v0.7.0

February 2024

Highlights

• Implementations of the SASTTransformer and SASTClassifier (@frankl1)

• fit_predict/fit_predict_proba methods in BaseClassifier for (sensibly) estimating accuracy on train data (@MatthewMiddlehurst)

• ResNet based auto-encoder to deep learning clustering module (@xiaopu222)

• Introducing BaseSegmenter for segmentation module (@TonyBagnall)

• Python 3.12 is now available!

• Various additions, documentation updates, and implementations for the visualisation module

• Many deprecations of old and unmaintained functionality (if you wish to help maintaining these, please get in touch!)

Benchmarking

Bug Fixes

• [BUG] Allow upper case cluster aliases (#1008) @TonyBagnall

Enhancements

• [ENH] Improvements to scatterplot when lower is better (#1010) @dguijo

• [ENH] Load the regression results from tsc.com (#1034) @TonyBagnall

• [ENH] speed up testing (#1096) @TonyBagnall

• [ENH] Data and results loading update (#1115) @TonyBagnall

• [ENH] Remove name RSTSF (#1128) @TonyBagnall

• [ENH] New plot for significance relationships (#954) @TonyBagnall

Other

• [ENH] Improve Pairwise-Visualizations in plot_scatter() (#1021) @patrickzib

Classification

Bug Fixes

• [ENH][BUG] Change predict_proba of WEASEL_v2 to the default behavior, as in RocketClassifier (#1022) @patrickzib

• [BUG] Fixed missing z-normalisation for SFA in RED CoMETS (#1057) @zy18811

• [BUG] remove base classifier shortcut and new test (#741) @hedeershowk

• [BUG] Wrong bound on sklearn for sparse argument (#1112) @julu98

Deprecation

• [DEP] Deprecate MatrixProfileClassifier (#1082) @MatthewMiddlehurst

• [ENH] Adds missing deprecation for parameter support_probabilities (#1028) @patrickzib

Documentation

• [DOC] Updated paper reference for RED CoMETS (#1058) @zy18811

Enhancements

• [ENH] ADD SAST transformer and SASTClassifier (#958) @frankl1

• [ENH] Test code for code in test directories (#989) @TonyBagnall

• [ENH] Reorganise mock estimators in utils (#1045) @TonyBagnall

• [ENH] The Default value for use_mini_batch_size should be set to False (#1065) @andregdmitri

• [ENH] Adds missing deprecation for parameter support_probabilities (#1028) @patrickzib

• [ENH] Classification _fit_predict mockup (#1107) @MatthewMiddlehurst

• [ENH] Remove conversion decorators from SignaturesClassifier (#1101) @TonyBagnall

• [ENH] Base classifier class docstrings (#1131) @TonyBagnall

• [ENH] Classifier pipeline rework, remove operator overload and delegate class (#401) @MatthewMiddlehurst

• [ENH] BaseClassifier testing for fit_predict/fit_predict_proba and classifier conversions to use said methods (#1116) @MatthewMiddlehurst

Maintenance

• [MNT] Fixes deprecation warnings for save_transformed_data in interval based classifiers (#1145) @TonyBagnall

Refactored

• [REF] Fixed the default use_mini_batch_size value in resnet (#1073) @jose-gilberto

Clustering

Bug Fixes

• [BUG] Allow upper case cluster aliases (#1008) @TonyBagnall

Documentation

• [DOC, REF] Add TimeSeriesKernelKMeans and TimeSeriesKShapes to clustering/__init__.py and also adds examples to dosctring (#1097) @andregdmitri

Enhancements

• [ENH] Clustering testing (#1003) @TonyBagnall

• [ENH] Introducing visualisation module (#1002) @TonyBagnall

• [ENH] Add clustering_accuracy_score to metrics (#1048) @MatthewMiddlehurst

• [ENH] speed up testing (#1096) @TonyBagnall

• [ENH] Add ResNet based auto-encoder to deep clustering module (#1124) @xiaopu222

Refactored

• [DOC, REF] Add TimeSeriesKernelKMeans and TimeSeriesKShapes to clustering/__init__.py and also adds examples to dosctring (#1097) @andregdmitri

Datasets

Enhancements

• [ENH] New regression problems (#1006) @TonyBagnall

• [ENH] Retrieve metadata for datasets from tsc.com (#1025) @TonyBagnall

• [ENH] Centralise testing data generation (#1105) @TonyBagnall

• [ENH] .arff writer function (#1023) @MatthewMiddlehurst

• [ENH] Data and results loading update (#1115) @TonyBagnall

Maintenance

• [MNT] remove distutils (#1084) @TonyBagnall

Other

• [BUG] Fix overnight loading test (#1127) @TonyBagnall

Distances

Enhancements

• [ENH] speed up testing (#1096) @TonyBagnall

Forecasting

Deprecation

• [ENH] Deprecate functors from performance metrics and replace with function calls (#1081) @TonyBagnall

