v0.10.0

July 2024

Highlights

• Dropped support for Python 3.8

• Announcement of deprecations for v1.0.0, includes removal and rework of the current forecasting and transformation frameworks

• New anomaly detection methods: DWT-MLEAD, K-Means, and an adapter for PyOD (@SebastianSchmidl)

• New data loaders for anomaly detection and segmentation benchmark archives (@SebastianSchmidl, @ermshaua)

• New Proximity Tree classifier using aeon distances, Proximity Forest to follow. (@itsdivya1309)

• New feature-based and dummy clusterers (@MatthewMiddlehurst, @aadya940)

• New FLUSS and BinSeg segmenters (@patrickzib)

• Addition of visualisation tools for shapelets (@baraline)

Anomaly Detection

Documentation

• [DOC] Improve anomaly detector documentation (#1647) @SebastianSchmidl

• [DOC] added reference to docstring of MERLIN anomaly detection (#1707) @adm-unl

Enhancements

• [MNT] Move haar wavelet transform to aeon.utils.numba package (#1579) @SebastianSchmidl

• [ENH] Convert stray and add max iterations for MERLIN (#1571) @MatthewMiddlehurst

• [ENH] DWT-MLEAD anomaly detection method (#1580) @SebastianSchmidl

• [ENH] Add sliding window utils for anomaly detection module (#1605) @SebastianSchmidl

• [ENH] K-Means anomaly detector (#1607) @SebastianSchmidl

• [ENH] Loader for TimeEval (anomaly detection) datasets (#1588) @SebastianSchmidl

• [ENH] Add PyOD adapter for anomaly detection (#1586) @SebastianSchmidl

Classification

Bug Fixes

• [BUG] Fixing wrong output activation in MLPClassifier (#1597) @hadifawaz1999

• [ENH,BUG] LearnedShapelets interfaced functions (#1687) @TonyBagnall

• [BUG] Adding modelcheckpoint callback by default to deep learners when user chooses their own callbacks list (#1763) @hadifawaz1999

• [BUG] Store n_timepoints and n_channels in fit to check in transform (#1758) @TonyBagnall

Deprecation

• [DEP] Rocket,RotationForest and FreshPrince deprecations (plus a bug fix) (#1648) @TonyBagnall

Documentation

• [DOC] remove ProximityTree from distance classifier notebook results display (#1745) @TonyBagnall

• [DOC] typo in convolutions notebook (#1769) @TonyBagnall

Enhancements

• [ENH,BUG] LearnedShapelets interfaced functions (#1687) @TonyBagnall

• [ENH] Tidy, deprecation actions and testing for sklearn estimators (#1701) @MatthewMiddlehurst

• [ENH] Implement Proximity Tree classifier using aeon distances (#1566) @itsdivya1309

• [ENH] REDCOMETS tweaks / fixes (#1722) @zy18811

• [ENH] change input type for MrSQMClassifier and deprecate prior to moving package (#1735) @TonyBagnall

• [ENH] Adds class_weight to those classifier that support it. Required for imbalanced datasets. (#1776) @patrickzib

Clustering

Bug Fixes

• [BUG] Remove feature extraction from TSFreshClusterer (#1716) @MatthewMiddlehurst

• [BUG] Cluster predict proba edge cases (#1739) @chrisholder

• [BUG] Adding modelcheckpoint callback by default to deep learners when user chooses their own callbacks list (#1763) @hadifawaz1999

Documentation

• [DOC] Wrong number of reference in Clara (#1640) @futuer-szd

Enhancements

• [ENH] Add DummyClusterer (#1567) @aadya940

• [ENH] Simple feature based clusterers (#1548) @MatthewMiddlehurst

• [ENH] removed unused attribute from k-means (#1727) @chrisholder

Datasets

Bug Fixes

• [BUG] Fix TSAD data loading tests and example notebooks (#1695) @SebastianSchmidl

Enhancements

• [ENH] Loader for TimeEval (anomaly detection) datasets (#1588) @SebastianSchmidl

• [ENH] Time Series Segmentation Benchmark + Human Activity Segmentation Challenge data loaders (#1755) @ermshaua

Maintenance

• [MNT] Set segmentation data loaders for overnight tests only (#1772) @TonyBagnall

Distances

Bug Fixes

• [BUG] lcss bounded below to 0 (#1610) @TonyBagnall

• [BUG] Bugfix MinDist doctests by setting word_length to 8 (#1690) @patrickzib

Enhancements

• [ENH] Lower Bounding Distances for SAX and SFA (#1622) @patrickzib

Forecasting

WARNING: The current forecasting framework is being removed in v1.0.0. To be replaced with a new framework. See the index page for more information.

