This is a Python implementation of the Unicode Collation Algorithm (UCA). It passes 100% of the UCA conformance tests for Unicode 5.2.0 (Python 2.7), Unicode 6.3.0 (Python 3.3+), Unicode 8.0.0 (Python 3.5+), Unicode 9.0.0 (Python 3.6+), and Unicode 10.0.0 (Python 3.7+) with a variable-weighting setting of Non-ignorable.

In short, sorting non-English strings properly.

The core of the algorithm involves multi-level comparison. For example, café comes before caff because at the primary level, the accent is ignored and the first word is treated as if it were cafe. The secondary level (which considers accents) only applies then to words that are equivalent at the primary level.

The Unicode Collation Algorithm and pyuca also support contraction and expansion. Contraction is where multiple letters are treated as a single unit. In Spanish, ch is treated as a letter coming between c and d so that, for example, words beginning ch should sort after all other words beginnings with c. Expansion is where a single letter is treated as though it were multiple letters. In German, ä is sorted as if it were ae, i.e. after ad but before af.

Here is how to use the pyuca module.

pip install pyuca

Usage example:

from pyuca import Collator
c = Collator()

assert sorted(["cafe", "caff", "café"]) == ["cafe", "caff", "café"]
assert sorted(["cafe", "caff", "café"], key=c.sort_key) == ["cafe", "café", "caff"]

Collator can also take an optional filename for specifying a custom collation element table.

You can also import collators for specific Unicode versions, e.g. from pyuca.collator import Collator_8_0_0. But just from pyuca import Collator will ensure that the collator version matches the version of unicodata provided by the standard library for your version of Python.

Tauber, J. K. (2016). pyuca: a Python implementation of the Unicode Collation Algorithm. The Journal of Open Source Software. DOI: 10.21105/joss.00021

Python code is made available under an MIT license (see LICENSE). allkeys.txt is made available under the similar license defined in LICENSE-allkeys.

If you have any problems, questions or suggestions, it's best to file an issue on GitHub although you can also contact me at jtauber@jtauber.com.

For more of my work on linguistics and Ancient Greek, see http://jktauber.com/.