hashable_collections 1.1

About

This library enables dictionaries and lists to be used as keys inside a dictionary. In other words, it makes them hashable.

While mutable objects (lists, dictionaries) should not be made hashable in general, the programmer is trusted not to abuse this power (you do test your code before shipping into production, don’t you?). An extension option to explicitly make the resulting hashable object immutable is provided below for convenience.

There are certain use cases - e.g. when trying to pass seemingly identical keys to NetworkX as different nodes, or when you want to query an object’s properties in place while making a mapping instead of calling a complicated identification function - where these may be useful and reduce boilerplate code.

Usage

>>> from hashable_collections.hashable_collections import hashable_dict
>>> hashed_dict = hashable_dict({'name':'foo'})
>>> from hashable_collections.hashable_collections import hashable_list
>>> hashed_list = hashable_list(['bar',1])
>>> output = {hashed_dict:0,hashed_list:1}
>>> output
{{'name':'foo'}: 0, ['bar', 1]: 1}

These hashable objects still retain mutability - in other words, it is still possible to assign hashed_dict['squiggle'] = 'squiggle' and have it execute without an issue. The hash will change correspondingly, effectively making it a different object from before.

Alternatively, one can subclass the objects defined here and override the __setitem__ method to make sure the resulting hashable dictionary/list is immutable:

class immutable_hashable_dict(hashable_dict):

    def __setitem__(self,key,value):
        raise ValueError('Immutable hashable dicts do not let you change values inside your dictionary.')

class immutable_hashable_list(hashable_list):

    def __setitem__(self,key,value):
        raise ValueError('Immutable hashable lists do not let you change valus inside your list.')

Caveats

The following is a list of known issues that may crop up with this technology.

• Creating a copy: If s is a hashable_dict, and one does m = s and then changes m[0], the corresponding value of s[0] is changed as well. Never assign an existing hashable_dict or hashable_list to another object unless you are sure the second object is not going to change.

Installation

pip install hashable_collections