pyunormalize

A pure Python implementation of the Unicode normalization algorithm independent of Python’s core Unicode database. This package conforms to version 16.0 of the Unicode standard, released in September 2024, and has been rigorously tested for accuracy using the official Unicode test file.

Installation and updates

To install the package, run:

pip install pyunormalize

To upgrade to the latest version, run:

pip install pyunormalize --upgrade

Unicode character database (UCD) version

To retrieve the version of the Unicode character database in use:

>>> from pyunormalize import UCD_VERSION
>>> UCD_VERSION
'16.0.0'

Example usage

>>> from pyunormalize import NFC, NFD, NFKC, NFKD
>>> s = "élève"  # "\u00E9\u006C\u00E8\u0076\u0065"
>>> nfc = NFC(s)
>>> nfd = NFD(s)
>>> nfc == s
True
>>> nfd == nfc
False
>>> " ".join([f"{ord(x):04X}" for x in nfc])
'00E9 006C 00E8 0076 0065'
>>> " ".join([f"{ord(x):04X}" for x in nfd])
'0065 0301 006C 0065 0300 0076 0065'
>>>
>>> s = "⑴ ffi ²"
>>> NFC(s), NFKC(s), NFD(s), NFKD(s)
('⑴ ffi ²', '(1) ffi 2', '⑴ ffi ²', '(1) ffi 2')

>>> from pyunormalize import normalize
>>> normalize("NFKD", "⑴ ffi ²")
'(1) ffi 2'
>>> forms = ["NFC", "NFD", "NFKC", "NFKD"]
>>> [normalize(f, "\u017F\u0307\u0323") for f in forms]
['ẛ̣', 'ẛ̣', 'ṩ', 'ṩ']

Related resources

This implementation is based on the following resources:

• Section 3.11, “Normalization Forms,” in the Unicode core specification, version 16.0.0
• Unicode Standard Annex #15: Unicode Normalization Forms, revision 56

Licenses

The code is licensed under the MIT license.

Usage of Unicode data files is subject to the UNICODE TERMS OF USE. Additional rights and restrictions regarding Unicode data files and software are outlined in the Unicode Data Files and Software License, a copy of which is included as UNICODE-LICENSE.