This is the base structure or the base class for the other data structure classes. It takes two arguments with the second being optional.
Signature: BaseStructure(redis, token = None)
from redistr import BaseStructure
from redis import Redis
import pickle
# if no token provided, a random 16 bytes will be generated
base = BaseStructure(Redis())
# generated or provided token can be accessed using `token` property
token = base.token
# the token will be used for data sharing between processes
base_other = BaseStructure(Redis(), token)
# the type for the redis data structure can be accessed
data_type = base_other.type # a `bytes` object
# the serializer can be acquired at `serialize` property
ser = base.serialize
# and can be changed using the same property
base.serialize = pickle
# `delete` method for deleting key from redis
base.delete() # True|False return
base.clear() # alias for `delete`
base.flush() # alias for `delete`
# a private static method for converting keys
conv_key = base._convert(1) # b'1'
conv_key = base._convert('test') # b'test'
# a private method for initializing
# it will check the redis data type for the key
# if type doesn't match it will delete it
base._initialize()
# two private methods for accessing serializer
base._loads(base._dumps({'test': 'case'})) # {'test': 'case'}
The dict <-> redis hash interface class, compatible with python dict
, refer to Python Docs: Dictionary.
Remember that the Dict
instance do not do dirty check, so modifications on mutable data should be done the same way when operating on a shelve
database.
from redistr import Dict
from redis import Redis
rem_dict = Dict(Redis(), 'dict_key')
rem_dict['test'] = 'case'
# length property for getting the length
length = rem_dict.length # 1
# popitem pops the key in sequence instead of randomly
rem_dict.popitem() # ('test', 'case')
rem_dict.update(test = 'case')
# additional feature, dynamic attributes
# only works if no methods or properties have the same name
value = rem_dict.test # 'case'
rem_dict.test = {'test': 'case'}
rem_dict.case = value
rem_dict.test # {'test': 'case'}
rem_dict.case # 'case;
# how to modify mutable data, refer to `shelve` docs
rem_dict['doc'] = {'test': 'case'}
value = rem_dict['doc']
value['case'] = 'test'
rem_dict['doc'] = value
rem_dict['doc'] # {'test': 'case', 'case': 'test'}
The redis HyperLogLog data structure, used to estimate how many items are in a collection using a very small memory footprint. Refer to Redis Docs: HyperLogLog and Redis author's blog on the subject for more information.
from redistr import HyperLogLog
from redis import Redis
hll = HyperLogLog(Redis(), 'hll_key')
# register an action to the structure
hll.register({'test': 'case'})
# plus sign overridden as alias for register method
hll + b'another action or item'
100 + hll
# get the estimate count for unique items
count = hll.cardinal() # 3
count = hll.count() # 3, alias for `cardinal` method
# `log` property for quicker action
hll.log = {'another': 'item'}
hll.log = ['yet', 'another', 'action']
count = hll.log # 5, get the unique item count
The interface for list
and set
are called List
and Set
. They behave the same as python list
and set
, refer to Python Docs: List and Python Docs: Set for more information.
Both structure provides _content
property for accessing raw content, and content
property for accessing parsed content.
from redistr import List
from redis import Redis
rem_list = List(Redis())
rem_set = List(Redis())
rem_list.append('item')
rem_list.extend('item')
# access the content
rem_list.content # ['item', 'i', 't', 'e', 'm']
# access the length
rem_list.length # 5
# circulate one item to the same/different list
# using RPOPLPUSH operation on redis structure
# Signature: rem_list.circulate(token = None)
value = rem_list.circulate() # 'm'
rem_list.content # ['m', 'item', 'i', 't', 'e']
value = rem_list.circulate('another_list_key')
rem_list.content # ['m', 'item', 'i', 't']
# get `bytes`, the serialized representation for 'e'
Redis().rpop('another_list_key') # `bytes` object
## additional features, methods and properties
# prepend an item to the left of the list
rem_list.prepend('item')
rem_list.unshift('item')
# popleft method and aliases for getting an item left
rem_list.popleft() # 'item'
rem_list.shift() # 'item'
from redistr import Set
from redis import Redis
rem_set = Set(Redis())
# behaves exactly like python `set`
rem_set.add(1)
rem_set.content # {1}
rem_set.length # 1
rem_set.add(2)
rem_set_1 = Set(Redis())
rem_set_1.add(2)
rem_set_1.add(3)
rem_set.union(rem_set_1) # {1, 2, 3}
rem_set | rem_set_1 # {1, 2, 3}
rem_set.difference({2,3,4,5}) # {1}
rem_set - {2,3,4,5} # {1}
# etc...
Queue
class is a subclass of List
, thus it has all the methods available to List
ready to be used. And since it is built on List
it can actually share the same redis key with a List
instance. It provides methods for both blocking and non-blocking use.
from redistr import Queue
from redis import Redis
queue = Queue(Redis, 'a_list_key')
# always hungery for more data
queue.full() # False, always False
# empty method for checking if empty
queue.empty() # True
# put an item to the left
queue.put('item')
queue.set('item2') # alias for put
queue.send('item 3') # alias for put
# put an item to the right
# have queue.set_right(), queue.send_right() aliases
queue.put_right('it 0')
# share the same methods with `List` class
queue.content # ['item 3', 'item2', 'item', 'it 0']
queue.length # 4
queue.pop() # 'it 0'
# blocking operations
# Signature: `get(block = True, timeout = 0)`
# Alias: `recv(block = True, timeout = 0)`
# Signature: `get_left(block = True, timeout = 0)`
# Alias: `recv_left(block = True, timeout = 0)`
# Signature: `circulate(token = None, block = True, timeout = 0)
queue.push('right')
queue.get() # 'right'
queue.get_left() # 'item 3'
queue.circulate() # 'item'
queue.content # ['item', 'item2']
# additional feature: `msg` property for quick accessing
queue.msg # 'item2'
queue.msg = 2000
queue.content # [2000, 'item']
queue.msg # 'item'
queue.msg = 1000
queue.content = [1000, 2000]
The SeCo
class is an independent class for serializing and compressing data, refer to its GitHub Repo for more information.
Signature: SeCo(serialize = None, compress = None, **kwards)
It can be initialized to use any combinations between ('json', 'msgpack', 'pickle') and ('zlib', 'bz2').
Recommend to use 'msgpack' and 'zlib' for the optimal speed and space efficiency; use 'pickle' for broadest Python type support; use 'bz2' for maximum space efficiency at the cost of time.
Change any of the structure's serializer using the following procedures.
from redistr import Queue
from redis import Redis
from seco import SeCo
import json, msgpack, pickle
queue = Queue(Redis())
# default serializer uses `msgpack` and `zlib`
queue.serialize # get the default serializer
# create new serializers
ser_json_bz2 = SeCo('json', 'bz2')
ser_pickle_zlib = SeCo('pickle')
# flush all stale data from redis
queue.flush() # or .delete(), or .clear()
## change the serializer
# use the `Serialize` instances
queue.serialize = ser_json_bz2
queue.serialize = ser_pickle_zlib
# or others with `loads` and`dumps` methods
# use this to avoid compression, etc.
queue.serialize = json
queue.serialize = msgpack
queue.serialize = pickle