biA`dZddlZddlmZmZmZmZmZddlZ ddl m Z ddl m Z mZddlmZmZddlmZdd lmZmZmZd d lmZmZmZd d lmZd eedeeddfdZddZGddeZ GddeZ!ee e!GddeeeZ"GddeZ#ee"e#GddeZ$GddeZ%ee$e%y)z8Xgboost pyspark integration submodule for estimator API.N)AnyListOptionalTypeUnion) keyword_only)ParamParams)HasProbabilityColHasRawPredictionCol)Config) XGBClassifier XGBRanker XGBRegressor)_ClassificationModel_SparkXGBEstimator_SparkXGBModel)get_class_name estimatormodelreturnc j}dtdtf fd dtdtddf fd }|j D]J}dt j d |}ttj|| }|||Lj}|j D]T}dt j d |}|d k(r|d z }ttj|| }|||Vj}|j D]J}dt j d|}ttj|| }|||Ly)zyThis function automatically infer to xgboost parameters and set them into corresponding pyspark estimators and modelsvrcXt|tjr#tj|j St|t r*|j Dcic]\}}||c}}St|tr|Dcgc] }| c}S|Scc}}wcc}wN) isinstancenpgenericarrayitemdictitemslist)rknvparam_value_converters R/home/cdr/jupyterlab/.venv/lib/python3.12/site-packages/xgboost/spark/estimator.pyr(z?_set_pyspark_xgb_cls_param_attrs..param_value_converters a $88A;##% % a >?ggiHUQA,R00H H a 89:")"-: :I:s "B! B' attr_nameparamNcH|_t||t||yr) typeConvertersetattr)r*r+rrr(s r)set_param_attrsz9_set_pyspark_xgb_cls_param_attrs..set_param_attrs)s#3 9e,y%(zRefer to XGBoost doc of z for this param )namedocz.fit() for this param callbackszThe callbacks can be arbitrary functions. It is saved using cloudpickle which is not a fully self-contained format. It may fail to load with different versions of dependencies.z.predict() for this param ) _get_xgb_params_defaultrstrr keysr_xgb_clsr _dummy_get_fit_params_default_get_predict_params_default) rr params_dictr/r1r2 param_objfit_params_dictpredict_params_dictr(s `` @r) _set_pyspark_xgb_cls_param_attrsr?s 335K)3)u))   ")&i002344DTF L !tE i() 779O$$& )&~i6H6H6J'K&L$TF ,  ;  6 C &--/#> i( )$??A#((*)&~i6H6H6J'K&L( 0 &--/#> i( )r0c8tjdty)Nz9`use_gpu` is deprecated since 2.0.0, use `device` instead)warningswarn FutureWarningr0r)_deprecated_use_gpurEPs MMC]r0c$eZdZdZedddddddddddddddd d eeeefd ed ed eedeedeedeede dee deede de de de dee de ddf"fdZ edeefdZededfdZd fd ZxZS)!SparkXGBRegressoraSparkXGBRegressor is a PySpark ML estimator. It implements the XGBoost regression algorithm based on XGBoost python library, and it can be used in PySpark Pipeline and PySpark ML meta algorithms like - :py:class:`~pyspark.ml.tuning.CrossValidator`/ - :py:class:`~pyspark.ml.tuning.TrainValidationSplit`/ - :py:class:`~pyspark.ml.classification.OneVsRest` SparkXGBRegressor automatically supports most of the parameters in :py:class:`xgboost.XGBRegressor` constructor and most of the parameters used in :py:meth:`xgboost.XGBRegressor.fit` and :py:meth:`xgboost.XGBRegressor.predict` method. To enable GPU support, set `device` to `cuda` or `gpu`. SparkXGBRegressor doesn't support setting `base_margin` explicitly as well, but support another param called `base_margin_col`. see doc below for more details. SparkXGBRegressor doesn't support `validate_features` and `output_margin` param. SparkXGBRegressor doesn't support setting `nthread` xgboost param, instead, the `nthread` param for each xgboost worker will be set equal to `spark.task.cpus` config value. Parameters ---------- features_col: When the value is string, it requires the features column name to be vector type. When the value is a list of string, it requires all the feature columns to be numeric types. label_col: Label column name. Default to "label". prediction_col: Prediction column name. Default to "prediction" pred_contrib_col: Contribution prediction column name. validation_indicator_col: For params related to `xgboost.XGBRegressor` training with evaluation dataset's supervision, set :py:attr:`xgboost.spark.SparkXGBRegressor.validation_indicator_col` parameter instead of setting the `eval_set` parameter in `xgboost.XGBRegressor` fit method. weight_col: To specify the weight of the training and