Changelog

Version 2.22.241011

Bug Fixes

• Fixed get_best_pipeline function sorting issue in massive AutoML.

• Fixed error message display issue in massive AutoML.

• Fixed the key display issue in new version of dataset report.

• Fixed importing issue in AutoML progress monitor and configuration.

Version 2.22.240917

New Functions

• Added Outlier Detection for regression OutlierDetectionRegression.

• Added Text Classification with model TextClassificationWithModel.

• Added Multi-task MLP Classification model MLPMultiTaskClassifier.

• Added Multi-task MLP Regression model MLPMultiTaskRegressor.

• Added Model Card creation fucntion create_model_card().

• Added Model Card parser parse_model_card().

• Added to_sqlalchemy function in ConnectionContext.

• Added Massive AutoML model MassiveAutomaticClassification, MassiveAutomaticRegression and MassiveAutomaticTimeSeries.

Enhancements

• Enhanced dataset report with new style.

• Enhanced model storage with model card integration.

• Enhanced make_future_dataframe with increment_type option.

• Enhanced hana_scheduler to support the schema of the output tables.

• Enhanced json pipeline input in Pipeline Module.

Bug Fixed

• Fixed numpy 2.0 incompatibility issue.

Version 2.21.240909

Bug Fixes

• Fixed the group id INT issue with special group id.

• Fixed display issue in dataset report when y-axis starts from 0.

• Fixed the massive model issue with empty params.

• Fixed cross join issue in dataframe function.

Version 2.21.240726

Bug Fixes

• Fixed the display order of hana_scheduler job.

• Fixed the missing schema support of output tables in hana_scheduler.

• Fixed the missing meta table issue in model storage import_model function.

• Fixed the timezone warning in hana_scheduler.

• Improved the Weibull distribution fitting with the location parameter.

Version 2.21.240712

Bug Fixes

• Fixed the AutoML config issue related to different runtime_platform.

• Fixed the model storage issue due to missing SQL training script.

• Fixed the AutoML Time Series parameter registration issue.

Version 2.21.240628

Bug Fixes

• Fixed melt function with multiple keys.

• Fixed dtypes issue to support REAL_VECTOR for hdbcli 2.21.

Version 2.21.240624

New Functions

• Added Hull-White Simulation function hull_white_simulate().

• Added Finetune Best Pipeline in AutoML AutomaticClassification and AutomaticRegression.

• Added Massive Text Mining in methods get_related_doc(), get_related_term(), get_relevant_doc(), get_relevant_term(), get_suggested_term().

• Added Benford Analysis benford_analysis().

• Added AutoML config Dict Visualizer AutoMLConfig.

• Added text classification with model TextClassificationWithModel.

Enhancements

• Enhanced DataFrame functions to support real vector type.

• Optimize default value for AutoML.

• Enhanced AutoML connections visualization with zoom in/out.

• Enhanced pipeline module with explainability.

• Enable get_permutation_importance() method for time series in ARIMA, AutoARIMA, AdditiveModelForecast, LTSF and BSTS

Bug Fixes

• Fixed SHAP display issues in RDT/DT/HGBT.

• Fixed EDA display issues.

Version 2.20.240426

Bug Fixes

• Fixed the placement issue of confusion matrix in model report.

• Fixed the shapely dependency installation issue.

• Fixed the concat function issue in hana_graph.

• Fixed the wrong thread_ratio parameter in LinkPredict.

• Fixed the pipeline module in model storage.

Version 2.20.240319

New Functions

• Added Bubble Plot hana_ml.visualizers.eda.bubble_plot() and Parallel Co-ordinate Plot hana_ml.visualizers.eda.parallel_coordinates() in EDA.

• Added time series permutation feature importance permutation_importance().

• Added MLP Recommender hana_ml.algorithms.pal.recommender.MLPRecommender.

Enhancements

• Enhanced progress monitor for AutoML to display at anytime. Especially for the scheduled job.

• Support algorithm-specific parameters in automl/pipeline predict. Effect in both pipline and automl module.

• Enhanced AutoML Auto SQL Content Integration for progress log management.

• Enhanced pipeline report to display connection scores.

• Enhanced AutoML config_dict templates with new operators.

• Enhanced AutoML with connection constraints option.

• Enhanced AUtoML with different logging levels.

• Enhanced AutoML with random/grid search option.

• Enhanced outlier_detection with voting logic.

• Enhanced online BCPD in massive mode.

Version 2.19.240221

Bug Fixes

• Fixed incorrect shap plot.

• Fixed incorrect dependence plot.

• Fixed wrong SQL generation result in fairml.

Version 2.19.240131

Default Value Changes

• Changed the default value of enable_plotly to True in all visualization module.

• Removed the matplotlib dependency.

Version 2.19.240124

Bug Fixes

• Fixed missing fairml metrics in binary_classification_debriefing.

• Fixed missing parameter issues in fairml.

Version 2.19.240115

Bug Fixes

• Fixed model_debriefing issue for RDT and HGBT.

Version 2.19.240104

Bug Fixes

• Fixed fairml prediction issue with fixed uuid.

• Fixed fairml issue when submodel is used.

• Fixed box_plot issue when plotly is enabled abd groupby is None.

• Fixed unified regression issue when key is not specified.

• Fixed massive IsolationForest without key issue.

• Fixed variable importance issue in model report.

• Fixed model storage issue when odbc is used.

Version 2.19.231207

New Functions

• Added FairMLClassification and FairMLRegression to provide fairness in machine learning.

• Added auto_cast() in dataframe module.

