CHANGELOG.md

0.29.0 (2024-11-26)

Features

• remove Image.__array__ (#964) (d2f59c9)
• rename count_row_if to count_rows_if (#960) (f5d6324)
• rename epoch_size to epoch_count (#962) (fa62e9b)
• rename flip_vertically and flip_horizontally methods (#946) (a3607d7), closes #945
• rename load_pretrained_model to from_pretrained_model (#957) (5b32acc)
• rename Row.get_value to Row.get_cell (#947) (e77cb6c)

Bug Fixes

• avoid performance warning when accessing column_count (#949) (c24b6d5)
• handle deprecations in third-party libraries (#963) (e10e6bb)
• index tensor on cpu (#961) (afafd43)

0.28.1 (2024-09-26)

Bug Fixes

• unwrap operators of cell operations (#932) (533a966)

0.28.0 (2024-09-17)

Features

• hyperparameter optimization for classical models (#843) (d8f7491), closes #264

0.27.0 (2024-07-19)

Features

• join (#870) (5764441), closes #745
• activation function for forward layer (#891) (5b5bb3f), closes #889
• add ImageDataset.split (#846) (3878751), closes #831
• add FunctionalTableTransformer (#901) (37905be), closes #858
• add InvalidFitDataError (#824) (487854c), closes #655
• add KNearestNeighborsImputer (#864) (fcdfecf), closes #743
• add moving average plot (#836) (abcf68a)
• add RobustScaler (#874) (62320a3), closes #650 #873
• add SequentialTableTransformer (#893) (e93299f), closes #802
• add temporal operations (#832) (06eab77)
• added 'histogram_2d' in TablePlotter (#903) (4e65ba9), closes #869 #798
• added from_str_to_temporal and continues prediction (#767) (35f468a), closes #806 #765 #740 #773
• added GRU layer (#845) (d33cb5d)
• Adds Dropout Layer (#868) (a76f0a1), closes #848
• dark mode for plots (#911) (5447551), closes #798
• easily create a baseline model (#811) (8e1b995), closes #710
• get first cell with value other than None (#904) (5a0cdb3), closes #799
• hyperparameter optimization for fnn models (#897) (c1f66e5), closes #861
• implement violin plots (#900) (9f5992a), closes #867
• plot decision tree (#876) (d3f81dc), closes #856
• prediction no longer takes a time series dataset only table (#838) (762e5c2), closes #837
• raise if remove_colums is called with unknown column by default (#852) (8f78163), closes #807
• regularization strength for logistic classifier (#866) (9f74e92), closes #750
• reorders parameters of RangeScaler and makes them keyword-only (#847) (2b82db7), closes #809
• replace seaborn with matplotlib for box_plot (#863) (4ef078e), closes #805 #849
• replaced seaborn with matplotlib for correlation_heatmap (#850) (d4680d4), closes #800 #849

Bug Fixes

• deps: bump urllib3 from 2.2.1 to 2.2.2 (#842) (b81bcd6), closes #3122 #3363 #3122 #3363 #3406 #3398 #3399 #3396 #3394 #3391 #3316 #3387 #3386
• labels of correlation heatmap (#894) (a88a609), closes #871
• make multi-processing in baseline models more consistent (#909) (fa24560), closes #907

Performance Improvements

• improved performance in various methods in Image and ImageList (#879) (134e7d8)

0.26.0 (2024-05-29)

Features

• Table.count_row_if (#788) (4137131), closes #786
• added method to load pretrained models from huggingface (#790) (dd8394b)
• infer input size of forward and LSTM layers (#808) (098a07f)
• outline around dots of scatterplot (#785) (ee8acf7)
• remove output conversions (#792) (46f2f5d), closes #732
• shorten some excessively long names (#787) (1c3ea59), closes #772
• specify column names in constructor of table transformers (#795) (69a780c)
• store window size and forecast horizon in dataset (#794) (f07bc5a)
• string operations on cells (#791) (4a17f76)

Bug Fixes

• handling of boolean columns in column statistics (#778) (f61cceb)
• sort x values of line plot (#782) (74d8649)

0.25.0 (2024-05-15)

