v1.3.0

What's Changed

Breaking Changes 🛠

• Raise minimum Python version to 3.10 by @tmke8 in #974

Exciting New Features 🎉

• Use bool mask in torch code by @tmke8 in #972
• Enable the package for Python 3.12 by @tmke8 in #1006

Other Changes

• Fix that cursed file again by @tmke8 in #971
• Make the other mask also bool by @tmke8 in #973
• Remove all from __future__ import annotations by @tmke8 in #975
• Clean up the CI by @tmke8 in #981
• Restore the use of python-type-stubs by @tmke8 in #983
• Convert a lot of if-elif cascades into match statements by @tmke8 in #984
• Replace darglint with pydoclint by @tmke8 in #985
• Make zip strict and avoid boolean traps by @tmke8 in #986
• Try to make mypy happy by @tmke8 in #993
• Use ints for the labels in the tests by @tmke8 in #994
• Normalize strings with black by @tmke8 in #995

Full Changelog: v1.2.3...v1.3.0

v1.2.3

What's Changed

• Update folktables by @tmke8 in #964

Full Changelog: v1.2.2...v1.2.3

v1.2.2

What's Changed

Bugfixes

• Fix multigrid plotting by @tmke8 in #958

Other Changes

• Disable number of args check by @tmke8 in #905
• Update to Poetry 1.5 by @tmke8 in #906
• Increase coverage of mypy by @tmke8 in #907
• Disable dependabot auto-merge by @tmke8 in #913
• Update type stubs by @tmke8 in #914
• Make ruff happy by @tmke8 in #927
• Make compatible with mypy 1.5 by @tmke8 in #932
• Some fixes by @tmke8 in #935
• Add explicit check=False by @tmke8 in #938
• Split off the dev dependencies for individual CI steps by @tmke8 in #939
• Modernize sphinx deployment by @tmke8 in #942
• Improve formatting slightly by @tmke8 in #949

Full Changelog: v1.2.1...v1.2.2

v1.2.1

What's Changed

Exciting New Features 🎉

• Update to Pandas 2.0 by @tmke8 in #887

Other Changes

• Update ruff and update config by @tmke8 in #886
• Address an ancient Seaborn deprecation warning by @tmke8 in #888
• Update ruff by @tmke8 in #892

Full Changelog: v1.2.0...v1.2.1

v1.2.0

What's Changed

Exciting New Features 🎉

• Raise highest supported Python version to 3.11 by @tmke8 in #879

Other Changes

• Use new setting in ruff to set the docstring convention by @tmke8 in #836
• Don't install the project itself with pdm by @tmke8 in #837
• Update ruff by @tmke8 in #838
• Update sklearn stubs by @tmke8 in #846
• Use kernel.value instead of kernel.name by @tmke8 in #847
• Update ruff by @tmke8 in #850
• Update to black 2023 by @tmke8 in #851
• Add more linters from ruff by @tmke8 in #852
• Update mypy by @tmke8 in #856
• Add naming lint by @tmke8 in #857
• Use Self type now that mypy supports it by @tmke8 in #858
• Update continuous_integration.yml by @tmke8 in #861
• Update ruff by @tmke8 in #860
• Add some ruff ignores by @tmke8 in #867
• Fix some type issues by @tmke8 in #868
• Try running the full CI only in the merge queue by @tmke8 in #874
• Update type stubs by @tmke8 in #884

Full Changelog: v1.1.0...v1.2.0

v1.1.0

What's Changed

Other Changes

• Improve pandas code by @thomkeh in #820
• Fix a problem with the darglint CI by @thomkeh in #821
• Fix a type error by @thomkeh in #823
• Add CPU version of torch to poetry by @thomkeh in #822
• Format with pytest-style by @thomkeh in #828
• Update Zafar to Python 3 by @thomkeh in #829
• Format with black preview by @thomkeh in #830
• Use ruff instead of isort by @thomkeh in #832
• Update Kamishima to Python3 by @thomkeh in #834

Full Changelog: v1.0.2...v1.1.0

v1.0.2

What's Changed

• Fix example by @thomkeh in #776
• Add tests for the column generating code by @thomkeh in #782
• Update the config for pyright by @thomkeh in #792
• Update to mypy 0.990 by @thomkeh in #799
• Use @override instead of @implements by @thomkeh in #807
• Remove unused imports by @thomkeh in #808
• Use more ruff checks by @thomkeh in #809
• Pyre config file by @thomkeh in #810
• Remove duplicate import by @thomkeh in #813
• Add "pyright: ignore" comments by @thomkeh in #814
• Update ruff to 176 by @thomkeh in #815
• Import folktables only on demand by @thomkeh in #819

Full Changelog: v1.0.1...v1.0.2

v1.0.1

What's Changed

Documentation improvements ✍️

• Restrict depth of TOC in the docs by @thomkeh in #769
• Update tutorial for EthicML 1.0 by @thomkeh in #770
• Fill in a lot of missing documentation by @thomkeh in #774

Other Changes

• Disallow implicit reexport with mypy by @thomkeh in #773
• Update release.yml by @thomkeh in #775

Full Changelog: v1.0.0...v1.0.1

v1.0.0

What's Changed

Breaking Changes 🛠

• Remove vision by @olliethomas in #547
• Change default of delete_prev to True by @thomkeh in #624
• Remove the lowercase variants of the dataset objects by @thomkeh in #643
• issue 647 by @olliethomas in #653
• Rename some things by @thomkeh in #687
• Enable run_metrics() to compute min and max in addition to diff and ratio by @thomkeh in #736
• Add two more namespaces: plot and run by @thomkeh in #753
• Move run_metrics by @thomkeh in #761