Bug Fixes

• [BUG] remove none argument for coerce to period (#1514) @TonyBagnall

• [BUG] unexclude mock forecaster (#1609) @TonyBagnall

Networks

Deprecation

• [DEP] Rename BaseDeepNetwork and deprecate inheritance of BaseObject (#1524) @MatthewMiddlehurst

Maintenance

• [MNT] Enhance networks module testing suite (#1631) @aadya940

Regression

Bug Fixes

• [BUG] Remove squaring distances in KNN regression (#1697) @TonyBagnall

• [BUG] Adding modelcheckpoint callback by default to deep learners when user chooses their own callbacks list (#1763) @hadifawaz1999

Deprecation

• [DEP] Rocket,RotationForest and FreshPrince deprecations (plus a bug fix) (#1648) @TonyBagnall

Documentation

• [DOC] Added All Regressors (2nd Attempt) (#1663) @Moonzyyy

Enhancements

• [ENH] Tidy, deprecation actions and testing for sklearn estimators (#1701) @MatthewMiddlehurst

Segmentation

Enhancements

• [ENH] axis=1 function default and removal of axis class default for series transformation and segmentation (#1625) @MatthewMiddlehurst

• [ENH] Add FLUSS segmenter from stumpy (#1693) @patrickzib

• [ENH] ClaSP: Adds parallelization for distance computations and numbarize function calls (#1692) @patrickzib

• [ENH] Adds BinSeg Segmenter from Ruptures library (#1694) @patrickzib

• [ENH] Time Series Segmentation Benchmark + Human Activity Segmentation Challenge data loaders (#1755) @ermshaua

Similarity Search

Enhancements

• [ENH] Update similarity search with new base classes : Query Search (#1508) @baraline

Transformations

WARNING: The current BaseTransformer framework is being removed in v1.0.0. To be replaced with individual frameworks for collection and series transformations. See the index page for more information.

Bug Fixes

• [BUG] Convert rockets to only use 32 bit floats internally (#1612) @TonyBagnall

• [BUG] add check y to BaseSeriesTransformer (#1624) @TonyBagnall

• [BUG] removed njit test exclusions (#1598) @TonyBagnall

• [BUG] fix padder fit_values (#1752) @TonyBagnall

• [BUG] Hydra bugfix (#1771) @MatthewMiddlehurst

Deprecation

• [DEP] Tidy up BaseTransformer testing (#1541) @TonyBagnall

• [DEP] Convert clear sky transformer to new base class (#1565) @MatthewMiddlehurst

• [DEP] Rocket,RotationForest and FreshPrince deprecations (plus a bug fix) (#1648) @TonyBagnall

• [DEP] deprecate PandasAdaptor (#1614) @TonyBagnall

• [DEP] remove Differencer (#1645) @TonyBagnall

• [DEP] Remove augmenters, binning, cos, sqrt (#1668) @TonyBagnall

• [DEP] Remove kalman filter and time since (#1669) @TonyBagnall

• [DEP] Remove ClearSky (#1673) @TonyBagnall

• [DEP] make BoxCox legacy (#1672) @TonyBagnall

• [DEP] remove PCA (#1674) @TonyBagnall

• [DEP] v0.10 move ColumnSelect and IndexSelect to legacy (#1675) @TonyBagnall

• [DEP] deprecate compose (#1679) @TonyBagnall

• [DEP] V0.10 remove ACF/PACF (#1678) @TonyBagnall

• [DEP] Dep fest (#1680) @TonyBagnall

• [DEP] dep fest2 (#1681) @TonyBagnall

• [DEP] dep fest3 (#1682) @TonyBagnall

• [ENH] Add missing import for deprecation (#1721) @TonyBagnall

• [DEP] V0.10 Move DOBIN to legacy (#1731) @TonyBagnall

• [DEP] Deprecate the update function in BaseCollectionTransformer (#1744) @TonyBagnall

• [DEP] Exclusion list (#1730) @TonyBagnall

Documentation

• [DOC] Add SlidingWindowSegmenter to API reference (#1600) @SebastianSchmidl

• [Doc] Images for “Transformations” (#1736) @Abhash297

• [DOC] Improve BaseCollectionTransformer docstrings (#1733) @TonyBagnall

Enhancements

• [MNT] Move haar wavelet transform to aeon.utils.numba package (#1579) @SebastianSchmidl

