NewYour coding agent can read the release notes before it upgrades.Set up the MCP server →
PyPI
PyPI, #949 most downloaded on PyPI
Python extension for computing string edit distances and similarities.
Latest stable
0.27.5
Last release 2 months ago
12 Sep 2026
Release timing varies
gaps range from 2 weeks to 9 months
Nearly every release is documented
notes for 36 of 36 stable releases
Nothing withdrawn
no release was ever pulled
6 years old
36 releases · first in 2021
Drag across the ruler to compare, or tap a release to open it.
avoid unnecessary len calls on iterables. This avoids a crash when changing the size of the iterable after the len call.
add type hints for StringMatcher
Your coding agent can read these notes before it upgrades. Set up the MCP server →
add missing pypy and freethreaded linux wheels
fix build with scikit-build-core v0.11
scikit-build-core v0.11Cython==3.0.12## Fixed - add missing mac os wheels
switch build system to scikit-build-core
scikit-build-coreCython==3.0.11fix potentially incorrect results of jaro_winkler when using high prefix weights
jaro_winkler when using high prefix weights## Changed - improve type hints
upgrade rapidfuzz-cpp to v3.0.0
rapidfuzz-cpp to v3.0.0added keyword argument pad to Hamming distance. This controls whether sequences of different length should be padded or lead to a ValueError
pad to Hamming distance. This controls whether sequences of different
length should be padded or lead to a ValueErrorCython==3.0.3## Changed - add support for Python 3.12 - drop support for Python 3.6 ## Added - add wheels for windows arm64
upgrade rapidfuzz-cpp to v2.0.0
rapidfuzz-cpp to v2.0.0relax dependency requirement on rapidfuzz
rapidfuzzfix function signature of get_requires_for_build_wheel
get_requires_for_build_wheeltype hints for editops/opcodes/matching_blocks did not allow any hashable sequence
editops/opcodes/matching_blocks did not allow any
hashable sequenceeditops/opcoded/matching_blocks did not allow any
hashable sequencetype hints did not get installed
fix incorrect result normalization in setratio and seqratio
setratio and seqratiofix support for cmake versions below 3.17
rapidfuzz-cpp when building against a previously installed versionmodernize cmake build to fix most conda-forge builds
## Changed - Added support for Python3.11
fix matching_blocks conversion for empty editops
fix broken matching_blocks conversion
use matching_blocks/apply/remove_subsequence/inverse implementation from RapidFuzz
matching_blocks/apply/remove_subsequence/inverse implementation from RapidFuzz## Added - add musllinux wheels
## Added - add missing type hints
#### Added - Add type hints
replace usage of deprecated Python APIs
Allow installation from system installed versions of rapidfuzz-cpp
rapidfuzz-cppIndel.normalized_similarity was broken in RapidFuzz v2.0.0 (see #20)
Fixed memory leak in error path of setratio
## Removed - removed support for Python 3.5
## Changed * Add support for Python 3.10
Update SequenceMatcher interface to support the autojunk parameter
Implement distance/ratio/hamming/jaro/jaro_winkler using rapidfuzz instead of providing a own implementation
Fix usage of deprecated Unicode APIs in distance, ratio and hamming
Fixed a bug in StringMatcher.StringMatcher.get_matching_blocks / extract_editops for Python 3; now allow only str editops on both Python 2 and Python
str editops on
both Python 2 and Python 3, for simpler and working code.One column per quarter.