• Added description and description_ext property of dataframe.

• Added QuantileTransform.

• Added summary().

• Added score function of AutoML module.

• Added fit(), transform() and score() functions in Pipeline module.

Enhancements

• Support STL decomposition method in seasonal_decompose() by offering new parameters decompose_method, stl_robust and stl_seasonal_average.

• Provide explainability support in LTSF.

• Offer ignore zero calculation support in UnifiedRegression when calculating MPE or MAPE.

• Provide top N verbose classes in predict () of UnifiedClassification by offering a new parameter verbose_top_n.

• Support Stock keeping oriented Prediction Error Costs (SPEC) in AutomaticTimeSeries.

• Enhanced AutoML logging with Auto SQL Content.

• Enhanced operation, mutate and filter in DataFrame.

• Enhanced reset_config_dict by allowing the custom config_dict.

• Boost dataset report with new PAL describe function.

• Enhanced proximal gradient support in HGBT including unified classification.

• Enhanced tf_analysis with enable_stopwords and keep_numeric parameters.

• Enhanced indicator for model deletion in model storage.

• Enhanced OneHotEncoding with onehot_min_frequency and onehot_max_categories parameters.

• Enhanced AutoML with connection optimization.

• Enhanced model list in model storage by adding partial info display option.

• Enhanced AutoML progress monitor with evaluating tab.

API Changes

• Added parameters decompose_method, stl_robust, stl_seasonal_average and smooth_method_non_seasonal in seasonal_decompose().

• Added parameters show_explainer and reference_dict to provide explainability support in LTSF.

• Added a parameter ignore_zero to offer ignore zero calculation support in UnifiedRegression when calculating MPE or MAPE.

• Added a parameter verbose_top_n in predict () of UnifiedClassification to present top N verbose classes.

Version 2.18.231114

Bug Fixes

• Fixed pipeline score issue.

• Fixed model table error in automl fit with reason code.

Version 2.18.231103

Bug Fixes

• Fixed key issue in ARIMA explainer.

• Fixed hana-ml parameter registration issue.

• Fixed PCA issue in pipeline fit.

• Fixed missing attributes issue in time series SHAP.

• Fixed NULL value issue in time series report.

• Fixed NULL value issue in SHAP visualizer.

• Fixed describe function issue with duplicate "unique" column.

• Fixed key issue in timeseries_box_plot when plotly is enabled.

• Fixed incompatibility issue with pandas 2.0.

• Fixed dataframe issue when hint clause is used.

• Fixed csrf token issue in amdp deployer.

Version 2.18.230927

Bug Fixes

• Fixed cancellation issue in AutoML progress monitor.

• Fixed log cleanup issue in AutoML.

• Fixed usage of concat in Graph.describe().

• Fixed temporary table issue by replacing with table variable.

• Fixed missing parameter issue in HGBT regression.

• Fixed syntax error in louvain.

Version 2.18.230914

New Functions

• Added make_future_dataframe().

• Added interval_quality().

• Added drop_view().

• Added hana_ml.graph.algorithms.CommunitiesLouvain.

Enhancements

• Enhanced the support of Portuguese in tf_analysis(), text_classification(), get_related_doc(), get_related_term(), get_relevant_doc(), get_relevant_term(), get_suggested_term functions().

• Enhanced the support of different type of network like NLinear, DLinear, XLinear, SCINet in LTSF.

• Support the massive mode of accuracy_measure().

• Support the massive mode of IsolationForest().

• Enhanced model storage method list_models with display option.

• Enhanced model selection with range support.

• Enhanced ShapleyExplainer with dependence plot.

• Enhanced AutoML with explanation visualization.

• Enhanced UnifiedRegression with predict interval and visualization.

• Enhanced UnifiedClassification with feature importance support.

• Enhanced OutlierDetectionTS with 'auto' mode.

• Simplified the AutoML fit with background_size.

• Enhanced AutoML and Pipeline modules with score function.

• Enhanced progress monitor in AutoML with the evaluating tab.

API Changes

• Added a parameter called network_type in LTSF for network selection.

• Enhanced HANA scheduler by removing manual parameters input.

Bug Fixes

• Fixed scatter plot error from ax.scatter c to cmap.

• Fixed BAS incompatibility issue.

• Fixed time diff error when creating new timeframe.

• Fixed date type issue in dataset report.

• Fixed time series report issue in changepoints_item.

Version 2.17.230808

Bug Fixes

• Fixed default cmap value in scatter_plot().

• Fixed display error for monthly data in tsa functions.

• Fixed index display error in seasonal_decompose().

• Fixed TimeSeriesReport index sort issue.

Version 2.17.230727

Bug Fixes

• Fixed Decimal issue in Explainer item and other related items in TimeSeriesReport.

• Fixed shadow option in EDA pie_plot().

• Fixed enable_stopwords issue in tf_analysis(). The same issue in wordcloud plot.

• Fixed year legend sort issue in TimeSeriesReport.

Version 2.17.230714

Bug Fixes

• Fixed wrong error message in HANAScheduler.

• Fixed corr issue that the column misses quotes.

• Fixed front-end connection reset issue in AutoML to avoid too many query from progress table.

• Fixed cron missing issue by adding NULL check.

• Fixed the Decimal issue in TimeSeriesReport.

• Fixed the x-axis order issue in TimeSeriesReport.

Version 2.17.230628

Bug Fixes

• Fixed CAP generation issues for APL.

• Fixed duplicated prefix for predict artifact in CAP generation.

• Fixed parameter checking for APL.