Exciting New Features 🎉

• Use sets of enums instead of flags for metrics aggregation by @thomkeh in #752

Other Changes

• Import metrics with star import by @thomkeh in #555
• Update type stubs by @thomkeh in #557
• Quick follow-up to the type stub update by @thomkeh in #558
• Remove TAP and reduce code duplication by @thomkeh in #559
• Add RobustAccuracy Metric by @olliethomas in #561
• PR labeling by @thomkeh in #566
• Remove PR labeler by @thomkeh in #570
• Exclude more paths from the full CI by @thomkeh in #571
• Use a tagged union of TypedDict for the flags by @thomkeh in #573
• Update sqf.py by @olliethomas in #562
• Merge main by @olliethomas in #580
• Trigger dependabot auto-merge only on main branch by @thomkeh in #582
• Update lock file after solving merge conflicts by @thomkeh in #581
• Merge main into next by @thomkeh in #583
• Fix docstring of SVM by @thomkeh in #588
• Turn SVM model class into a dataclass by @thomkeh in #589
• Automatically add label to PRs for the "next" branch by @thomkeh in #591
• Merge main into next by @thomkeh in #592
• Replace most Literals with Enums by @olliethomas in #590
• Replace ray with joblib.Parallel by @thomkeh in #598
• Convert docstrings from google format to sphinx format by @thomkeh in #599
• Use GitHub Copilot to generate docstrings by @thomkeh in #600
• Remove AIF360 dependency by @thomkeh in #602
• Extend gitignore by @thomkeh in #607
• Turn Agarwal into a dataclass by @thomkeh in #608
• Store s and y as pd.Series by @thomkeh in #609
• More OmegaConf tests by @thomkeh in #610
• Pass seed to run by @thomkeh in #612
• Move pydocstyle config into pyproject.toml by @thomkeh in #614
• add option to specify test set size by @olliethomas in #615
• Clean-ups from the rename by @thomkeh in #620
• Try to fix evaluate_models() by @thomkeh in #622
• add graceful failure... well at least complete and don't crash by @olliethomas in #626
• closes #623. by @olliethomas in #625
• tight layout by @olliethomas in #632
• Check whether docs can be built when running CI by @thomkeh in #636
• Fix example in docs by @thomkeh in #637
• rename train_test_split.py file to splits.py by @olliethomas in #639
• Add aliases for dataset splits within the dataset classes by @thomkeh in #641
• Use an abstract base class instead of Protocol for metrics by @thomkeh in #644
• Actually use the model seed in evaluate_models by @thomkeh in #642
• Update fairlearn to 0.7.0 by @thomkeh in #645
• Make model path absolute by @thomkeh in #649
• Update OmegaConf and remove PyTorch from dependencies by @thomkeh in #650
• Run poetry update by @thomkeh in #651
• Fix a few things that I noticed by @thomkeh in #652
• Hgr method by @olliethomas in #658
• Fix a long-standing type error by @thomkeh in #660
• add FairDummies method by @olliethomas in #646
• add new models to main __init__ by @olliethomas in #662
• Use single DataFrame as underlying storage for DataTuple by @thomkeh in #659
• Don't let DataTuple inherit from TestTuple by @thomkeh in #664
• Various improvements by @thomkeh in #668
• Add an optional name argument to replace_data() by @thomkeh in #669
• Parallel pytest by @thomkeh in #670
• Re-order tests to make better use of parallel execution by @thomkeh in #671
• Major reorg by @thomkeh in #672
• Improve docs structure by @thomkeh in #673
• Modularize the dataset definition by @thomkeh in #675
• Rename to_npz and from_npz to not mention npz anymore by @thomkeh in #678
• Fix code syntax for bibtex in law.py by @thomkeh in #679
• Improve docs and update dev branch version by @thomkeh in #681
• Add migration guide to 1.0 by @thomkeh in #682
• Add settings for auto-generated release notes by @thomkeh in #683
• Add a tutorial for the Adult dataset by @thomkeh in #684
• Use patched nbsphinx (because of dark mode) by @thomkeh in #685
• Change dependencies for building docs by @thomkeh in #688
• Change the return type of run_metrics() by @thomkeh in #693
• When constructing a datatuple check for overlapping columns by @thomkeh in #692
• Sort out which files get passed to the preprocessing models by @thomkeh in #694
• Use official pandas type stubs by @thomkeh in #695
• Fix usage of nan_to_num by @thomkeh in #698
• Make a method private and mark classes as abstract by @thomkeh in #700
• update pandas stubs by @thomkeh in #716
• Use a new caching strategy for the CI by @thomkeh in #717
• Set python version with an env variable in the CI by @thomkeh in #718
• Check docstrings with darglint by @thomkeh in #721
• Fix a random assortment of things by @thomkeh in #727
• Prepare for upcoming mypy release by @thomkeh in #735
• Update dependabot.yml by @olliethomas in #740
• Fix two long-standing problems by @thomkeh in #743
• Use tilde for long reference by @thomkeh in #747
• tqdm is not actually a dependency of EthicML by @thomkeh in #748
• Fix docs by @thomkeh in #754
• Update mypy by @thomkeh in #767
• Update gitignore by @thomkeh in #768

Full Changelog: v0.7.1...v1.0.0

v1.0.0-beta.2

What's Changed

Breaking Changes 🛠

• Add two more namespaces: plot and run by @thomkeh in #753

Exciting New Features 🎉

• Use sets of enums instead of flags for metrics aggregation by @thomkeh in #752

Other Changes

• Use tilde for long reference by @thomkeh in #747
• tqdm is not actually a dependency of EthicML by @thomkeh in #748

Full Changelog: v1.0.0-beta.1...v1.0.0-beta.2