All Versions
Latest Version
Avg Release Cycle
87 days
Latest Release
1521 days ago

Changelog History
Page 1

  • v0.18.0

    February 13, 2020
  • v0.17.0 Changes

    August 20, 2018
    • 👉 Make benchmarks script Py3 compatible. [Stefan Behnel]

    • ➕ Add Go lang port. [iddober]

    • ➕ Add reference to C# port. [ericcoleman]

    • 🚚 Chore: remove license header from files. [Jose Diaz-Gonzalez]

    The files should all inherit the projects license.

    • 🛠 Fix README title style. [Thomas Grainger]

    • ➕ Add readme check. [Thomas Grainger]

    install docutils and Pygments

    • Cache pip. [Thomas Grainger]

    • ⬆️ Upgrade pip/setuptools for hypothesis. [Thomas Grainger]

    • 👍 Feat: drop py26 and py33 support from tox. [Jose Diaz-Gonzalez]

    • ✅ Feat: drop support for 2.6 in [Jose Diaz-Gonzalez]

    • Feat: drop reference to 2.4 from readme. [Jose Diaz-Gonzalez]

    • Feat: drop py2.6 and py3.3 classifiers. [Jose Diaz-Gonzalez]

    • 👍 Feat: drop 2.6 and 3.3 support. [Jose Diaz-Gonzalez]

    These are no longer supported. Please upgrade your python version if you are using either version.

    • Fuzz: _token_sort: check for equivalence. [Ralf Ramsauer]

    If we don't have to full_process the strings, we can safely assume to return 100 in case both candidates equal.

    Signed-off-by: Ralf Ramsauer [email protected]

    • ✅ Test: add more test cases. [Ralf Ramsauer]

    Signed-off-by: Ralf Ramsauer [email protected]

    • Utils: add and use check_for_equivalence decorator. [Ralf Ramsauer]

    And decorate basic scoring functions.

    The check_for_equivalence decorator MUST be used after the check_for_none decorator, as otherwise ratio(None, None) will get a score of 100.

    This fixes the first part of the recently introduced changes in the test set.

    Signed-off-by: Ralf Ramsauer [email protected]

    • ✅ Tests: add some corner cases. [Ralf Ramsauer]

    '' and '' are equal, so are '{' and '{'. Test if fuzzywuzzy gives them a score of 100.

    For the moment, this patch breaks tests, fixes in fuzzywuzzy follow.

    Signed-off-by: Ralf Ramsauer [email protected]

    • 🚚 Utils: remove superfluous check. [Ralf Ramsauer]

    Decorators make sure that only non None-values are passed. We can safely assume that None will never get here.

    Other than that, None's shouldn't simply be ignored and erroneously changed to empty strings. Better let users fail.

    This commit doesn't break any tests.

    Signed-off-by: Ralf Ramsauer [email protected]

    • README: add missing requirements. [Ralf Ramsauer]

    pycodestyle and hypothesis are required for automatic testing. Add them to README's requirement section.

    Signed-off-by: Ralf Ramsauer [email protected]

    • ✂ Remove empty document. [Ralf Ramsauer]

    Signed-off-by: Ralf Ramsauer [email protected]

  • v0.16.0 Changes

    December 18, 2017

    ➕ Add punctuation characters back in so process does something.

    Simpler alphabet and even fewer examples. [davidcellis]

    Fewer examples and larger deadlines for Hypothesis. [davidcellis]

    Slightly more examples. [davidcellis]

    ✅ Attempt to fix the failing 2.7 and 3.6 python tests. [davidcellis]

    Readme: add link to C++ port. [Lizard]

    🛠 Fix tests on Python 3.3. [Jon Banafato]

    Modify tox.ini and .travis.yml to install enum34 when running with
    ✅ Python 3.3 to allow hypothesis tests to pass.

    Normalize Python versions. [Jon Banafato]

    • Enable Travis-CI tests for Python 3.6
    • Enable tests for all supported Python versions in tox.ini
    • Add Trove classifiers for Python 3.4 - 3.6 to

    👍 Note: Python 2.6 and 3.3 are no longer supported by the Python core
    👍 team. Support for these can likely be dropped, but that's out of scope
    for this change set.