Version 2.17.230622

New Functions

• Added set_scale_out() to enable APL functions execution in scaling out environment.

• Added OnlineBCPD.

• Added HANAGeneratorForCAP.

• Added PowerTransform.

• Added HANAScheduler.

Enhancements

• Enhanced the support of plotly for eda functions like quarter_plot(), seasonal_plot() ...

• Enhanced the support of spectral clustering in UnifiedClustering

• Enhanced HANA artifacts generation for pipeline module.

• Enhanced AutoML with reason code option.

• Enhanced TimeSeriesReport with confidence interval.

• Enhanced RDT with prediction interval in UnifiedRegression.

• Enhanced ModelStorage with server-side scheduler.

• Enhanced unified API for pivoted input data.

• Enhanced diff() to support datetime column.

Version 2.16.230601

Bug Fixes

• Fixed model load issue for Pipeline module in model storage.

• Fixed parameter missing in HybridGradientBoostingClassifier and HybridGradientBoostingRegressor.

Version 2.16.230526

Bug Fixes

• Fixed pipeline missing evaluation function.

• Fixed tips and chart width for model report.

• Fixed built-in operation missing in pipeline module.

• Fixed WordCloud issues to disable stopwords.

Version 2.16.230519

Bug Fixes

• Fixed AutomaticTimeSeries config_dict template.

• Fixed progress logging in auto-ml module.

• Fixed the progress monitor GeneralProgressStatusMonitor issue when early_stop is enabled.

• Fixed KNN NaN issue due to the pandas new changes.

• Fixed describe() function to support SMALLINT.

Version 2.16.230508

Bug Fixes

• Fixed pipeline module for model storage.

• Fixed stuck in progress monitor when progress table is not empty.

• Fixed quote issue in serializing pipeline object.

• Fixed parameter missing in FACCM.

• Fixed dependency issue for pydotplus.

• Fixed tail() function with default rel_col.

• Fixed NaN in KNN optimal parameter collect.

Version 2.16.230413

Bug Fixes

• Fixed unsupported issue of precomputed distance matrix in predict() of UnifiedClustering

• Fixed wrong error message when PAL functions are executed in HANA SPS07.

• Fixed no timestamp support in plot_time_series_outlier().

• Fixed timestamp error in progress monitor in AutomaticClassification, AutomaticRegression, AutomaticTimeSeries.

• Fixed polynomial feature generation category undefined issue.

Version 2.16.230323

Bug Fixes

• Fixed wrong error message when "hint" has been used.

• Fixed load_model issue for APL model.

Version 2.16.230316

New Functions

• Added model report items to TimeSeriesReport.

• Added time-series report to UnifiedReport.

• Added TimeSeriesClassification.

• Added TUDF code generation function.

• Added AMDP generator create_amdp_class() for Pipeline.

• Added import_csv_from() function for importing csv file from the cloud storage locations like Azure, Amazon(AWS) Google Cloud, SAP HANA Cloud, Data Lake Files(HDLFS).

• Added set_scale_out() to enable APL functions execution in scaling out environment.

Enhancements

• Enhanced model report with new framework.

• Enhanced LinearRegression with prediction/confidence interval.

• Enhanced accuracy_measure() with SPEC measure.

• Enhanced UnifiedExponentialSmoothing with reason code control parameter.

• Enhanced precalculated distances matrix input for KMEDOIDS in UnifiedClustering.

• Enhanced AutomaticClassification and AutomaticRegression with successive halving.

API Changes

• Added a parameter called decom_state in UnifiedExponentialSmoothing for the control of reason code display.

• Added a parameter called lang in WordCloud for language selection.

Version 2.15.230217

Bug Fixes

• Fixed cmap issues in eda visualizer.

• Fixed FFM label bug.

• Fixed missing word_cloud module issue.

Version 2.15.230111

Bug Fixes

• Fixed the blank chart issue in dataset report.

• Fixed dataset report crash due to empty column.

Version 2.15.221223

Bug Fixes

• Fixed detected season change-points missing error of BCPD.

• Fixed Change points Chart in TimeSeriesReport.

Version 2.15.221216

New Functions

• Added nullif() function in dataframe.

• Added long-term time series forecast algorithm LTSF.

• Added plot_change_points() to plot change points.

• Added plot_psd() to plot power spectral density.

• Added periodogram() to perform power spectral density estimate of the input signal.

• Added change_point detection item to TimeSeriesReport.

Enhancements

• Enhanced successive halving/hyperband in FRM and ALS.

• Enhanced OutlierDetectionTS with IsolationForest and DBSCAN.

• Enhanced Data Type Timestamp support in BCPD.

• Enhanced UnifiedClustering with massive mode.

• Enhanced UnifiedReport with scoring report.

• Enhanced ModelStorage with model export/import function.

• Enhanced MLFlow integration with pipeline module and model export.

• Enhanced UnifiedReport with model debriefing and pipeline report.

• Enhanced force_plot in ShapleyExplainer to handle missing features.

• Enhanced HybridGradientBoostingClassifier and HybridGradientBoostingRegressor with early_stop option.

• Enhanced pipeline and auto-ml module (AutomaticClassification, AutomaticRegression and AutomaticTimeSeries) with predefined output tables.

API Changes

• Change the "JSON" column to NCLOB table type in ModelStorage.

Version 2.14.221208

Bug Fixes

• Fixed dependency issues in dataset report.

• Fixed documentation link in PyPI portal.

• Fixed replace() function to support NULL replacement.

Version 2.14.221201

Bug Fixes

• Fixed dataset report display issue:

  • wrong binning method for distribution plot.