validation dataset, set :py:attr:`xgboost.spark.SparkXGBRegressor.weight_col` parameter instead of setting `sample_weight` and `sample_weight_eval_set` parameter in `xgboost.XGBRegressor` fit method. base_margin_col: To specify the base margins of the training and validation dataset, set :py:attr:`xgboost.spark.SparkXGBRegressor.base_margin_col` parameter instead of setting `base_margin` and `base_margin_eval_set` in the `xgboost.XGBRegressor` fit method. num_workers: How many XGBoost workers to be used to train. Each XGBoost worker corresponds to one spark task. use_gpu: .. deprecated:: 2.0.0 Use `device` instead. device: .. versionadded:: 2.0.0 Device for XGBoost workers, available options are `cpu`, `cuda`, and `gpu`. force_repartition: Boolean value to specify if forcing the input dataset to be repartitioned before XGBoost training. repartition_random_shuffle: Boolean value to specify if randomly shuffling the dataset when repartitioning is required. enable_sparse_data_optim: Boolean value to specify if enabling sparse data optimization, if True, Xgboost DMatrix object will be constructed from sparse matrix instead of dense matrix. launch_tracker_on_driver: Boolean value to indicate whether the tracker should be launched on the driver side or the executor side. coll_cfg: The collective configuration. See :py:class:`~xgboost.collective.Config` kwargs: A dictionary of xgboost parameters, please refer to https://xgboost.readthedocs.io/en/stable/parameter.html Note ---- The Parameters chart above contains parameters that need special handling. For a full list of parameters, see entries with `Param(parent=...` below. This API is experimental. Examples -------- >>> from xgboost.spark import SparkXGBRegressor >>> from pyspark.ml.linalg import Vectors >>> df_train = spark.createDataFrame([ ... (Vectors.dense(1.0, 2.0, 3.0), 0, False, 1.0), ... (Vectors.sparse(3, {1: 1.0, 2: 5.5}), 1, False, 2.0), ... (Vectors.dense(4.0, 5.0, 6.0), 2, True, 1.0), ... (Vectors.sparse(3, {1: 6.0, 2: 7.5}), 3, True, 2.0), ... ], ["features", "label", "isVal", "weight"]) >>> df_test = spark.createDataFrame([ ... (Vectors.dense(1.0, 2.0, 3.0), ), ... (Vectors.sparse(3, {1: 1.0, 2: 5.5}), ) ... ], ["features"]) >>> xgb_regressor = SparkXGBRegressor(max_depth=5, missing=0.0, ... validation_indicator_col='isVal', weight_col='weight', ... early_stopping_rounds=1, eval_metric='rmse') >>> xgb_reg_model = xgb_regressor.fit(df_train) >>> xgb_reg_model.transform(df_test) featureslabel predictionNrFT) features_col label_colprediction_colpred_contrib_colvalidation_indicator_col weight_colbase_margin_col num_workersuse_gpudeviceforce_repartitionrepartition_random_shuffleenable_sparse_data_optimlaunch_tracker_on_drivercoll_cfgrKrLrMrNrOrPrQrRrSrTrUrVrWrXrYkwargsrc vt||j}| r t|jdi|yNrDsuper__init__ _input_kwargsrE setParams)selfrKrLrMrNrOrPrQrRrSrTrUrVrWrXrYrZ input_kwargs __class__s r)r_zSparkXGBRegressor.__init__s5* ))   !&&r0ctSrrclss r)r7zSparkXGBRegressor._xgb_clsr0SparkXGBRegressorModelctSr)rjrgs r)_pyspark_model_clsz$SparkXGBRegressor._pyspark_model_clss%%r0cnt||j|jr t dy)NzCSpark Xgboost regressor estimator does not support `qid_col` param.r^_validate_params isDefinedqid_col ValueErrorrbrds r)roz"SparkXGBRegressor._validate_paramss3  " >>$,, 'U  (r0rN)__name__ __module__ __qualname____doc__rrr5rrintboolrrr_ classmethodrrr7rlro __classcell__rds@r)rGrGVsltl/9 **.26$()-"& $"'+0).)