Features

• major API redesign (WIP) (#752) (8e781f9), closes #694 #699 #714 #748
• move NN converters and layers to separate packages (#759) (c6a4073)
• remove operations without replacement from tabular containers (#747) (0e5a54b)
• specify partial order in label encoder (#763) (6fbe537), closes #639

Bug Fixes

• Conversion of tabular dataset to tensors (#757) (9e40b65)
• fixed devices with new polars implementation (#756) (e72339e)

Performance Improvements

• implement one hot encoder and imputer using polars (#768) (e993c17)

0.24.0 (2024-05-09)

Features

• Column.plot_histogram() using Table.plot_histograms for consistent results (#726) (576492c)
• Regressor.summarize_metrics and Classifier.summarize_metrics (#729) (1cc14b1), closes #713
• Table.keep_only_rows (#721) (923a6c2)
• Table.remove_rows (#720) (a1cdaef), closes #698
• Add ImageDataset and Layer for ConvolutionalNeuralNetworks (#645) (5b6d219), closes #579 #580 #581
• added load_percentage parameter to ImageList.from_files to load a subset of the given files (#739) (0564b52), closes #736
• added rnn layer and TimeSeries conversion (#615) (6cad203), closes #614 #648 #656 #601
• Basic implementation of cell with polars (#734) (004630b), closes #712
• deprecate Table.add_column and Table.add_row (#723) (5dd9d02), closes #722
• deprecated Table.from_excel_file and Table.to_excel_file (#728) (c89e0bf), closes #727
• Larger histogram plot if table only has one column (#716) (31ffd12)
• polars implementation of a column (#738) (732aa48), closes #712
• polars implementation of a row (#733) (ff627f6), closes #712
• polars implementation of table (#744) (fc49895), closes #638 #641 #649 #712
• regularization for decision trees and random forests (#730) (102de2d), closes #700
• Remove device information in image class (#735) (d783caa), closes #524
• return fitted transformer and transformed table from fit_and_transform (#724) (2960d35), closes #613

Bug Fixes

• make Image.clone internal (#725) (215a472), closes #626

Performance Improvements

• improved performance of TabularDataset.__eq__ by a factor of up to 2 (#697) (cd7f55b)

0.23.0 (2024-05-04)

Features

• add Column.to_table (#705) (36e4a7a), closes #695
• added Column.summarize_statistics() (#715) (71730a9), closes #701
• replace other values than NaN with imputer (#707) (4a059e0), closes #643

Bug Fixes

• use UTF-8 encoding when opening files (#704) (f8c27bc)

0.22.1 (2024-05-02)

Bug Fixes

• make _Layer, _InputConversion, _OutputConversion public (#692) (eb226f4), closes #690 #691

0.22.0 (2024-05-01)

Features

• is_fitted is now always a property (#662) (b1db881), closes #586
• add Column.missing_value_count (#682) (f084916), closes #642
• Add InputConversion & OutputConversion for nn interface (#625) (fd723f7), closes #621
• Add hash,eq and sizeof in ForwardLayer (#634) (72f7fde), closes #633
• allow using tables that already contain target for prediction (#687) (e9f1cfb), closes #636
• callback Row.sort_columns takes four parameters instead of two tuples (#683) (9c3e3de), closes #584
• rename group_rows_by in Table to group_rows (#661) (c1644b7), closes #611
• rename number_of_column in Row to number_of_columns (#660) (0a08296), closes #646
• rework TaggedTable (#680) (db2b613), closes #647
• show missing value count/ratio in summarized statistics (#684) (74b8a35), closes #619
• specify extras instead of features in to_tabular_dataset (#685) (841657f), closes #623

Bug Fixes

• actually use kernel of support vector machines for training (#681) (09c5082), closes #602

Performance Improvements

• Faster plot_histograms and more reliable plots (#659) (b5f0a12)

0.21.0 (2024-04-17)

Features

• add ARIMA model (#577) (8b9c7a9), closes #570
• Add ImageList class (#534) (3cb74a2), closes #528 #599 #600
• more hash, sizeof and eq implementations (#609) (2bc0b0a)