Enhancements

• [ENH] Reorganise mock estimators in utils (#1045) @TonyBagnall

• [ENH] Exclude MockUnivariateForecasterLogger from test_non_state_changing_method_contract (#1068) @TonyBagnall

• [ENH] Tidy up the utils directory (#1042) @TonyBagnall

• [ENH] Remove tensor flow probability, replace with scipy (#1087) @TonyBagnall

• [ENH] Deprecate functors from performance metrics and replace with function calls (#1081) @TonyBagnall

Networks

Enhancements

• [ENH] Some fixes network module (#1122) @hadifawaz1999

• [ENH] Add ResNet based auto-encoder to deep clustering module (#1124) @xiaopu222

Regression

Enhancements

• [ENH] New regression problems (#1006) @TonyBagnall

• [ENH] The Default value for use_mini_batch_size should be set to False (#1065) @andregdmitri

Refactored

• [REF] Fixed the default use_mini_batch_size value in resnet (#1073) @jose-gilberto

Segmentation

Enhancements

• [ENH] Introducing BaseSegmenter for segmentation module (#946) @TonyBagnall

Maintenance

• [MNT] Add upper bounds for main dependencies and replace attrs (#1106) @MatthewMiddlehurst

Similarity Search

Bug Fixes

• [ENH, BUG, DOC] Refactor distance profiles, introduce STD threshold for numba (#957) @baraline

Documentation

• [ENH, BUG, DOC] Refactor distance profiles, introduce STD threshold for numba (#957) @baraline

• [ENH, DOC] Add Euclidean distance with convolution for similarity search (#1009) @baraline

Enhancements

• [ENH, BUG, DOC] Refactor distance profiles, introduce STD threshold for numba (#957) @baraline

• [ENH, DOC] Add Euclidean distance with convolution for similarity search (#1009) @baraline

Transformations

Bug Fixes

• [ENH, BUG, DOC] Refactor distance profiles, introduce STD threshold for numba (#957) @baraline

• [BUG] Collection transformer wrapper bugfixing and PlateauFinder refactor (#804) @MatthewMiddlehurst

Documentation

• [ENH, BUG, DOC] Refactor distance profiles, introduce STD threshold for numba (#957) @baraline

• [ENH] Removes a whole load of references to “scitype” and “mtype” (#1035) @TonyBagnall

Enhancements

• [ENH] remove unnecessary call to check_X in SFA and SFA_Fast (#994) @TonyBagnall

• [ENH] ADD SAST transformer and SASTClassifier (#958) @frankl1

• [ENH, BUG, DOC] Refactor distance profiles, introduce STD threshold for numba (#957) @baraline

• [ENH] Introducing visualisation module (#1002) @TonyBagnall

• [ENH] Removes a whole load of references to “scitype” and “mtype” (#1035) @TonyBagnall

• [ENH] Deprecate Filter transformer (#1110) @TonyBagnall

• [ENH] Remove conversion decorators from SignaturesClassifier (#1101) @TonyBagnall

• [ENH] New base classes: base series estimator , segmenters and base series transformers (#996) @TonyBagnall

Maintenance

• [MNT] remove is_numeric (#1083) @TonyBagnall

• [MNT] remove distutils (#1084) @TonyBagnall

Unit Testing

Bug Fixes

• [BUG] Correct imports for visualisation testing (#1090) @MatthewMiddlehurst

Enhancements

• [ENH] Clustering testing (#1003) @TonyBagnall

• [ENH] Reorganise mock estimators in utils (#1045) @TonyBagnall

• [ENH] Exclude MockUnivariateForecasterLogger from test_non_state_changing_method_contract (#1068) @TonyBagnall

• [ENH] Testing and basic notebooks for series and learning task specific plotting functions (#1051) @MatthewMiddlehurst

• [ENH] Testing for estimator specific plotting functions (#1071) @MatthewMiddlehurst

• [ENH] Centralise testing data generation (#1105) @TonyBagnall

• [ENH] Testing and changes for results plotting functions (#1092) @MatthewMiddlehurst

• [ENH] BaseClassifier testing for fit_predict/fit_predict_proba and classifier conversions to use said methods (#1116) @MatthewMiddlehurst