    🛠 Fix typos. [Sven-Hendrik Haase]

  • v0.15.1 Changes

    July 19, 2017

    🛠 Fix (addresses #155) [Paul O'Leary McCann]

    🔀 Merge remote-tracking branch 'upstream/master' into
    extract_optimizations. [nolan]

    👀 Seed random before generating benchmark strings. [nolan]

    Cleaner implementation of same idea without new param, but adding
    existing full_process param to Q,W,UQ,UW. [nolan]

    🛠 Fix benchmark only generate list once. [nolan]

    Only run util.full_process once on query when using extract functions,
    ➕ add new benchmarks. [nolan]

  • v0.15.0 Changes

    February 20, 2017

    ➕ Add extras require to install python-levenshtein optionally. [Rolando

    This allows to install python-levenshtein as dependency.

    🛠 Fix link formatting in the README. [Alex Chan]

    ➕ Add fuzzball.js JavaScript port link. [nolan]

    ➕ Added Rust Port link. [Logan Collins]

    📄 Validate_string docstring. [davidcellis]

    ✅ For full comparisons test that ONLY exact matches (after processing)
    are added. [davidcellis]

    ➕ Add detailed docstrings to WRatio and QRatio comparisons.

  • v0.14.0 Changes

    November 04, 2016
    • ✅ Possible PEP-8 fix + make pep-8 warnings appear in test. [davidcellis]

    • Possible PEP-8 fix. [davidcellis]

    • Possible PEP-8 fix. [davidcellis]

    • ✅ Test for stderr log instead of warning. [davidcellis]

    • ⚠ Convert warning.warn to logging.warning. [davidcellis]

    • ➕ Additional details for empty string warning from process. [davidcellis]

    String formatting fix for python 2.6

    • ⚠ Enclose warnings.simplefilter() inside a with statement. [samkennerly]
  • v0.13.0 Changes

    November 01, 2016
    • 👌 Support alternate git status output. [Jose Diaz-Gonzalez]

    • ✅ Split warning test into new test file, added to travis execution on 2.6 / pypy3. [davidcellis]

    • ✂ Remove hypothesis examples database from gitignore. [davidcellis]

    • ➕ Add check for warning to tests. [davidcellis]

    Reordered test imports

    • 🚚 Check processor and warn before scorer may remove processor. [davidcellis]

    • ✅ Renamed test - tidied docstring. [davidcellis]

    • ➕ Add token ratios to the list of scorers that skip running full_process as a processor. [davidcellis]

    • Added tokex_sort, token_set to test. [davidcellis]

    • ✅ Test docstrings/comments. [davidcellis]

    Removed redundant check from test.

    • ➕ Added py.test .cache/ removed duplicated build from gitignore. [davidcellis]

    • 0️⃣ Added default_scorer, default_processor parameters to make it easier to change in the future. [davidcellis]

    Added warning if the processor reduces the input query to an empty string.

    • 0️⃣ Rewrote extracts to explicitly use default values for processor and scorer. [davidcellis]

    • 🔄 Changed Hypothesis tests to use pytest parameters. [davidcellis]

    • ➕ Added Hypothesis based tests for identical strings. [Ducksual]

    Added support for hypothesis to travis config. Hypothesis based tests are skipped on Python 2.6 and pypy3.

    Added .hypothesis/ folder to gitignore

    • ➕ Added test for simple 'a, b' string on process.extractOne. [Ducksual]

    • 🖨 Process the query in process.extractWithoutOrder when using a scorer which does not do so. [Ducksual]

    Closes 139

    • Mention that difflib and levenshtein results may differ. [Jose Diaz- Gonzalez]

    Closes #128

  • v0.12.0 Changes

    September 14, 2016
    • 👍 Declare support for universal wheels. [Thomas Grainger]

    • Clarify that license is GPLv2. [Gareth Tan]

  • v0.11.1 Changes

    July 27, 2016
    • ➕ Add editorconfig. [Jose Diaz-Gonzalez]

    • ➕ Added tox.ini cofig file for easy local multi-environment testing changed travis config to use py.test like tox updated use of pep8 module to pycodestyle. [Pedro Rodrigues]

  • v0.11.0 Changes

    June 30, 2016

    Clean-up. [desmaisons_david]

    🐎 Improving performance. [desmaisons_david]

    🐎 Performance Improvement. [desmaisons_david]

    🛠 Fix link to Levenshtein. [Brian J. McGuirk]

    🛠 Fix readme links. [Brian J. McGuirk]

    ➕ Add license to [Jose Diaz-Gonzalez]

    Closes #113