Performance Improvements

• Add special case to Table.add_rows to increase performance (#608) (ffb8304), closes #606
• improve performance of model & forward layer (#616) (e856cd5), closes #610
• lazily import our modules and external libraries (#624) (20fc313)
• treat Tables specially when calling add_rows (#606) (e555b85), closes #575

0.20.0 (2024-04-03)

Features

• add deterministic hash methods to all types (#573) (f6a3ca7)
• add fnn functionality (#529) (ce53153), closes #522
• add suffixes to models to indicate their task (#588) (d490dee)
• added lag_plot (#548) (0fb38d2), closes #519
• added normal plot for time series (#550) (dbdf11e), closes #549
• when using from table to time series feature must be given (#572) (ca23f0f), closes #571

Bug Fixes

• incorrect type hint for number_of_bins parameter (#567) (b434e53)
• mark various API elements as internal (#587) (ea176fc), closes #582 #585

0.19.0 (2024-02-06)

Features

• return the correct size for custom container objects (#547) (f44c34d)

0.18.0 (2024-02-03)

Features

• Add adjust_color_balance method in Image (#530) (dba23f9), closes #525
• Add find_edges method in Image (#531) (d728eb6), closes #523
• class for time series (#508) (73cdfb1), closes #481

0.17.1 (2024-01-11)

Bug Fixes

• update torch (#527) (6934d1d)

0.17.0 (2024-01-11)

Features

• change image class to use PyTorch tensors and methods (#506) (efa2b61), closes #505

0.16.0 (2023-11-22)

Features

• drop Python 3.10 and add Python 3.12 (#478) (5bf0e75)
• enable copy-on-write for pandas dataframes (#494) (9a19389), closes #428

Bug Fixes

• deps-dev: Bump urllib3 from 1.26.17 to 1.26.18 (#480) (a592d2c), closes #3159

Performance Improvements

• remove unneeded copy operations in table transformers (#496) (6443beb), closes #494

0.15.0 (2023-07-13)

Features

• Add copy method for tables (#405) (72e87f0), closes #275
• add gaussian noise to image (#430) (925a505), closes #381
• add schema conversions when adding new rows to a table and schema conversion when creating a new table (#432) (6e9ff69), closes #404 #322 #127 #322 #127
• add test for empty tables for the method Table.sort_rows (#431) (f94b768), closes #402
• added color adjustment feature (#409) (2cbee36), closes #380
• added test_repr table tests (#410) (cb77790), closes #349
• discretize table (#327) (5e3da8d), closes #143
• Improve error handling of TaggedTable (#450) (c5da544), closes #150
• Maintain tagging in methods inherited from Table class (#332) (bc73a6c), closes #58
• new error class OutOfBoundsError (#438) (1f37e4a), closes #262
• rename several Table methods for consistency (#445) (9954986), closes #439
• suggest similar columns if column gets accessed that doesnt exist (#385) (6a097a4), closes #203

Bug Fixes

• added the missing ids in parameterized tests (#412) (dab6419), closes #362
• don't warn if Imputer transforms column without missing values (#448) (f0cb6a5)
• Warnings raised by underlying seaborn and numpy libraries (#425) (c4143af), closes #357

0.14.0 (2023-06-30)

Features

• Add find_edges method to Image class (#383) (d14b6ce), closes #288
• Add StandardScaler transformer (#316) (57b0572), closes #142
• Add docstrings to the getter methods for hyperparameters in Regression and Classification models (#371) (9073f04), closes #313
• Added Table.group_by to group a table by a given key (#343) (afb98be), closes #160
• Added and improved errors and warnings in the table transformers (#372) (544e307), closes #152
• added crop() method in image and tests (#365) (eba8163)
• added invert_colors method (#367) (1e4d110)
• adjust brightness and contrast of image (#368) (1752feb), closes #289 #291
• blur Image method (#363) (c642176)
• check that methods of table can handle an empty table (#314) (686c2e7), closes #123
• convert image to grayscale (#366) (1312fe7), closes #287
• enhance replace_column to accept a list of new columns (#312) (d50c5b5), closes #301
• Explicitly throw UnknownColumnNameError in TaggedTable._from_table (#334) (498999f), closes #333
• flip images / eq method for image (#360) (54f4ae1), closes #280
• improve table.summary. Catch ValueError thrown by column.stability (#390) (dbbe0e3), closes #320
• improve error handling of column.stability when given a column that contains only None (#388) (1da2499), closes #319
• Improve Error Handling of classifiers and regressors (#355) (66f5f64), closes #153
• properties for width-height of image (#359) (d9ebdc1), closes #290
• Resize image (#354) (3a971ca), closes #283
• rotate_left and rotate_right added to Image (#361) (c877530), closes #281
• set kernel of support vector machine (#350) (1326f40), closes #172
• sharpen image (#364) (3444700), closes #286

