Various tools for working with objects and classes in Python


Keywords
library
License
MIT
Install
pip install objecttools==1.0.1

Documentation

objecttools

Various tools for working with objects and classes in Python

Cached properties

Works just like a normal property, but returned values are cached:

from objecttools import CachedProperty

class ExpensiveOperations(object):
    @CachedProperty
    def expensive_attribute(self):
        return self.calculate()
    
    # To make it settable
    
    @expensive_attribute.setter
    def expensive_attribute(self, value):
        pass
    
    # To make it deletable
    
    @expensive_attribute.deleter
    def expensive_attribute(self):
        pass

e = ExpensiveOperations()
e.other_attribute = 1
print(e.expensive_attribute)  # Takes a long time.
print(e.expensive_attribute)  # Very quick; just retrieve from cache
v = e.expensive_attribute

e.other_attribute = 2  # expensive_attribute should be different now!
print(e.expensive_attribute)  # Old value that is wrong.
del e.expensive_attribute
print(e.expensive_attribute)  # Takes a long time, but returns new value.
e.other_attribute = 1
# Reset to known value
e.expensive_attribute = v
print(e.expensive_attribute)  # Correct value!

Singletons

from objecttools import Singleton

Sentinel = Singleton.create('Sentinel')

Sentinel() is Sentinel()  # True

d.get('missing_value', Sentinel()) is Sentinel()  # True

class GlobalState(dict, metaclass=Singleton):
    attr = 0


gs = GlobalState()
gs['value'] = 7
gs.attr = 1

print(GlobalState()['value'] + GlobalState().attr)  # 8

For Python 2 and 3 compatibility, use it as a decorator:

@Singleton.as_decorator
class Class(object):
    pass

ObjectProxy

Create a weak-referenceable proxy to an object that supports the same operations:

from objecttools import ObjectProxy
 
ls = [1, 2, 3]
proxy = ObjectProxy(ls)
 
proxy.append(4)
proxy += [5, 6]
proxy[2:4] = [7, 8, 9]
print(ls)  # [1, 2, 7, 8, 9, 5, 6]

serializable

Create serializable forms of objects (For pickling)

from objecttools import SerializableFunction, SerializableConstant
import pickle
 
file = open('file.txt', 'w')
 
f = lambda file, a: file.write(a)
 
try:
    # Cannot pickle files, even though it is a global constant
    gile = pickle.loads(pickle.dumps(file))
except TypeError:
    gile = pickle.loads(pickle.dumps(SerializableConstant('file', __name__))).value
 
try:
    # Cannot pickle functions if they are lambdas
    # Or are inner functions
    # Or are deleted after creation
    # Or are methods
    g = pickle.loads(pickle.dumps(f))
except pickle.PicklingError:
    g = pickle.loads(pickle.dumps(SerializableFunction(f)))
 
g(gile, 'data')  # Works

Installing

From PyPI

$ pip install objecttools

From source

$ git clone 'https://github.com/MitalAshok/objecttools.git'
$ python ./objecttools/setup.py install