-%)#'CcN+' '  ' #3- '#+3-'SM'"#''$' ' '%)'#'' #'!'"6"#'$%'& '''4l+&4(@#A&&r0rGc,eZdZdZedeefdZy)rjzt The model returned by :func:`xgboost.spark.SparkXGBRegressor.fit` .. Note:: This API is experimental. rctSrrfrgs r)r7zSparkXGBRegressorModel._xgb_clsrir0N)rurvrwrxr{rrr7rDr0r)rjrjs' l+r0rjc(eZdZdZeddddddddddddd d d d dd d eeeefd ededededeedeedeedeede dee deede de de de dee de ddf&fdZ edeefd Zeded!fd"Zd$fd# ZxZS)%SparkXGBClassifieraSparkXGBClassifier is a PySpark ML estimator. It implements the XGBoost classification algorithm based on XGBoost python library, and it can be used in PySpark Pipeline and PySpark ML meta algorithms like - :py:class:`~pyspark.ml.tuning.CrossValidator`/ - :py:class:`~pyspark.ml.tuning.TrainValidationSplit`/ - :py:class:`~pyspark.ml.classification.OneVsRest` SparkXGBClassifier automatically supports most of the parameters in :py:class:`xgboost.XGBClassifier` constructor and most of the parameters used in :py:meth:`xgboost.XGBClassifier.fit` and :py:meth:`xgboost.XGBClassifier.predict` method. To enable GPU support, set `device` to `cuda` or `gpu`. SparkXGBClassifier doesn't support setting `base_margin` explicitly as well, but support another param called `base_margin_col`. see doc below for more details. SparkXGBClassifier doesn't support setting `output_margin`, but we can get output margin from the raw prediction column. See `raw_prediction_col` param doc below for more details. SparkXGBClassifier doesn't support `validate_features` and `output_margin` param. SparkXGBClassifier doesn't support setting `nthread` xgboost param, instead, the `nthread` param for each xgboost worker will be set equal to `spark.task.cpus` config value. Parameters ---------- features_col: When the value is string, it requires the features column name to be vector type. When the value is a list of string, it requires all the feature columns to be numeric types. label_col: Label column name. Default to "label". prediction_col: Prediction column name. Default to "prediction" probability_col: Column name for predicted class conditional probabilities. Default to probabilityCol raw_prediction_col: The `output_margin=True` is implicitly supported by the `rawPredictionCol` output column, which is always returned with the predicted margin values. pred_contrib_col: Contribution prediction column name. validation_indicator_col: For params related to `xgboost.XGBClassifier` training with evaluation dataset's supervision, set :py:attr:`xgboost.spark.SparkXGBClassifier.validation_indicator_col` parameter instead of setting the `eval_set` parameter in `xgboost.XGBClassifier` fit method. weight_col: To specify the weight of the training and validation dataset, set :py:attr:`xgboost.spark.SparkXGBClassifier.weight_col` parameter instead of setting `sample_weight` and `sample_weight_eval_set` parameter in `xgboost.XGBClassifier` fit method. base_margin_col: To specify the base margins of the training and validation dataset, set :py:attr:`xgboost.spark.SparkXGBClassifier.base_margin_col` parameter instead of setting `base_margin` and `base_margin_eval_set` in the `xgboost.XGBClassifier` fit method. num_workers: How many XGBoost workers to be used to train. Each XGBoost worker corresponds to one spark task. use_gpu: .. deprecated:: 2.0.0 Use `device` instead. device: .. versionadded:: 2.0.0 Device for XGBoost workers, available options are `cpu`, `cuda`, and `gpu`. force_repartition: Boolean value to specify if forcing the input dataset to be repartitioned before XGBoost training. repartition_random_shuffle: Boolean value to specify if randomly shuffling the dataset when repartitioning is required. enable_sparse_data_optim: Boolean value to specify if enabling sparse data optimization, if True, Xgboost DMatrix object will be constructed from sparse matrix instead of dense matrix. launch_tracker_on_driver: Boolean value to indicate whether the tracker should be launched on the driver side or the executor side. coll_cfg: The collective configuration. See :py:class:`~xgboost.collective.Config` kwargs: A dictionary of xgboost parameters, please refer to https://xgboost.readthedocs.io/en/stable/parameter.html Note ---- The Parameters chart above contains parameters that need special handling. For a full list of parameters, see entries with `Param(parent=...