  • NA handling issue in scatter matrix.

• Fixed SQL generation issue in hana_ml.algorithms.pal.pipeline.Pipeline module.

• Fixed model state creation in KNN.

• Fixed missing parameters in AutomaticTimeSeries.

• Fixed wrong type of split_method in AutomaticTimeSeries.

Version 2.14.221028

Bug Fixes

• Fixed pipeline monitor when password contains ','.

• Fixed message not defined error in auto-ml.

• Fixed pipeline error for PCA, DT and FN when disable_hana_execution() is executed.

• Fixed json pipeline generation for HGBT and RDT.

• Fixed parameter name typos for DT and RDT in UnifiedClassification.

• Fixed execute_statement parser when parameters contain special characters.

Version 2.14.221014

Bug Fixes

• Fixed legend issues in forecast_line_plot().

• Fixed duplicated outputs issue in artifact generator.

• Fixed best pipeline report that points exceed the chart.

• Fixed progress bar counter issue for AutomaticTimeSeries.

• Fixed predefined partition in unified API.

Version 2.14.220923

Bug Fixes

• Fixed degree_values issue Polynomial regression in UnifiedRegression.

• Fixed legend order in seasonal_plot().

• Fixed cross validation parameters in AutomaticTimeSeries forecast.

Version 2.14.220918

New Functions

• Added replace() function in dataframe.

• Added the class OutlierDetectionTS for time series outlier detection algorithm called.

• Added the class AutomaticTimeSeries.

• Added function make_future_dataframe().

• Added force_plot() for SHAP explainer.

• Added class ImputeTS.

• Added time series data report.

• Added function ks_test().

• Added function create_dataframe_from_spark().

• Added set_model_state function.

• Added outlier profiling.

• Added plot_time_series_outlier() in EDA.

Enhancements

• Enhanced model storage

  • support pipeline in auto-ml.

  • support model report.

• MLFlow integration for auto-ml.

• Pipeline module enhancement with PAL_PIPELINE_FIT and PAL_PIPELINE_PREDICT.

• Enhanced DataFrame function with enable_abap_sql().

• Successive halving for HGBT, KNN, SVM and MLR.

• Enhanced auto-ml with lightweight config dict option.

• Enhanced JSON model support in Multi-class LogisticRegression` and LinearRegression.

• Enhanced the support of French and Russian in tf_analysis(), hana_ml.text.tm.text_classification(), get_related_doc(), get_related_term(), get_relevant_doc(), get_relevant_term(), get_suggested_term() functions.

• Enhanced the support of pre-defined period setting in seasonal_decompose() with a new parameter periods.

API Changes

• Added initialization parameters handling_missing and json_export in LinearRegression.

• Added initialization parameters json_export, precompute_lms_sketch, stable_sketch_alg, sparse_sketch_alg in Multi-class LogisticRession.

• Added parameter periods in seasonal_decompose().

• Added parameters for APL segmented modeling, segmented forecast and parallel apply: max_tasks and segment_column_name (see APL 2209 and APL 2211 release notes).

Version 2.13.220722

Bug Fixes

• Fixed early_stop in auto-ml.

• Fixed display issue in unified report for APL.

Version 2.13.220715

Bug Fixes

• Fixed class_map0, class_map1 issue in UnifiedClassification.

• Fixed early_stop parameter missing in AutomaticClassification and AutomaticRegression.

• Fixed binary_classification_debriefing(): divided by zero issue.

Version 2.13.220701

Bug Fixes

• Fixed table name too long in model storage save_model() function.

• Fixed mlflow autologging with additional fit parameters.

• Fixed no mlflow model info display issue.

• Fixed metric sampling for model report.

• Fixed wrong schedule template in ModelStorage.

Version 2.13.220608

Bug Fixes

• fixed identifier length too long issue for function outputs.

Version 2.13.220511

New Functions

• Added upsert_streams_data() and update_streams_data() for DataFrame.

• Added stationary_test() function.

• Added class CrostonTSB.

• Added get_temporary_tables() and clean_up_temporary_tables() functions.

• Added Pipeline class json outputs for auto-ml pipeline_fit.

• Added EDA for time series data.

  • Added plot_pacf() and plot_acf()

  • Added plot_moving_average()

  • Added plot_rolling_stddev()

  • Added seasonal_plot()

  • Added timeseries_box_plot()

  • Added plot_seasonal_decompose()

  • Added quarter_plot()

• Added rolling window in generate_feature() function.

• Added get_connection_id(), restart_session() and cancel_session_operation() for class ConnectionContext.

Enhancements

• Added support of the following collection of new parameters for HGBT in UnifiedClassification and UnifiedRegression: replacemissing, default_missing_direction, feature_grouping, tol_rate, compression, max_bin_num.

• Improved the performance of box_plot().

• Enhanced the massive mode support of UninfiedClassification, UnifiedRegression, ARIMA, AutoARIMA, AdditiveModelForecast.

• Enhanced MLFlow autologging for UninfiedClassification and UnifiedRegression.

API Changes

• Added parameter interpret in predict() method of KNNClassifier and KNNRegressor for enabling procedure PAL_KNN_INTERPRET.

• Added parameters sample_size, top_k_attributions, random_state in predict() method of KNNClassifier and KNNRegressor for generating local interpretation result.

• Enabled missing value handling for input data by adding imputation related parameters in fit(), predict() and score() functions of both UninfiedClassification and UnifiedRegression.

• Added parameter model_type in GARCH for allowing variant GARCH models.

