membank

Python library for storing data in persistent memory (sqlite, postgresql, berkeley db)

goals

Provide interface to database storage that automates heavy lifting of database setup, migration, table definition, query construction.

quick intro

add items to persistent storage

import dataclasses as data # Python standard library

from membank import LoadMemory

@data.dataclass
class Dog():
    breed: str
    color: str = "black"
    weight: float = 0
    data: dict = data.field(default_factory=dict)
    picture: bytes = b''
    aliases: list = data.field(default_factory=list)

@data.dataclass
class DogWithID():
    id: str = data.field(default=None, metadata={"key": True})
    breed: str
    color: str = "black"
    weight: float = 0
    data: dict = data.field(default_factory=dict)
    picture: bytes = b''
    alive: bool = True

memory = LoadMemory() # defaults to sqlite memory
memory.put(Dog('Puli')) # stores object into database
dog = memory.get.dog() # retrieves first object found
assert dog.breed == 'Puli'
dog.color = "white"
memory.put(dog) # be carefull you store another dog
dog = memory.put(DogWithID("AB1234", "Puli"))
dog = memory.get.dogwithid(id="AB1234")
dog.color = "white"
memory.put(dog) # now you update existing dog

retrieve those after

memory = LoadMemory() # to make this work in new process, don't use sqlite memory
dog = memory.get.dog() # get first found or None
assert dog.color == 'black'
dogs = memory.get("dog") # always returns list (empty or with items)
assert len(dogs) >= 0

editing returned objects

dog = memory.get.dog()
dog.breed = 'Labdrador'
memory.put(dog) # stores edited object back

filter objects

dog = memory.get.dog(breed='Labdrador')
assert dog.breed == 'Labrador'

Known issues

• return class is stored(pickled) into underlying database table with reference to the module to load it back. If module name or class name is changed, membank won't always be able to load previous data as object pointer could be missing