` below. This API is experimental. Examples -------- >>> from xgboost.spark import SparkXGBClassifier >>> from pyspark.ml.linalg import Vectors >>> df_train = spark.createDataFrame([ ... (Vectors.dense(1.0, 2.0, 3.0), 0, False, 1.0), ... (Vectors.sparse(3, {1: 1.0, 2: 5.5}), 1, False, 2.0), ... (Vectors.dense(4.0, 5.0, 6.0), 0, True, 1.0), ... (Vectors.sparse(3, {1: 6.0, 2: 7.5}), 1, True, 2.0), ... ], ["features", "label", "isVal", "weight"]) >>> df_test = spark.createDataFrame([ ... (Vectors.dense(1.0, 2.0, 3.0), ), ... ], ["features"]) >>> xgb_classifier = SparkXGBClassifier(max_depth=5, missing=0.0, ... validation_indicator_col='isVal', weight_col='weight', ... early_stopping_rounds=1, eval_metric='logloss') >>> xgb_clf_model = xgb_classifier.fit(df_train) >>> xgb_clf_model.transform(df_test).show() rHrIrJ probability rawPredictionNrFT)rKrLrMprobability_colraw_prediction_colrNrOrPrQrRrSrTrUrVrWrXrYrKrLrMrrrNrOrPrQrRrSrTrUrVrWrXrYrZrc t||j}| r t|jdi||j dy)N) objectiverD)r^r_r`rEra _setDefault)rbrKrLrMrrrNrOrPrQrRrSrTrUrVrWrXrYrZrcrds r)r_zSparkXGBClassifier.__init__sH.  ))   !&& 4(r0ctSrrrgs r)r7zSparkXGBClassifier._xgb_clsr0SparkXGBClassifierModelctSr)rrgs r)rlz%SparkXGBClassifier._pyspark_model_clss&&r0ct||j|jr t d|j dr t dy)NzDSpark Xgboost classifier estimator does not support `qid_col` param.rzHSetting custom 'objective' param is not allowed in 'SparkXGBClassifier'.)r^rorprqrr getOrDefaultrss r)roz#SparkXGBClassifier._validate_paramssS  " >>$,, 'V    [ )Z  *r0rt)rurvrwrxrrr5rrryrzrrr_r{rrr7rlror|r}s@r)rrs||/9 *,"1*.26$()-"& $"'+0).)-%)')CcN+) )  )  ) )#3-)#+3-)SM)"#))$) ) ) %)!)"#'#)$#'%)&6"')())* +))Bm,'4(A#B''  r0rc,eZdZdZedeefdZy)rzu The model returned by :func:`xgboost.spark.SparkXGBClassifier.fit` .. Note:: This API is experimental. rctSrrrgs r)r7z SparkXGBClassifierModel._xgb_clsrr0N)rurvrwrxr{rrr7rDr0r)rrs' m,r0rc&eZdZdZeddddddddddddddddd d eeeefd ed ed eedeedeedeedeede dee deede de de de dee de ddf$fdZ edeefdZededfdZd!fd ZxZS)"SparkXGBRankeraSparkXGBRanker is a PySpark ML estimator. It implements the XGBoost ranking algorithm based on XGBoost python library, and it can be used in PySpark Pipeline and PySpark ML meta algorithms like :py:class:`~pyspark.ml.tuning.CrossValidator`/ :py:class:`~pyspark.ml.tuning.TrainValidationSplit`/ :py:class:`~pyspark.ml.classification.OneVsRest` SparkXGBRanker automatically supports most of the parameters in :py:class:`xgboost.XGBRanker` constructor and most of the parameters used in :py:meth:`xgboost.XGBRanker.fit` and :py:meth:`xgboost.XGBRanker.predict` method. To enable GPU support, set `device` to `cuda` or `gpu`. SparkXGBRanker doesn't support setting `base_margin` explicitly as well, but support another param called `base_margin_col`. see doc below for more details. SparkXGBRanker doesn't support setting `output_margin`, but we can get output margin from the raw prediction column. See `raw_prediction_col` param doc below for more details. SparkXGBRanker doesn't support `validate_features` and `output_margin` param. SparkXGBRanker doesn't support setting `nthread` xgboost param, instead, the `nthread` param for each xgboost worker will be set equal to `spark.task.cpus` config value. Parameters ---------- features_col: When the value is string, it requires the features column name to be vector type. When the value is a list of string, it requires all the feature columns to be numeric types. label_col: Label column name. Default to "label". prediction_col: Prediction column name. Default to "prediction" pred_contrib_col: Contribution prediction column name. validation_indicator_col: For params related to `xgboost.XGBRanker` training with evaluation dataset's supervision, set :py:attr:`xgboost.spark.SparkXGBRanker.validation_indicator_col` parameter