Bug Fixes

• Fixed key error bug for parameter param_values in DecisionTreeClassifier and DecisionTreeRegressor.

• Fixed the encoding error of imputation strategy of NONE type in Imputer.

• Fixed the key error bug when enabling AFL states for clustering algorithms.

Version 2.12.220428

Bug Fixes

• Adapted the auto-ml logging according to the PAL function changes.

Version 2.12.220425

Bug Fixes

• Fixed the display issue for the pipeline report.

• Fixed the missing ptype issue in AutoML evaluate function.

• Fixed the transform issue in pipeline fit_predict`() function.

Version 2.12.220408

Bug Fixes

• Fixed cancellation button in auto_ml.

• Fixed pivot_table() for handling NULL values.

• Fixed TreeModelDebriefing dot visualizer for Decision Trees.

• Fixed the display issue for dataset report with NULL values in dataset report.

Version 2.12.220325

New Functions

• Added IsolationForest.

• Added auto_ml including AutomaticClassification, AutomaticRegression and Preprocessing.

• Added progress monitor called PipelineProgressStatusMonitor for AutomaticClassification and AutomaticRegression

• Added best pipeline report called BestPipelineReport for AutomaticClassification and AutomaticRegression

• Added to_datetime(), to_tail() methods in DataFrame.

Enhancements

• Added validation procedure for n_components in CATPCA.

• Improved display name in pivot_table().

• Added compression and thresholding in hana_ml.algorithms.pal.tsa.wavelet.wavedec().

• Moved generate_feature() to DataFrame.

• Enhanced create_dataframe_from_pandas() with upsert option.

• Added ignore_scatter_matrix option in build() in dataset report.

• Expose APL variable selection parameters.

• Enhanced text mining module tm with German support.

• Support more objective functions(see obj_func) in HybridGradientBoostingClassifier and HybridGradientBoostingRegressor.

• Enhanced white_noise_test() with an option: the degree of freedom, model_df.

• Enhanced GRUAttention with local interpretability of model.

• Enhanced integer index support for explain_arima_model() for ARIMA and AutoARIMA.

• Added procomputed affinity for AgglomerateHierarchicalClustering.

• Added model compression related parameters for HybridGradientBoostingClassifier and HybridGradientBoostingRegressor.

Bug Fixes

• Fixed m4_sampling() with lowercase column name.

• Fixed inconsistent IDs assigned to solvers in LogisticRegression() between LOGR and M_LOGR.

• Fixed a parameter naming error in fft(): flattop_model --> flattop_mode.

• Fixed a validation error for endog parameter in predict() in Attention.

API Changes

• Added parameter model_df in the white_noise_test() for selecting the degree of freedom.

• Added parameter explain_mode in predict() of GRUAttention for selecting the mechanism for generating the reason code for inference results.

Version 2.11.220209

Bug Fixes

• Fixed wrong arg check for 'histogram' in parameter split_method of HybridGradientBoostingClassifier and HybridGradientBoostingRegressor.

• Fixed bug in deploy_class with transport_request.

Version 2.11.220107

Bug Fixes

• Fixed box_plot() with lower case column name.

• Fixed add_id() when the rel_col input is of list type.

• Fixed shortest_path and shortest_path_one_to_all type cast error.

• Fixed the alignment error in fast_dtw().

• Position correction for random search times in LogisticRegression.

• Fixed HANA hint script generation for resource restriction.

Version 2.11.211211

New Functions

• Added FeatureSelection.

• Added BSTS.

• Added WordCloud.

• Added hdbprocedure generation in PALBase and applied to all functions.

• Added GARCH.

• APL classification, regression, clustering: a new method, 'export_apply_code', generates code which can be used to apply a trained model outside APL.

Enhancements

• Enhanced Preprocessing with FeatureSelection.

• Enhanced the ModelStorage with fit parameters in json format.

• Enhanced PCA categorical support.

• Enhanced ModelStorage with fit parameters info.

• Enhanced UnifiedExponentialSmoothing with massive mode.

• Enhanced UnifiedClassification with AMDP generation as a function.

• Enhanced ARIMA with an explainer in the predict() function.

• Enhanced AdditiveModelForecast with an explainer in the predict() function.

• Enhanced UnifiedClassification with continued training of a trained HybridGradientBoostingClassifier model.

• Enhanced APL AutoTimeSeries with advanced predict outputs: the 'APL/ApplyExtraMode' parameter can be set in 'extra_applyout_settings'.

• Enhanced the stored procedure information retrieval.

• Enhanced fillna() to support non-numeric columns.

• Enhanced dataset report to convert PAL unsupported type.

API Changes

• Added initialization parameter background_size, and thread_ratio, top_k_attributions, trend_mod, trend_width, seasonal_width in the predict() method of ARIMA and AutoARIMA.

• Added parameters show_explainer, decompose_seasonality, decompose_holiday in the predict() function of AdditiveModelForecast.

• Added warm_start in the fit() method of HybridGradientBoostingClassifier as well as the fit() method of HybridGradientBoostingRegressor for continued training with existing model.

Bug Fixes

• Fixed index creation bug in on-premise text_classification api.

• Fixed multi-class LogisticRegression init check bug.

• Fixed has_table() error for local temporary tables.

Version 2.10.210918

New Functions

• Added dtw() for generic dynamic time warping with predefined and custom defined step pattern.

• Added wavedec() for multi-level discrete wavelet transformation, and waverec() for the corresponding inverse transformation.

• Added wpdec() and wprec() for multi-level (discrete) wavelet packet transformation and inverse.

• Added OnlineMultiLogisticRegression which is the online version of Multi-Class Logistic Regression.