Bug Fixes

• Keeping no columns with Table.keep_only_columns results in an empty Table with a row count above 0 (#386) (15dab06), closes #318
• remove default value of positive_class parameter of classifier metrics (#382) (58fc09e)
• remove default value of radius parameter of blur (#378) (7f07f29)

0.13.0 (2023-06-01)

Features

• add Choice class for possible values of hyperparameter (#325) (d511c3e), closes #264
• Add RangeScaler transformer (#310) (f687840), closes #141
• Add methods that tell which columns would be affected by a transformer (#304) (3933b45), closes #190
• Getters for hyperparameters of Regression and Classification models (#306) (5c7a662), closes #260
• improve error handling of table (#308) (ef87cc4), closes #147
• Remove warnings thrown in new Transformer methods (#324) (ca046c4), closes #323

0.12.0 (2023-05-11)

Features

• add learning_rate to AdaBoost classifier and regressor. (#251) (7f74440), closes #167
• add alpha parameter to lasso_regression (#232) (b5050b9), closes #163
• add parameter lasso_ratio to ElasticNetRegression (#237) (4a1a736), closes #166
• Add parameter number_of_tree to RandomForest classifier and regressor (#230) (414336a), closes #161
• Added Table.plot_boxplots to plot a boxplot for each numerical column in the table (#254) (0203a0c), closes #156 #239
• Added Table.plot_histograms to plot a histogram for each column in the table (#252) (e27d410), closes #157
• Added Table.transform_table method which returns the transformed Table (#229) (0a9ce72), closes #110
• Added alpha parameter to RidgeRegression (#231) (1ddc948), closes #164
• Added Column#transform (#270) (40fb756), closes #255
• Added method Table.inverse_transform_table which returns the original table (#227) (846bf23), closes #111
• Added parameter c to SupportVectorMachines (#267) (a88eb8b), closes #169
• Added parameter maximum_number_of_learner and learner to AdaBoost (#269) (bb5a07e), closes #171 #173
• Added parameter number_of_trees to GradientBoosting (#268) (766f2ff), closes #170
• Allow arguments of type pathlib.Path for file I/O methods (#228) (2b58c82), closes #146
• convert Schema to dict and format it nicely in a notebook (#244) (ad1cac5), closes #151
• Convert between Excel file and Table (#233) (0d7a998), closes #138 #139
• convert containers for tabular data to HTML (#243) (683c279), closes #140
• make Column a subclass of Sequence (#245) (a35b943)
• mark optional hyperparameters as keyword only (#296) (44a41eb), closes #278
• move exceptions back to common package (#295) (a91172c), closes #177 #262
• precision metric for classification (#272) (5adadad), closes #185
• Raise error if an untagged table is used instead of a TaggedTable (#234) (8eea3dd), closes #192
• recall and F1-score metrics for classification (#277) (2cf93cc), closes #187 #186
• replace prefix n with number_of (#250) (f4f44a6), closes #171
• set alpha parameter for regularization of ElasticNetRegression (#238) (e642d1d), closes #165
• Set column_names in fit methods of table transformers to be required (#225) (2856296), closes #179
• set learning rate of Gradient Boosting models (#253) (9ffaf55), closes #168
• Support vector machine for regression and for classification (#236) (7f6c3bd), closes #154
• usable constructor for Table (#294) (56a1fc4), closes #266
• usable constructor for TaggedTable (#299) (01c3ad9), closes #293

Bug Fixes

• OneHotEncoder no longer creates duplicate column names (#271) (f604666), closes #201
• selectively ignore one warning instead of all warnings (#235) (3aad07d)

0.11.0 (2023-04-21)