• [ENH] Tidy up data_gen (#1129) @TonyBagnall

Maintenance

• [MNT] Add pytest plugin to retry failed tests (#1043) @MatthewMiddlehurst

• [MNT] Legacy deprecation (#1114) @TonyBagnall

Refactored

• [REF] Rename tests to testing and move utilities for generating expected results to the folder. (#1041) @MatthewMiddlehurst

• [REF] Refactor utils._testing to testing.utils (#1047) @MatthewMiddlehurst

Visualisations

Bug Fixes

• [BUG] Correct imports for visualisation testing (#1090) @MatthewMiddlehurst

Deprecation

• [MNT] Deprecate old plot_series functions and policy clarification (#1054) @MatthewMiddlehurst

Enhancements

• [ENH] scatter plots enhancement. (#1033) @dguijo

• [ENH,REF] Refactor visualisation and add temporal importance curves function (#1050) @MatthewMiddlehurst

• [ENH] Testing and basic notebooks for series and learning task specific plotting functions (#1051) @MatthewMiddlehurst

• [ENH] Testing for estimator specific plotting functions (#1071) @MatthewMiddlehurst

• [ENH] Testing and changes for results plotting functions (#1092) @MatthewMiddlehurst

• [ENH, MNT] Changes to docs in significance plots and stats. (#1141) @dguijo

Maintenance

• [MNT] Deprecate old plot_series functions and policy clarification (#1054) @MatthewMiddlehurst

• [ENH, MNT] Changes to docs in significance plots and stats. (#1141) @dguijo

Refactored

• [ENH,REF] Refactor visualisation and add temporal importance curves function (#1050) @MatthewMiddlehurst

Other

Bug Fixes

• [BUG] fix tag errors, refactor comments np.array -> np.ndarray (#1014) @TonyBagnall

• [MNT, BUG] Modify njit func generator to support python functions (#1044) @baraline

Deprecation

• [MNT] Deprecate mlflow functionality (#968) @MatthewMiddlehurst

Documentation

• [DOC] remove plot_critical_difference from benchmarking API (#1040) @TonyBagnall

• [DOC,MNT] Tidy up GitHub txt links and update dockerfile version (#1075) @MatthewMiddlehurst

• [DOC] Update index page and images (#1080) @MatthewMiddlehurst

• [DOC] aeon mentoring/project webpage (#1085) @baraline

• [DOC] API tidy (#1140) @TonyBagnall

• [DOC] Tidy up developers guide (#1139) @TonyBagnall

Enhancements

• [ENH] purge todo (#995) @TonyBagnall

• [ENH] More purging of datatypes (#993) @TonyBagnall

• [ENH] Purge datatypes from utils (#1015) @TonyBagnall

• [ENH] Remove the PODS adaptor (#1038) @TonyBagnall

• [ENH] Update scikit-learn version (#1093) @MatthewMiddlehurst

• [MNT] Allow Python 3.12 (#861) @MatthewMiddlehurst

• [DOC] Tidy up developers guide (#1139) @TonyBagnall

• [ENH] deprecates conversion functions in datatypes (#1143) @TonyBagnall

• [ENH] Split validation and conversion in utils (#1134) @TonyBagnall

Maintenance

• [MNT] Release workflow typo fix (#991) @MatthewMiddlehurst

• [MNT] Remove ! marks (#1027) @TonyBagnall

• [MNT, BUG] Modify njit func generator to support python functions (#1044) @baraline

• [MNT] Set upper bound for tensorflow (#1070) @MatthewMiddlehurst

• [DOC,MNT] Tidy up GitHub txt links and update dockerfile version (#1075) @MatthewMiddlehurst

• [MNT] Stop the bot from commenting on other bots PRs (#1079) @MatthewMiddlehurst

• [MNT] Add pyupgrade hook (#1056) @MatthewMiddlehurst

• [MNT] Update numba bound (#1133) @MatthewMiddlehurst

• [MNT] Disable sweep and other small maintenance bits (#1132) @MatthewMiddlehurst

• [MNT] Allow Python 3.12 (#861) @MatthewMiddlehurst

• [MNT] Deprecate mlflow functionality (#968) @MatthewMiddlehurst

• [MNT] Further isolation of datatypes (#1117) @TonyBagnall

Contributors

The following have contributed to this release through a collective 91 GitHub Pull Requests:

@andregdmitri, @baraline, @dguijo, @frankl1, @hadifawaz1999, @hedeershowk, @jose-gilberto, @julu98, @MatthewMiddlehurst, @patrickzib, @TonyBagnall, @xiaopu222, @zy18811