• Added SpectralClustering.

• Added LSTM with attention.

• Added OneHotEncoding.

• Added unified preprocessor.

• Added plot() method for Pipeline.

• Added UnifiedExponentialSmoothing.

Enhancements

• Enhanced the model storage support for OnlineLinearRegression.

• Enhanced multi-threading in tm functions.

• Enhanced HDL container option.

• Enhanced timestamp support for ARIMA, AutoARIMA, VectorARIMA, OnlineARIMA, SingleExponentialSmoothing, DoubleExponentialSmoothing, TripleExponentialSmoothing, AutoExponentialSmoothing, BrownExponentialSmoothing, Croston, LR_seasonal_adjust.

• Enhanced new distributions for mcmc() sampling.

• Support multiple accuracy_measure methods in SingleExponentialSmoothing, DoubleExponentialSmoothing, TripleExponentialSmoothing, AutoExponentialSmoothing, BrownExponentialSmoothing, Croston, and LR_seasonal_adjust.

• Added plotly support.

API Changes

• Added parameters key, endog, exog, categorical_variable in the fit() function of AdditiveModelForecast.

• Added initialization parameters prediction_confidence_1 and prediction_confidence_2 in BrownExponentialSmoothing.

Version 2.9.210726

Bug Fixes

• Fixed load model initialized error in model storage service.

• Fixed bad link in pypi portal.

Version 2.9.210709

Bug Fixes

• Fixed missing WeaklyConnectedComponents in hana_ml.graph.algorithms.

• Fixed missing statistics in hana_ml.graph.Graph.describe.

• Fixed a bug, where the Graph object creation and discover_graph_workspace() and Graph.describe() do not work on an on-premise system

Version 2.9.210630

Bug Fixes

• Fixed accuracy_measure issue in SingleExponentialSmoothing, DoubleExponentialSmoothing, TripleExponentialSmoothing, and AutoExponentialSmoothing.

• Fixed empty input table error in Croston.

• Fixed class_map error for multiclass logisticregreesion in UnifiedClassification.

Version 2.9.210619

Enhancements

• Constants for directions used in graph functions can be found in hana_ml.graph.constants.DIRECTION_*

• Following functions and objects are now available in hana_ml.graph for import

  • Graph object

  • create_graph_from_dataframes and create_graph_from_hana_dataframes factory methods

  • discover_graph_workspaces

  • discover_graph_workspace

• The geometries do not need to be to be specified when creating a DataFrame instance anymore. The geometries are analyzed automatically.

• Support list of targets and trans_param in feature_tool.

• Enhanced unified report for UnifiedRegression to view feature importance.

• Enhanced join() to support list of DataFrame.

• Enhanced union() to support list of DataFrame.

• Streamlined the create_dataframe_from_pandas() geo parameters. Now there is only one list of geo_cols, which supports column references as well as (lon, lat) tuples, and one SRID parameter for all columns

• When you call create_dataframe_from_pandas`() and pass a GeoPandas DataFrame, the geometry column will be detected automatically and processed as a geometry. You don't need to add it manually to geo_cols

• The Graph constructor is simplified. You can instantiate a graph simply by the workspace name.

• Enhanced ModelStorage for APL to support HANA Data Lake.

New Functions

• Introduced hana_ml.graph.algorithms which contains all graph algorithms in the future. The package provides a AlgorithmBase class which can be used to build additional algorithms for a graph.

• Add hana_ml.graph.algorithms.ShortestPath, which replaces Graph.shortest_path

• Add hana_ml.graph.algorithms.Neighbors, which replaces Graph.neighbors

• Add hana_ml.graph.algorithms.NeighborsSubgraph, which replaces Graph.neighbors_with_edges

• Add hana_ml.graph.algorithms.KShortestPaths

• Add hana_ml.graph.algorithms.ShortestPathsOneToAll

• Add hana_ml.graph.discovery.discover_graph_workspace, which reads the metadata of a graph

• Add hana_ml.graph.create_graph_from_edges_dataframe

• Add hana_ml.graph.Graph.has_vertices, to check if a list of vertices exist in a graph

• Add hana_ml.graph.Graph.subgraph, to create a vertices or edges induced subgraph

• Add hana_ml.graph.Graph.describe, to get some statistics

• Add hana_ml.graph.Graph.degree_distribution

• Add hana_ml.DataFrame.srids, which returns the SRS of each geometry column

• Add hana_ml.DataFrame.geometries, which returns the geometry columns if there are any

• Add hana_ml.spatial package, that contains

  • create_predefined_srs

  • is_srs_created

  • get_created_srses

• Add hana_ml.docstore package, that contains

  • create_collection_from_elements

• Added BCPD for Bayesian change point detection.

• Added shape() method in DataFrame.

• Added sort_values(), sort_index() in DataFrame.

• Added scheduler for model renew in ModelStorage.

• Added min(), max(), mean(), median(), sum(), value_counts() in DataFrame.

• Added SHAP support for UnifiedClassification.

• Added data lake support in model_storage.

• Added data lake support in dataframe functions.

• Added line plot for time series forecast.

• Added split_column() method in DataFrame.

• Added concat_columns() method in DataFrame.

• Added outlier_detection_kmeans(), which detects outliers in datasets based on the result of k-means clustering.

• Added intermittent_forecast() for forecasting intermittent demand data(time-series).

• Added OnlineLinearRegression which is an online version of the Linear Regression.

