted sets with range queries, and more.MongoDB
_ is a cross-platform document-oriented database program. Classified
as a NoSQL database program, MongoDB uses JSON-like documents with
schema. PyMongo is the recommended way to work with MongoDB from Python.LMDB
_ is a lightning-fast, memory-mapped database. With memory-mapped
files, it has the read performance of a pure in-memory database while
retaining the persistence of standard disk-based databases.BerkeleyDB
_ is a software library intended to provide a high-performance
embedded database for key/value data. Berkeley DB is a programmatic toolkit
that provides built-in database support for desktop and server applications.LevelDB
_ is a fast key-value storage library written at Google that
provides an ordered mapping from string keys to string values. Data is stored
sorted by key and users can provide a custom comparison function... _Memcached
: https://pypi.org/project/python-memcached/
.. _MongoDB
: https://api.mongodb.com/python/current/
.. _Redis
: https://redis.io/clients#python
.. _LMDB
: https://lmdb.readthedocs.io/
.. _BerkeleyDB
: https://pypi.org/project/bsddb3/
.. _LevelDB
: https://plyvel.readthedocs.io/
DiskCache Documentation
_DiskCache at PyPI
_DiskCache at GitHub
_DiskCache Issue Tracker
_.. _DiskCache Documentation
: http://www.grantjenks.com/docs/diskcache/
.. _DiskCache at PyPI
: https://pypi.python.org/pypi/diskcache/
.. _DiskCache at GitHub
: https://github.com/grantjenks/python-diskcache/
.. _DiskCache Issue Tracker
: https://github.com/grantjenks/python-diskcache/issues/
Copyright 2016-2023 Grant Jenks
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License.
.. _DiskCache
: http://www.grantjenks.com/docs/diskcache/