pydicti 1.1.4

Case insensitive derivable dictionary

Installation

You can install the newest version of pydicti from PyPI:

pip install pydicti

Alternatively, you can just take the file pydicti.py and redistribute it with your application.

Overview

• class dicti: default case insensitive dictionary type
• class odicti: ordered case insensitive dictionary type
• def build_dicti: create a case insensitive dictionary class
• def Dicti: create a case insensitive copy of a dictionary

dicti

Objects of type dicti are dictionaries that feature case insensitive item access:

>>> d = dicti(Hello='foo', world='bar')
>>> d['heLLO']
'foo'
>>> 'WOrld' in d
True

Internally however, the keys retain their original case:

>>> sorted(d.keys())
['Hello', 'world']

odicti

The type odicti instanciates order-preserving case insensitive dictionaries:

>>> odicti(zip('abc', range(3)))
Dicti(OrderedDict([('a', 0), ('b', 1), ('c', 2)]))

build_dicti

With build_dicti you can create custom case insensitive dictionaries. This function is what is used to create the pydicti.dicti and pydicti.odicti types. Note that calling build_dicti several times with the same argument will result in identical types:

>>> build_dicti(dict) is dicti
True
>>> build_dicti(OrderedDict) is odicti
True

build_dicti uses subclassing to inherit the semantics of the given base dictionary type:

>>> issubclass(odicti, OrderedDict)
True

Dicti

The function Dicti is convenient for creating case insensitive copies of dictionary instances:

>>> o = OrderedDict(zip('abcdefg', range(7)))
>>> oi = Dicti(o)
>>> type(oi) is odicti
True

JSON

The subclassing approach allows to plug your dictionary instance into places where typechecking with isinstance is used, like in the json module:

>>> import json
>>> d == json.loads(json.dumps(d), object_hook=dicti)
True

You can use json.loads(s, object_pairs_hook=odicti) to deserialize ordered dictionaries.

Pitfalls

The equality comparison tries preserves the semantics of the base type as well as reflexitivity. This has impact on the transitivity of the comparison operator:

>>> i = dicti(oi)
>>> roi = odicti(reversed(list(oi.items())))
>>> roi == i and i == oi
True
>>> oi != roi and roi != oi  # NOT transitive!
True
>>> oi == i and i == oi      # reflexive
True

The coercion rules in python allow this to work pretty well when performing comparisons between types that are subclasses of each other. Be careful otherwise, however.

License

Copyright © 2013 Thomas Gläßle <[e-mail]>

This work is free. You can redistribute it and/or modify it under the terms of the Do What The Fuck You Want To Public License, Version 2, as published by Sam Hocevar. See the COPYING file for more details.

This program is free software. It comes without any warranty, to the extent permitted by applicable law.

CHANGELOG

1.1.4

Date: 17.10.2020

• use str.casefold() on python3
• make normalization function a parameter of build_dict, so that user-defined normalization functions can be passed

1.1.3

Date: 28.06.2019

• avoid key recomputation in __setitem__

1.1.2

Date: 28.06.2019

• leave item order invariant under assignment in odicti (#2)
• leave key case invariant under assignment

1.1.1

Date: 25.03.2019

• fix deprecated MutableMapping import (error on py38)

1.1.0

Date: 19.03.2019

• drop py2.6 support
• fix version number in long_description

1.0.0

Date: 19.03.2019

• make str representation more dict-like
• misc cleanup

0.0.6

Date: 08.09.2016

• fix UnicodeDecodeError in setup when UTF-8 is not the default encoding

0.0.5

Date: 18.05.2016

• fix pickling on py 3.5
• the ‘name’ parameter to build_dicti can now be a qualname

0.0.4

Date: 01.02.2014

• add coverage reports
• use more extensive unit tests
• add support for pickle

0.0.3

Date: 26.01.2014

• add support for python26
• make dependency on OrderedDict optional
• migrate to setuptools in order to use testing commands
• support ordereddict.OrderedDict as fallback

0.0.2

Date: 29.12.2013

• fix dicti.pop
• support deepcopy(dicti)
• make nosetest automatically execute the doctests