API Changes

• Removed geo_cols from dataframe.create_dataframe_from_shapefile

• Removed geo_cols from ConnectionContext.sql()

• Removed geo_cols from ConnectionContext.table()

• Removed Graph.neighbors and Graph.neighbors_with_edges

• Removed Graph.shortest_path

• Removed hana_ml.graph.Path. This is not used anymore

• Removed hana_ml.graph.create_hana_graph_from_existing_workspace. This is replaced by a simplified Graph object constructor.

• Renamed hana_ml.graph.create_hana_graph_from_vertex_and_edge_frames to create_graph_from_dataframes

• Changed the type of geo_cols in create_dataframe_from_pandas to list, which supports direct column references or (lon, lat) tuples for generating POINT geometries

Bug Fixes

• Fixed inflexible default locations of selected columns of input data, e.g. key, features and endog.

• Fixed model report's feature importance when it has 0 importance.

Version 2.8.210421

Version 2.8.210421 supports SAP HANA SPS05 and SAP HANA Cloud

Bug Fixes

• Fixed model report's feature importance when it has 0 importance.

• Fixed pivot_table() with multiple index issue.

• Fixed the shap display for categorical columns.

Version 2.8.210321

Version 2.8.210321 supports SAP HANA SPS05 and SAP HANA Cloud

Enhancements

• Enhanced sql() to enable multiline execution.

• Enhanced save() to add append option.

• Enhanced diff() to enable negative input.

• Enhanced model report functionality of UnifiedClassification with added model and data visualization.

• Enhanced dataset report module with a optimized process of report generation and better user experience.

• Enhanced UnifiedClustering() to support parameter distance_level in AgglomerateHierarchicalClustering and DBSCAN functions. Please refer to documentation for details.

• Enhanced model storage to support unified report.

New Functions

• Added generate_html_report() and generate_notebook_iframe_report() functions for UnifiedRegression which could display the output, e.g. statistic and model.

• APL Gradient Boosting: the other_params parameter is now supported.

• APL all models: a new method, get_model_info, is created, allowing users to retrieve the summary and the performance metrics of a saved model.

• APL all models: users can now specify the weight of explanatory variables via the weight parameter.

• Added LSTM.

• Added Text Mining functions support for both SAP HANA on-premise and cloud version.

  • tf_analysis()

  • text_classification()

  • get_related_doc()

  • get_related_term()

  • get_relevant_doc()

  • get_relevant_term()

  • get_suggested_term()

• Added UnifiedReport.

New Dependency

• Added new dependency 'htmlmin' for generating dataset and model report.

API Changes

• Added parameters use_fast_library and use_float to KMeans.

• Added parameter build_report to UnifiedRegression.

• Added parameter distance_level in UnifiedClustering when func is AgglomerateHierarchicalClustering and DBSCAN. Please refer to documentation for details.

• Renamed batch_size by chunk_size in create_dataframe_from_pandas().

• Added initialization parameters random_state and random_initialization for OnlineARIMA, and its partial_fit() method now supports two additional parameters -- learning_rate and epsilon for updating the values in the input model.

Bug Fixes

• Fixed model storage support issue for OnlineARIMA.

• Fixed inflexible default locations of selected columns of input data, e.g. key, features and endog.

• Fixed accuracy_measure issue in AutoExponentialSmoothing.

Version 2.6.210126

Version 2.6.210126 supports SAP HANA SPS05 and SAP HANA Cloud

Bug Fixes

• Fixed uuid issue for Python 3.8.

• Fixed wrong legend for the model report of UnifiedClassification.

• Fixed dataset report to handle the dataset with missing value.

Version 2.6.210113

Version 2.6.210113 supports SAP HANA SPS05 and SAP HANA Cloud

Bug Fixes

• Fixed load_model issue for KMeans clustering.

• Removed pypi installation of Shapely for windows user.

• Fixed duplicate rows bug in save() function.

• Fixed loading issue in model report.

• Replaced the option batch_size with chunk_size in create_dataframe_from_pandas().

Version 2.6.201209

Version 2.6.201209 supports SAP HANA SPS05 and SAP HANA Cloud

Bug Fixes

• Remove shap from installation.

• Fixed bugs in ConnectionContext when autocommit=False.

• Fixed font properties bugs in hana_ml.visualizers.eda functions.

• APL Documentation: other_train_apl_aliases is now documented.

• APL Gradient Boosting Classification: the target variable won't be displayed in prediction if it is not given in input.

• APL Gradient Boosting: the default parameter values are now set in the APL backend level. They won't be set in the Python API level.

• Fixed handling of geometry columns in the context of collect() calls.

• Fixed shapely not being a required dependency.

• Fixed the displacement of parameter dispersion in CPD.

Version 2.6.201106

Version 2.6.201116 supports SAP HANA SPS05 and SAP HANA Cloud

Enhancements

• Enhanced the performance of collect() method for large datasets.

• Enhanced the performance of create_dataframe_from_pandas() for large datasets.

New Functions

• Added kdeplot() for 1D and 2D kde plotting.

• Added SHAPLEY visualization module shap().

Bug Fixes

• Fixed incompatibility issue with matplotlib>=3.3.0.

Version 2.6.201016(2.6.200928)

Version 2.6.201016 supports SAP HANA SPS05 and SAP HANA Cloud

API Changes

• HybridGradientBoostingClassifier and HybridGradientBoostingRegressor: added a parameter adopt_prior to indicate whether to adopt the prior distribution of the target as the initial point.

• Added parameters compression, max_bits, max_quantization_iter for the following SVM classes:

  • SVC

  • SVR

  • SVRanking

  • OneClassSVM

• RDTClassifier: added parameters compression, max_bits, quantize_rate for model compression.