• [ENH] Lower Bounding Distances for SAX and SFA (#1622) @patrickzib

• [ENH] axis=1 function default and removal of axis class default for series transformation and segmentation (#1625) @MatthewMiddlehurst

• [ENH] Remove two hard coded arrays from minirocket (#1698) @TonyBagnall

• [ENH] Merge MiniRocket and MiniRocketMultivariate (#1708) @TonyBagnall

• [ENH] Merge MultiRocket and MultiRocketMultivariate (#1711) @TonyBagnall

• [ENH] Channel selection rocket wrapper (#1504) @TonyBagnall

• [ENH] Shapelet visualization tools (#1715) @baraline

• [ENH] Revert mini rocket to separate functions for univariate and multivariate (#1781) @TonyBagnall

Unit Testing

Bug Fixes

• [BUG] Fix TSAD data loading tests and example notebooks (#1695) @SebastianSchmidl

• [BUG] Use almost equal in random state testing of deep learning (#1773) @hadifawaz1999

Deprecation

• [ENH] Exclude deprecated estimators (#1547) @TonyBagnall

Enhancements

• [ENH] Exclude deprecated estimators (#1547) @TonyBagnall

• [ENH] Improve and refactor collection test data generation utilities (#1564) @MatthewMiddlehurst

• [ENH] Moves a test file (#1665) @TonyBagnall

• [ENH] Rework series test data generators (#1592) @MatthewMiddlehurst

Maintenance

• [MNT] Add threads limit for numba and TF (#1604) @baraline

• [MNT] Drop Python 3.8 (#1659) @MatthewMiddlehurst

• [MNT] Replace Intel macOS runners with m1 macOS runners in CI (#1585) @chrisholder

• [MNT] Unit testing revamp part 1: check_estimator (#1479) @MatthewMiddlehurst

Refactored

• [ENH] Improve and refactor collection test data generation utilities (#1564) @MatthewMiddlehurst

• [REF] Refactor testing.utils.data_gen into testing.data_generation (#1589) @MatthewMiddlehurst

Visualisations

Enhancements

• [ENH] Shapelet visualization tools (#1715) @baraline

Other

Bug Fixes

• [BUG] Fix broken import causing CI to fail (#1717) @MatthewMiddlehurst

Deprecation

• [DEP] Remove probablistic class based metrics (#1630) @TonyBagnall

• [DEP] remove public VectorizedDF (#1666) @TonyBagnall

• [DEP] Remove _serialize (#1667) @TonyBagnall

• [DEP] v0.10 move ThetaLinesTransformer to legacy (#1676) @TonyBagnall

Documentation

• [DOC] Fixed a typo in mentoring.md (#1700) @futuer-szd

• [MNT,DOC] Remove extension_templates (#1657) @MatthewMiddlehurst

• [DOCS] Add any papers (#1526) @TonyBagnall

• [DOC,MNT] Documentation updates and deprecation warning for v1.0.0 (#1743) @MatthewMiddlehurst

Maintenance

• [MNT] Remove stale branches (#1543) @MatthewMiddlehurst

• [MNT] Fix pre-commit pushes for forks (hopefully) and easier CI label interactions (#1601) @MatthewMiddlehurst

• [MNT] Change pre-commit trigger type (#1661) @MatthewMiddlehurst

• [MNT] Disable pytest threading by default for local runs (#1656) @MatthewMiddlehurst

• [MNT,DOC] Remove extension_templates (#1657) @MatthewMiddlehurst

• [MNT] Exclude docs and examples from release source (#1658) @MatthewMiddlehurst

• [MNT] Change pre-commit workflow checkout ref (#1704) @MatthewMiddlehurst

• [MNT] Set keras bound to fix CI (#1723) @MatthewMiddlehurst

• [DOC,MNT] Documentation updates and deprecation warning for v1.0.0 (#1743) @MatthewMiddlehurst

• [MNT] Removing tensorflow_addons dependency (#1421) @MatthewMiddlehurst

• [MNT] Fix certain workflow actions for forks and add button for resetting CI (#1764) @MatthewMiddlehurst

Other

• [DEP] Simple removals for deprecation list for 0.10 (#1644) @TonyBagnall

• [MNT] Raise tensorflow bound to >2.13 (#1765) @TonyBagnall

Contributors

The following have contributed to this release through a collective 107 GitHub Pull Requests:

@aadya940, @Abhash297, @adm-unl, @baraline, @chrisholder, @SebastianSchmidl, @ermshaua, @futuer-szd, @hadifawaz1999, @itsdivya1309, @MatthewMiddlehurst, @Moonzyyy, @patrickzib, @TonyBagnall, @zy18811