instead of setting the `eval_set` parameter in :py:class:`xgboost.XGBRanker` fit method. weight_col: To specify the weight of the training and validation dataset, set :py:attr:`xgboost.spark.SparkXGBRanker.weight_col` parameter instead of setting `sample_weight` and `sample_weight_eval_set` parameter in :py:class:`xgboost.XGBRanker` fit method. base_margin_col: To specify the base margins of the training and validation dataset, set :py:attr:`xgboost.spark.SparkXGBRanker.base_margin_col` parameter instead of setting `base_margin` and `base_margin_eval_set` in the :py:class:`xgboost.XGBRanker` fit method. qid_col: Query id column name. num_workers: How many XGBoost workers to be used to train. Each XGBoost worker corresponds to one spark task. use_gpu: .. deprecated:: 2.0.0 Use `device` instead. device: .. versionadded:: 2.0.0 Device for XGBoost workers, available options are `cpu`, `cuda`, and `gpu`. force_repartition: Boolean value to specify if forcing the input dataset to be repartitioned before XGBoost training. repartition_random_shuffle: Boolean value to specify if randomly shuffling the dataset when repartitioning is required. enable_sparse_data_optim: Boolean value to specify if enabling sparse data optimization, if True, Xgboost DMatrix object will be constructed from sparse matrix instead of dense matrix. launch_tracker_on_driver: Boolean value to indicate whether the tracker should be launched on the driver side or the executor side. coll_cfg: The collective configuration. See :py:class:`~xgboost.collective.Config` kwargs: A dictionary of xgboost parameters, please refer to https://xgboost.readthedocs.io/en/stable/parameter.html .. Note:: The Parameters chart above contains parameters that need special handling. For a full list of parameters, see entries with `Param(parent=...` below. .. Note:: This API is experimental. Examples -------- >>> from xgboost.spark import SparkXGBRanker >>> from pyspark.ml.linalg import Vectors >>> ranker = SparkXGBRanker(qid_col="qid") >>> df_train = spark.createDataFrame( ... [ ... (Vectors.dense(1.0, 2.0, 3.0), 0, 0), ... (Vectors.dense(4.0, 5.0, 6.0), 1, 0), ... (Vectors.dense(9.0, 4.0, 8.0), 2, 0), ... (Vectors.sparse(3, {1: 1.0, 2: 5.5}), 0, 1), ... (Vectors.sparse(3, {1: 6.0, 2: 7.5}), 1, 1), ... (Vectors.sparse(3, {1: 8.0, 2: 9.5}), 2, 1), ... ], ... ["features", "label", "qid"], ... ) >>> df_test = spark.createDataFrame( ... [ ... (Vectors.dense(1.5, 2.0, 3.0), 0), ... (Vectors.dense(4.5, 5.0, 6.0), 0), ... (Vectors.dense(9.0, 4.5, 8.0), 0), ... (Vectors.sparse(3, {1: 1.0, 2: 6.0}), 1), ... (Vectors.sparse(3, {1: 6.0, 2: 7.0}), 1), ... (Vectors.sparse(3, {1: 8.0, 2: 10.5}), 1), ... ], ... ["features", "qid"], ... ) >>> model = ranker.fit(df_train) >>> model.transform(df_test).show() rHrIrJNrFT)rKrLrMrNrOrPrQrqrRrSrTrUrVrWrXrYrKrLrMrNrOrPrQrqrRrSrTrUrVrWrXrYrZrc vt||j}| r t|jdi|yr\r])rbrKrLrMrNrOrPrQrqrRrSrTrUrVrWrXrYrZrcrds r)r_zSparkXGBRanker.__init__Ks5, ))   !&&r0ctSrrrgs r)r7zSparkXGBRanker._xgb_clsgr0SparkXGBRankerModelctSr)rrgs r)rlz!SparkXGBRanker._pyspark_model_clsks""r0cnt||j|js t dy)Nz@Spark Xgboost ranker estimator requires setting `qid_col` param.rnrss r)rozSparkXGBRanker._validate_paramsos3  "~~dll+R ,r0rt)rurvrwrxrrr5rrryrzrrr_r{rrr7rlror|r}s@r)rrs|}~/9 **.26$()-!%"& $"'+0).)-%)%'CcN+' '  ' #3- '#+3-'SM'"#'#''$' ' '%)' #'!'"#'#'$6"%'&''( )''6i#4(=#>##r0rc,eZdZdZedeefdZy)rzq The model returned by :func:`xgboost.spark.SparkXGBRanker.fit` .. Note:: This API is experimental. rctSrrrgs r)r7zSparkXGBRankerModel._xgb_cls~rr0N)rurvrwrxr{rrr7rDr0r)rrws& ir0rrt)&rxrAtypingrrrrrnumpyrpysparkrpyspark.ml.paramr r pyspark.ml.param.sharedr r collectiversklearnrrrcorerrrutilsrr?rErGrjrrrrrDr0r)rs> 33 *J<< "5)&'5)04^0D5) 5)p _*_D ^ !!24JKr+->@Srj 2 !!35LMi'iX . !1DEr0