• RDTRegressor: added parameters compression, max_bits, quantize_rate, fittings_quantization for model compression.

• In predict() method function ARIMA and AutoARIMA, a new value 'truncation_algorithm' of parameter forecast_method is introduced to improve the prediction performance.

• New intialization parameters string_variable, variable_weight are added to KNNClassifier, KNNRegressor and DBSCAN to enable distance calculation based on String distance.

• New parameters extrapolation, smooth_width, auxiliary_normalitytest are added to seasonal_decompose() function.

New functions

• Added dataset manager.

• Added graph and spatial modules.

• Added dataset report.

• Added clustering function: SlightSilhouette().

• Added native storage support in model storage service and dataset manager.

• Added VectorARIMA.

• Added UnifiedRegression.

• Added UnifiedClustering.

Bug Fixes

• Fixed ROC curve display in model report with disordered points.

• Fixed load_model() for UnifiedClassification in model storage service.

• Fixed model_selection for UnifiedClassification.

Version 2.5.200626

Version 2.5.200626 supports SAP HANA SPS05 and SAP HANA Cloud

API Changes

• Removed parameter ConnectionContext in PAL functions.

• Updated parameter algorithm from mandatory to optional in DecisionTreeClassifier and DecisionTreeRegressor, with default value 'cart'.

• Added parameter key in fit() method of ARIMA and AutoARIMA.

• Added parameter decompose_type in seasonal_decompose().

• Added parameter save_alignment and a new output statistic table in fast_dtw().

• Added parameter table_structure in create_dataframe_from_pandas().

• Added parameters resampling_method and param_search_strategy for HybridGradientBoostingClassifier and HybridGradientBoostingRegressor.

New Functions

• Added functions melt() and read_pickle().

• Added functionalities for UnifiedClassication. Especially, generate_html_report() and generate_notebook_iframe_report() are provided to visualize the output, e.g. confusion matrix and ROC curve.

• Added mcmc().

• Added model selection services as follows:

  • ParamSearchCV

  • GridSearchCV

  • RandomSearchCV

• Added visualizers for model debriefing.

Enhancements

• Enhanced smart sampling for visualizers.

• Enhanced import function to SAP HANA.

• Enhanced bytes, TIMESTAMP and BIGINT support in create_dataframe_from_pandas().

• Enhanced TIMESTAMP and DATE support in describe().

• Predictions made with APL gradient boosting can now be complemented with the reasons that led to these predictions: number of top or bottom explanatory variables, strength values, etc.

• Supported more data types, SMALLINT, DECIMAL, TINYINT, BIGINT, CLOB and BLOB in dtypes(), generate_table_type() and is_numeric().

• Enhanced the missing value handling ability in the groupby column by creating a new class for missing values for the following EDAVisualizer functions:

  • bar_plot()

  • box_plot()

  • pie_plot()

• Predictions made with APL gradient boosting can now be complemented with the reasons that led to these predictions: number of top or bottom explanatory variables, strength values, etc.

• APL gradient boosting can provide metrics about feature interactions strength.

• The connection parameter is no longer required for APL model creation.

Bug Fixes

• Fixed wrong ID issue in fit function by adding key initialization parameter in ARIMA and AutoARIMA.

• Fixed CLOB type issue in create_dataframe_from_pandas() by adding parameters table_structure and drop_exit_tab.

• Fixed pivot_table() index naming bug.

• Fixed temporary view from temporary table issue in APL time series function by adding sort_data and get_horizon_wide_metric.

• Fixed bugs in create_dataframe_from_pandas() if the table is temporary.

• Fixed bugs for data type of init centers in GaussianMixture.

• Fixed bugs when some data types, e.g. SMALLINT, DECIMAL or TINYINT are not supported in dtypes(), generate_table_type() and is_numeric().

• Fixed bugs when data types, e.g. DATE and TIMESTAMP, are not supported in describe().

• Fixed the table overwrite bug in save() if the table name is duplicate.

• Fixed missing quotation mark in column name bugs in EDA.

• Users can set 'Cutting Strategy' in APL Gradient Boosting.

• APL models are saved correctly.

Deprecated Functions

• GradientBoostingClassifier.

• GradientBoostingRegressor.

Version 1.0.8

Version 1.0.8 supports SAP HANA SP04 (100% coverage for SAP HANA SPS04 PAL algorithms)

New Functions

• Preprocessing: MDS (Multidimensional Scaling), SMOTE (Synthetic Minority Over-Sampling Technique, only supported in SAP HANA SPS05), Sampling, variance_test().

• Statistics: condition_index(), cdf() (Cumulative Distribution Function), distribution_fit(), quantile() (Distribution Quantile), entropy(), ftest_equal_var() (Equal Variance Test), factor_analysis(), grubbs_test(), kaplan_meier_survival_analysis(), KDE (Kernel Density), median_test_1samp() (One-Sample Median Test), wilcoxon() (Wilcox Signed Rank Test).

• Time-Series Analysis: LR_seasonal_adjust (Linear Regression with Damped Trend and Seasonal Adjust), AdditiveModelForecast, Hierarchical_Forecast, correlation() (Correlation Function), online algorithms and fast_dtw() (fast dynamic-time-warping).

• Miscellaneous: abc_analysis(), TSNE (T-distributed Stochastic Neighbour Embedding), weighted_score_table().

• Added functions in dataframe.py: data_manipulation().

• Added cross-validation options to SAP HANA PAL functions.

• Added visualizers (EDA profiler).

• Added model storage services.