RedisWorks, the Pythonic Redis Client.

Table of Contents:


As a person who loves dynamic typing, I have always been frustrated with Py-Redis. It is a great library, but it is too low level. It maps the actual Redis commands directly.

I always wanted to just say the.path = "something" and have it magically saved into Redis. Then somewhere else say print(the.path) and get the value from Redis. As simple as that! And here it is: RedisWorks


pip install redisworks

Note that RedisWorks needs Redis server 2.4+.


let’s say if you want all the keys in Redis to start with the word root. Then you:

root = Root()  # connects to Redis on local host by default

Or if you want to be more specific:

root = Root(host='localhost', port=6379, db=0)

Saving to Redis

Saving to Redis is as simple as assigning objects to attributes of root or attributes of attributes of root (you can go as deep as you want.) Make sure you are not using any Python’s reserved words in the key’s name.


>>> from redisworks import Root
>>> import datetime
>>> root = Root()
>>> = [1, 3, 4]
>>> = [1, [2, 2]]
>>> some_date = datetime.datetime(2016, 8, 22, 10, 3, 19)
>>> root.time = some_date
>>> root.the.mapping.example = {1:1, "a": {"b": 10}}

Redis works will automatically convert your object to the proper Redis type and immediately write it to Redis as soon as you assign an element!

The respective keys for the above items will be just like what you type:, root.time, root.the.mapping.example:

If you use redis-cli, you will notice that the data is saved in the proper Redis data type:> scan 0
1) "0"
2) 1) "root.the.mapping.example"
   2) "root.time"
   3) ""> type root.the.mapping.example
hash> type root.time
string> type

Reading from Redis

Reading the data is as simple as if it was just saved in Python memory!

Redis works returns Lazy queries just like how Django returns lazy queries. In fact the lazy objects code is borrowed from Django!

If you ran the example from Saving to Redis, run a flush root.flush() to empty Redisworks Cache. This is so it goes and gets the objects from Redis instead of reading its own current copy of data:

>>> from redisworks import Root
>>> import datetime
>>> root = Root()
>>> thetime = root.time
>>> thelist =
>>> mydict = root.the.mapping.example
>>> mydict  # is not evalurated yet!
<Lazy object: root.the.mapping.example>
>>> print(mydict)
{1:1, "a": {"b": 10}}  # Now all the 3 objects are read from Redis!
>>> mydict
{1:1, "a": {"b": 10}}
[1, 3, 4]
[1, [2, 2]]
>>> root.time
2016-08-22 10:03:19

Changing root key name

Every key name by default starts with the word root. If you want to use another name, simple subclass Root:

>>> from redisworks import Root
>>> post=Post()
>>> post.item1 = "something"  # saves to Redis
>>> print(post.item1)  # loads from Redis

Numbers as attribute names

Let’s say you want root.1 as a key name. Python does not allow attribute names start with numbers.

All you need to do is start the number with the character i so Redisworks takes care of it for you:

>>> root.i1 = 10
>>> print(root.i1)

The actual key in Redis will be root.1

Use it, let me know how it goes!

I’m pretty excited about this library. I learned a lot while making it and I will be speaking about it in the SoCal Python meetup on Aug 30 2016. Check it out if you are around!



Sep Dehpour is a principal engineer at ChowNow.

Los Angeles, California, USA