pyxDamerauLevenshtein

1.10.1 verified Mon Apr 27 auth: no python

High-performance Damerau-Levenshtein edit distance computation for Python, implemented in Cython. Currently at version 1.10.1, with wheels for CPython 3.9–3.14 on Linux, macOS, and Windows. Releases are stable and infrequent.

pip install pyxdameraulevenshtein

Common errors

error ModuleNotFoundError: No module named 'pyxdameraulevenshtein' ↓

cause Library not installed or wrong import name (common typo: pyxdameraulevenshtein vs pyxdamerau-levenshtein).

fix

Run: pip install pyxdameraulevenshtein

error TypeError: expected string or bytes-like object ↓

cause Passed non-string (e.g., int, list) to distance function.

fix

Convert arguments to strings: str(arg1), str(arg2)

error ImportError: cannot import name 'damerau_levenshtein_distance' from 'pyxdameraulevenshtein' ↓

cause Old library version (<1.0.0) may have different API. Current version 1.10.1 has this function.

fix

Upgrade: pip install --upgrade pyxdameraulevenshtein

Warnings

gotcha The library uses Cython and ships prebuilt wheels for most platforms. On unsupported architectures (e.g., ARM Linux without manylinux), install from source requires a C compiler. ↓

fix Install via pip; if source build fails, install a C compiler (gcc, clang) and ensure Python development headers are available.

gotcha Input strings must be of the same type (both str or both bytes). Mixing types may raise a TypeError. ↓

fix Ensure both arguments are strings (or both bytes) before calling the distance functions.

gotcha The normalized distance function returns 0.0 for identical strings and 1.0 for completely different strings. It is not a symmetric normalized version of edit distance in the typical Levenshtein sense; it uses max length of the two strings as denominator. ↓

fix Understand that normalized_damerau_levenshtein_distance computes distance / max(len(s1), len(s2)).

Imports

damerau_levenshtein_distance
wrong
```
import pyxdameraulevenshtein
```
correct
```
from pyxdameraulevenshtein import damerau_levenshtein_distance
```
Direct import without function name requires qualifying the call, but most users expect the function directly.

normalized_damerau_levenshtein_distance

wrong

from pyxdameraulevenshtein import normalized_distance

correct

from pyxdameraulevenshtein import normalized_damerau_levenshtein_distance

Function name is unique; common typo omits 'damerau'.

Quickstart

Compute Damerau-Levenshtein distance between two strings. Works for any hashable sequences (e.g., lists of tokens).

from pyxdameraulevenshtein import damerau_levenshtein_distance

string1 = 'hello'
string2 = 'hola'
dist = damerau_levenshtein_distance(string1, string2)
print(f"DL distance: {dist}")

# Normalized distance (0.0 to 1.0)
from pyxdameraulevenshtein import normalized_damerau_levenshtein_distance
norm_dist = normalized_damerau_levenshtein_distance(string1, string2)
print(f"Normalized DL distance: {norm_dist:.4f}")