Features

• OneHotEncoder.inverse_transform now maintains the column order from the original table (#195) (3ec0041), closes #109
• add plot_ prefix back to plotting methods (#212) (e50c3b0), closes #211
• adjust Column, Schema and Table to changes in Row (#216) (ca3eebb)
• back Row by a polars.DataFrame (#214) (62ca34d), closes #196 #149
• clean up Row class (#215) (b12fc68)
• convert between Row and dict (#206) (e98b653), closes #204
• convert between a dict and a Table (#198) (2a5089e), closes #197
• create column types for polars data types (#208) (e18b362), closes #196
• dataframe interchange protocol (#200) (bea976a), closes #199
• move existing ML solutions into safeds.ml.classical package (#213) (655f07f), closes #210

Bug Fixes

• table.keep_only_columns now maps column names to correct data (#194) (459ab75), closes #115
• typo in type hint (#184) (e79727d), closes #180

0.10.0 (2023-04-13)

Features

• move exceptions into subpackages (#177) (10b17fd)

0.9.0 (2023-04-04)

Features

• container for images (#159) (ed7ae34), closes #158
• improve error handling for predict (#145) (a5ff11c), closes #9
• move ImputerStrategy to safeds.data.tabular.typing (#174) (205c8e2)
• rename n_neighbors to number_of_neighbors (#162) (526b96e)

Bug Fixes

• export TableTransformer and InvertibleTableTransformer (#135) (81c3695)

0.8.0 (2023-03-31)

Features

• create empty Table without schema (#128) (ddd3f59), closes #127
• improve ColumnTypes (#132) (1786a87), closes #113
• infer schema of row if not passed explicitly (#134) (c5869bb), closes #15
• new method is_fitted to check whether a model is fitted (#130) (8e1c3ea)
• new method is_fitted to check whether a transformer is fitted (#131) (e20954f)
• rename drop_XY methods of Table to remove_XY (#122) (98d76a4)
• rename fit_transform to fit_and_transform (#119) (76a7112), closes #112
• rename shuffle to shuffle_rows (#125) (ea21928)
• rename slice to slice_rows (#126) (20d21c2)
• rename TableSchema to Schema (#133) (1419d25)

0.7.0 (2023-03-29)

Features

• sort_rows of a Table (#104) (20aaf5e), closes #14
• add _file suffix to methods interacting with files (#103) (ec011e4)
• improve transformers for tabular data (#108) (b18a06d), closes #61 #90
• remove OrdinalEncoder (#107) (b92bba5), closes #61
• specify features and target when creating a TaggedTable (#114) (95e1fc7), closes #27
• swap name and data parameters of Column (#105) (c2f8da5)

0.6.0 (2023-03-27)

Features

• allow calling correlation_heatmap with non-numerical columns (#92) (b960214), closes #89
• function to drop columns with non-numerical values from Table (#96) (8f14d65), closes #13
• function to drop columns/rows with missing values (#97) (05d771c), closes #10
• remove list_columns_with_XY methods from Table (#100) (a0c56ad), closes #94
• rename keep_columns to keep_only_columns (#99) (de42169)
• rename remove_outliers to drop_rows_with_outliers (#95) (7bad2e3), closes #93
• return new model when calling fit (#91) (165c97c), closes #69

Bug Fixes

• handling of missing values when dropping rows with outliers (#101) (0a5e853), closes #7

0.5.0 (2023-03-26)

Features

• move plotting methods into Column and Table classes (#88) (5ec6189), closes #62

0.4.0 (2023-03-26)

Features

• better names for properties of TaggedTable (#74) (fee398b)
• change the name of a Column (#76) (ec539eb)
• metrics as methods of models (#77) (bc63693), closes #64
• optionally pass type to column (#79) (64aa429), closes #78
• remove target_name parameter of predict (#70) (b513454)
• rename tagged_table parameter of fit to training_set (#71) (8655521)
• return TaggedTable from predict (#73) (5d5f5a6)

0.3.0 (2023-03-24)

Features

• make Column and Row iterable (#55) (74eea1f), closes #47

Bug Fixes

• "UserWarning: X has feature names" when predicting (#53) (74b0753), closes #51