Changelog

All notable changes to text_sensitivity will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

Unreleased

0.3.4 - 2024-11-20

Changed

  • Moved to GitHub

Fixed

  • Various security fixes

0.3.3 - 2023-01-28

Added

  • Ensured full documentation

Changed

  • Moved documentation to sphinx

  • Added generation of unique attributes when using from_pattern

  • Only show outer tqdm in input_space_robustness

Fixed

  • Bugfix with lazy loading of locales

0.3.2 - 2022-03-21

Fixed

  • Bugfix when Notebook UI

0.3.1 - 2022-03-16

Changed

  • Custom sorting of metrics in return types

  • Requires genbase>=0.2.4

  • Requires text_explainability>=0.6.1

  • Moved version information

0.3.0 - 2022-03-04

Added

  • Fairness metrics

Changed

  • Requires genbase>=0.2.2

  • Requires text_explainability>=0.6.0

  • Renamed pyproject.toml to .portray to avoid build errors

  • Do not render conf_mat for model metrics

Fixed

  • Added person provider to city generation in English

0.2.6 - 2021-12-06

Added

  • Return type and notebook UI rendering for mean score

  • Return type and notebook UI for invariance test

Changed

  • Requires genbase>=0.1.14

  • Requires text_explainability>=0.5.8

0.2.5 - 2021-12-02

Added

  • Return types for sensitivity tests

  • Rendering of sensitivity tests using genbase.ui

Changed

  • Moved SeedMixin and CaseMixin to genbase

  • Use genbase.internationalization

  • Requires genbase>=0.1.13

0.2.4 - 2021-11-16

Fixed

  • Bugfix in OneToOnePerturbation.from_dictionary

  • Bugfix in compare_metric

  • Bugfix in one_to_one_dictionary_mapping

0.2.3 - 2021-11-16

Fixed

  • Bugfix in oneway_dictionary_mapping

0.2.2 - 2021-11-15

Fixed

  • Bugfix in one_to_one_dictionary_mapping

0.2.1 - 2021-11-03

Added

  • Invariance testing

  • Comparison of mean scores (labelwise)

0.2.0 - 2021-10-08

Added

  • Random license plate generation

  • Added SeedMixin to WordList

  • Added CaseMixin to WordList

  • Robustness testing for random inputs

  • Generate data from patterns

  • Example usage for robustness testing and data generation

Changed

  • Ability to generate items from WordList

0.1.10 - 2021-10-07

Added

  • Perturbation imports (character, word, sentence) to text_sensitivity.perturbation

  • Examples in README.md

  • Attribute renaming in text_sensitivity.data.random.entity

Changed

  • Updated usage with text_explainability==0.5.0

  • Updated usage with faker==8.16.0

0.1.9 - 2021-10-02

Fixed

  • Bugfix in reading .csv files

0.1.8 - 2021-10-02

Removed

  • Removed cities from wordlists

0.1.7 - 2021-10-02

Added

  • MANIFEST.in

  • Security tests with bandit

  • Ability to make random entities lowercase, uppercase or sentencecase

  • Tests for text_sensitivity.data.random.string

  • Tests for text_sensitivity.data.random.entity

  • Additional documentation

  • Ability to generate addresses/cities in a country with a likelihood based on their population

Removed

  • Removed countries from wordlists

Fixed

  • Bugfixes in OneToOnePerturbation and OneToManyPerturbation

0.1.6 - 2021-10-02

Changed

  • Moved random string data generation from text_sensitivity.data.random to text_sensitivity.data.random.string

  • Renamed RandomData to RandomString

  • Seed behavior generalized in SeedMixin, only requiring a self._seed and self._original_seed to work with a class

Added

  • Random multilingual entity generation with Python package faker

  • Documentation and example usages for random entity generation

0.1.5 - 2021-10-01

Added

  • Internationalization support

  • Name of countries by language word list

  • Top 100 most populous cities by country word list

0.1.4 - 2021-09-30

Added

  • Citation information

  • Documentation styling

  • Generation of random Cyrillic text

0.1.3 - 2021-09-27

Added

  • Documentation

  • Ability to make OneToOnePerturbation from unordered list

  • Extended one-to-one and one-to-many dictionary mappings

0.1.2 - 2021-09-24

Changed

  • Proper n-times application of function with OneToManyPerturbation

Fixed

  • Bugfix in character generation

0.1.1 - 2021-09-24

Added

  • Example usage

  • Sensitivity testing wrapper functions (compare accuracy, precision, recall)

0.1.0 - 2021-09-24

Added

  • Random data generation

  • One to one perturbation

  • One to many perturbation

  • Example perturbation functions

  • README.md

  • LICENSE

  • CI/CD pipeline for flake8 testing

  • setup.py