Repository: incubator-hivemall Updated Branches: refs/heads/master 6b462ae45 -> e158f58ac
Close #44: [HIVEMALL-65] Update define-all.spark and import-packages.spark Project: http://git-wip-us.apache.org/repos/asf/incubator-hivemall/repo Commit: http://git-wip-us.apache.org/repos/asf/incubator-hivemall/commit/e158f58a Tree: http://git-wip-us.apache.org/repos/asf/incubator-hivemall/tree/e158f58a Diff: http://git-wip-us.apache.org/repos/asf/incubator-hivemall/diff/e158f58a Branch: refs/heads/master Commit: e158f58ac9680d71a91095118d0c6ff0d8684ed7 Parents: 6b462ae Author: Yuming Wang <wgy...@gmail.com> Authored: Thu Feb 16 01:10:00 2017 +0900 Committer: myui <yuin...@gmail.com> Committed: Thu Feb 16 01:10:00 2017 +0900 ---------------------------------------------------------------------- resources/ddl/define-all.spark | 176 +++++++++++++++++++++++-------- resources/ddl/import-packages.spark | 4 +- 2 files changed, 133 insertions(+), 47 deletions(-) ---------------------------------------------------------------------- http://git-wip-us.apache.org/repos/asf/incubator-hivemall/blob/e158f58a/resources/ddl/define-all.spark ---------------------------------------------------------------------- diff --git a/resources/ddl/define-all.spark b/resources/ddl/define-all.spark index e49a711..70385bc 100644 --- a/resources/ddl/define-all.spark +++ b/resources/ddl/define-all.spark @@ -32,8 +32,20 @@ sqlContext.sql("CREATE TEMPORARY FUNCTION train_arow AS 'hivemall.classifier.ARO sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS train_arowh") sqlContext.sql("CREATE TEMPORARY FUNCTION train_arowh AS 'hivemall.classifier.AROWClassifierUDTF$AROWh'") +sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS train_scw") +sqlContext.sql("CREATE TEMPORARY FUNCTION train_scw AS 'hivemall.classifier.SoftConfideceWeightedUDTF$SCW1'") + +sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS train_scw2") +sqlContext.sql("CREATE TEMPORARY FUNCTION train_scw2 AS 'hivemall.classifier.SoftConfideceWeightedUDTF$SCW2'") + sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS train_adagrad_rda") -sqlContext.sql("CREATE TEMPORARY FUNCTION train_arowh AS 'hivemall.classifier.AdaGradRDAUDTF'") +sqlContext.sql("CREATE TEMPORARY FUNCTION train_adagrad_rda AS 'hivemall.classifier.AdaGradRDAUDTF'") + +sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS train_kpa") +sqlContext.sql("CREATE TEMPORARY FUNCTION train_kpa AS 'hivemall.classifier.KernelExpansionPassiveAggressiveUDTF'") + +sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS kpa_predict") +sqlContext.sql("CREATE TEMPORARY FUNCTION kpa_predict AS 'hivemall.classifier.KPAPredictUDAF'") /** * Multiclass classification @@ -54,6 +66,12 @@ sqlContext.sql("CREATE TEMPORARY FUNCTION train_multiclass_pa2 AS 'hivemall.clas sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS train_multiclass_cw") sqlContext.sql("CREATE TEMPORARY FUNCTION train_multiclass_cw AS 'hivemall.classifier.multiclass.MulticlassConfidenceWeightedUDTF'") +sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS train_multiclass_arow") +sqlContext.sql("CREATE TEMPORARY FUNCTION train_multiclass_arow AS 'hivemall.classifier.multiclass.MulticlassAROWClassifierUDTF'") + +sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS train_multiclass_arowh") +sqlContext.sql("CREATE TEMPORARY FUNCTION train_multiclass_arowh AS 'hivemall.classifier.multiclass.MulticlassAROWClassifierUDTF$AROWh'") + sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS train_multiclass_scw") sqlContext.sql("CREATE TEMPORARY FUNCTION train_multiclass_scw AS 'hivemall.classifier.multiclass.MulticlassSoftConfidenceWeightedUDTF$SCW1'") @@ -64,10 +82,10 @@ sqlContext.sql("CREATE TEMPORARY FUNCTION train_multiclass_scw2 AS 'hivemall.cla * Similarity functions */ -sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS cosine_sim") +sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS cosine_similarity") sqlContext.sql("CREATE TEMPORARY FUNCTION cosine_similarity AS 'hivemall.knn.similarity.CosineSimilarityUDF'") -sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS jaccard") +sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS jaccard_similarity") sqlContext.sql("CREATE TEMPORARY FUNCTION jaccard_similarity AS 'hivemall.knn.similarity.JaccardIndexUDF'") sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS angular_similarity") @@ -83,15 +101,15 @@ sqlContext.sql("CREATE TEMPORARY FUNCTION distance2similarity AS 'hivemall.knn.s * Distance functions */ -sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS homming_distance") -sqlContext.sql("CREATE TEMPORARY FUNCTION hamming_distance AS 'hivemall.knn.distance.HammingDistanceUDF'") - sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS popcnt") sqlContext.sql("CREATE TEMPORARY FUNCTION popcnt AS 'hivemall.knn.distance.PopcountUDF'") sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS kld") sqlContext.sql("CREATE TEMPORARY FUNCTION kld AS 'hivemall.knn.distance.KLDivergenceUDF'") +sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS hamming_distance") +sqlContext.sql("CREATE TEMPORARY FUNCTION hamming_distance AS 'hivemall.knn.distance.HammingDistanceUDF'") + sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS euclid_distance") sqlContext.sql("CREATE TEMPORARY FUNCTION euclid_distance AS 'hivemall.knn.distance.EuclidDistanceUDF'") @@ -162,6 +180,9 @@ sqlContext.sql("CREATE TEMPORARY FUNCTION array_hash_values AS 'hivemall.ftvec.h sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS prefixed_hash_values") sqlContext.sql("CREATE TEMPORARY FUNCTION prefixed_hash_values AS 'hivemall.ftvec.hashing.ArrayPrefixedHashValuesUDF'") +sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS feature_hashing") +sqlContext.sql("CREATE TEMPORARY FUNCTION feature_hashing AS 'hivemall.ftvec.hashing.FeatureHashingUDF'") + /** * Feature pairing functions */ @@ -172,6 +193,9 @@ sqlContext.sql("CREATE TEMPORARY FUNCTION polynominal_features AS 'hivemall.ftve sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS powered_features") sqlContext.sql("CREATE TEMPORARY FUNCTION powered_features AS 'hivemall.ftvec.pairing.PoweredFeaturesUDF'") +sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS feature_pairs") +sqlContext.sql("CREATE TEMPORARY FUNCTION feature_pairs AS 'hivemall.ftvec.pairing.FeaturePairsUDTF'") + /** * Feature scaling functions */ @@ -182,8 +206,8 @@ sqlContext.sql("CREATE TEMPORARY FUNCTION rescale AS 'hivemall.ftvec.scaling.Res sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS zscore") sqlContext.sql("CREATE TEMPORARY FUNCTION zscore AS 'hivemall.ftvec.scaling.ZScoreUDF'") -sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS normalize") -sqlContext.sql("CREATE TEMPORARY FUNCTION normalize AS 'hivemall.ftvec.scaling.L2NormalizationUDF'") +sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS l2_normalize") +sqlContext.sql("CREATE TEMPORARY FUNCTION l2_normalize AS 'hivemall.ftvec.scaling.L2NormalizationUDF'") /** * Feature selection functions @@ -218,7 +242,13 @@ sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS extract_weight") sqlContext.sql("CREATE TEMPORARY FUNCTION extract_weight AS 'hivemall.ftvec.ExtractWeightUDFWrapper'") sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS add_feature_index") -sqlContext.sql("CREATE TEMPORARY FUNCTION add_feature_index AS 'hivemall.ftvec.AddFeatureIndexUDFWrapper'") +sqlContext.sql("CREATE TEMPORARY FUNCTION add_feature_index AS 'hivemall.ftvec.AddFeatureIndexUDF'") + +sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS feature") +sqlContext.sql("CREATE TEMPORARY FUNCTION feature AS 'hivemall.ftvec.FeatureUDF'") + +sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS feature_index") +sqlContext.sql("CREATE TEMPORARY FUNCTION feature_index AS 'hivemall.ftvec.FeatureIndexUDF'") /** * feature converting functions @@ -230,9 +260,15 @@ sqlContext.sql("CREATE TEMPORARY FUNCTION conv2dense AS 'hivemall.ftvec.conv.Con sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS to_dense_features") sqlContext.sql("CREATE TEMPORARY FUNCTION to_dense_features AS 'hivemall.ftvec.conv.ToDenseFeaturesUDF'") +sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS to_dense") +sqlContext.sql("CREATE TEMPORARY FUNCTION to_dense AS 'hivemall.ftvec.conv.ToDenseFeaturesUDF'") + sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS to_sparse_features") sqlContext.sql("CREATE TEMPORARY FUNCTION to_sparse_features AS 'hivemall.ftvec.conv.ToSparseFeaturesUDF'") +sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS to_sparse") +sqlContext.sql("CREATE TEMPORARY FUNCTION to_sparse AS 'hivemall.ftvec.conv.ToSparseFeaturesUDF'") + sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS quantify") sqlContext.sql("CREATE TEMPORARY FUNCTION quantify AS 'hivemall.ftvec.conv.QuantifyColumnsUDTF'") @@ -252,6 +288,9 @@ sqlContext.sql("CREATE TEMPORARY FUNCTION vectorize_features AS 'hivemall.ftvec. sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS categorical_features") sqlContext.sql("CREATE TEMPORARY FUNCTION categorical_features AS 'hivemall.ftvec.trans.CategoricalFeaturesUDF'") +sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS ffm_features") +sqlContext.sql("CREATE TEMPORARY FUNCTION ffm_features AS 'hivemall.ftvec.trans.FFMFeaturesUDF'") + sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS indexed_features") sqlContext.sql("CREATE TEMPORARY FUNCTION indexed_features AS 'hivemall.ftvec.trans.IndexedFeatures'") @@ -261,6 +300,25 @@ sqlContext.sql("CREATE TEMPORARY FUNCTION quantified_features AS 'hivemall.ftvec sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS quantitative_features") sqlContext.sql("CREATE TEMPORARY FUNCTION quantitative_features AS 'hivemall.ftvec.trans.QuantitativeFeaturesUDF'") +sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS binarize_label") +sqlContext.sql("CREATE TEMPORARY FUNCTION binarize_label AS 'hivemall.ftvec.trans.BinarizeLabelUDTF'") + +sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS onehot_encoding") +sqlContext.sql("CREATE TEMPORARY FUNCTION onehot_encoding AS 'hivemall.ftvec.trans.OnehotEncodingUDAF'") + +/** + * ranking helper functions + */ + +sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS bpr_sampling") +sqlContext.sql("CREATE TEMPORARY FUNCTION bpr_sampling AS 'hivemall.ftvec.ranking.BprSamplingUDTF'") + +sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS item_pairs_sampling") +sqlContext.sql("CREATE TEMPORARY FUNCTION item_pairs_sampling AS 'hivemall.ftvec.ranking.ItemPairsSamplingUDTF'") + +sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS populate_not_in") +sqlContext.sql("CREATE TEMPORARY FUNCTION populate_not_in AS 'hivemall.ftvec.ranking.PopulateNotInUDTF'") + /** * ftvec/text functions */ @@ -272,8 +330,11 @@ sqlContext.sql("CREATE TEMPORARY FUNCTION tf AS 'hivemall.ftvec.text.TermFrequen * Regression functions */ -sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS train_logregr") -sqlContext.sql("CREATE TEMPORARY FUNCTION train_logregr AS 'hivemall.regression.LogressUDTF'") +sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS logress") +sqlContext.sql("CREATE TEMPORARY FUNCTION logress AS 'hivemall.regression.LogressUDTF'") + +sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS train_logistic_regr") +sqlContext.sql("CREATE TEMPORARY FUNCTION train_logistic_regr AS 'hivemall.regression.LogressUDTF'") sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS train_pa1_regr") sqlContext.sql("CREATE TEMPORARY FUNCTION train_pa1_regr AS 'hivemall.regression.PassiveAggressiveRegressionUDTF'") @@ -287,20 +348,20 @@ sqlContext.sql("CREATE TEMPORARY FUNCTION train_pa2_regr AS 'hivemall.regression sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS train_pa2a_regr") sqlContext.sql("CREATE TEMPORARY FUNCTION train_pa2a_regr AS 'hivemall.regression.PassiveAggressiveRegressionUDTF$PA2a'") -sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS train_adadelta") -sqlContext.sql("CREATE TEMPORARY FUNCTION train_adadelta AS 'hivemall.regression.AdaDeltaUDTF'") - -sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS train_adagrad") -sqlContext.sql("CREATE TEMPORARY FUNCTION train_adagrad AS 'hivemall.regression.AdaGradUDTF'") - sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS train_arow_regr") sqlContext.sql("CREATE TEMPORARY FUNCTION train_arow_regr AS 'hivemall.regression.AROWRegressionUDTF'") sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS train_arowe_regr") -sqlContext.sql("CREATE TEMPORARY FUNCTION train_arow_regr AS 'hivemall.regression.AROWRegressionUDTF$AROWe'") +sqlContext.sql("CREATE TEMPORARY FUNCTION train_arowe_regr AS 'hivemall.regression.AROWRegressionUDTF$AROWe'") sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS train_arowe2_regr") -sqlContext.sql("CREATE TEMPORARY FUNCTION train_arow_regr AS 'hivemall.regression.AROWRegressionUDTF$AROWe2'") +sqlContext.sql("CREATE TEMPORARY FUNCTION train_arowe2_regr AS 'hivemall.regression.AROWRegressionUDTF$AROWe2'") + +sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS train_adagrad_regr") +sqlContext.sql("CREATE TEMPORARY FUNCTION train_adagrad_regr AS 'hivemall.regression.AdaGradUDTF'") + +sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS train_adadelta_regr") +sqlContext.sql("CREATE TEMPORARY FUNCTION train_adadelta_regr AS 'hivemall.regression.AdaDeltaUDTF'") /** * Array functions @@ -321,11 +382,8 @@ sqlContext.sql("CREATE TEMPORARY FUNCTION subarray_endwith AS 'hivemall.tools.ar sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS subarray_startwith") sqlContext.sql("CREATE TEMPORARY FUNCTION subarray_startwith AS 'hivemall.tools.array.SubarrayStartWithUDF'") -sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS collect_all") -sqlContext.sql("CREATE TEMPORARY FUNCTION collect_all AS 'hivemall.tools.array.CollectAllUDAF'") - -sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS concat_array") -sqlContext.sql("CREATE TEMPORARY FUNCTION concat_array AS 'hivemall.tools.array.ConcatArrayUDF'") +sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS array_concat") +sqlContext.sql("CREATE TEMPORARY FUNCTION array_concat AS 'hivemall.tools.array.ArrayConcatUDF'") sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS subarray") sqlContext.sql("CREATE TEMPORARY FUNCTION subarray AS 'hivemall.tools.array.SubarrayUDF'") @@ -333,10 +391,35 @@ sqlContext.sql("CREATE TEMPORARY FUNCTION subarray AS 'hivemall.tools.array.Suba sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS array_avg") sqlContext.sql("CREATE TEMPORARY FUNCTION array_avg AS 'hivemall.tools.array.ArrayAvgGenericUDAF'") +sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS array_sum") +sqlContext.sql("CREATE TEMPORARY FUNCTION array_sum AS 'hivemall.tools.array.ArraySumUDAF'") + +sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS to_string_array") +sqlContext.sql("CREATE TEMPORARY FUNCTION to_string_array AS 'hivemall.tools.array.ToStringArrayUDF'") + +sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS array_intersect") +sqlContext.sql("CREATE TEMPORARY FUNCTION array_intersect AS 'hivemall.tools.array.ArrayIntersectUDF'") + sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS select_k_best") sqlContext.sql("CREATE TEMPORARY FUNCTION select_k_best AS 'hivemall.tools.array.SelectKBestUDF'") /** + * Bit operation functions + */ + +sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS bits_collect") +sqlContext.sql("CREATE TEMPORARY FUNCTION bits_collect AS 'hivemall.tools.bits.BitsCollectUDAF'") + +sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS to_bits") +sqlContext.sql("CREATE TEMPORARY FUNCTION to_bits AS 'hivemall.tools.bits.ToBitsUDF'") + +sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS unbits") +sqlContext.sql("CREATE TEMPORARY FUNCTION unbits AS 'hivemall.tools.bits.UnBitsUDF'") + +sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS bits_or") +sqlContext.sql("CREATE TEMPORARY FUNCTION bits_or AS 'hivemall.tools.bits.BitsORUDF'") + +/** * Compression functions */ @@ -444,21 +527,17 @@ sqlContext.sql("CREATE TEMPORARY FUNCTION mse AS 'hivemall.evaluation.MeanSquare sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS rmse") sqlContext.sql("CREATE TEMPORARY FUNCTION rmse AS 'hivemall.evaluation.RootMeanSquaredErrorUDAF'") -/** - * Matrix Factorization - */ - -sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS mf_predict") -sqlContext.sql("CREATE TEMPORARY FUNCTION mf_predict AS 'hivemall.mf.MFPredictionUDF'") +sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS r2") +sqlContext.sql("CREATE TEMPORARY FUNCTION r2 AS 'hivemall.evaluation.R2UDAF'") -sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS train_mf_sgd") -sqlContext.sql("CREATE TEMPORARY FUNCTION train_mf_sgd AS 'hivemall.mf.MatrixFactorizationSGDUDTF'") +sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS ndcg") +sqlContext.sql("CREATE TEMPORARY FUNCTION ndcg AS 'hivemall.evaluation.NDCGUDAF'") -sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS train_mf_adagrad") -sqlContext.sql("CREATE TEMPORARY FUNCTION train_mf_adagrad AS 'hivemall.mf.MatrixFactorizationAdaGradUDTF'") +sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS logloss") +sqlContext.sql("CREATE TEMPORARY FUNCTION logloss AS 'hivemall.evaluation.LogarithmicLossUDAF'") /** - * Matrix factorization functions + * Matrix Factorization */ sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS mf_predict") @@ -470,6 +549,12 @@ sqlContext.sql("CREATE TEMPORARY FUNCTION train_mf_sgd AS 'hivemall.mf.MatrixFac sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS train_mf_adagrad") sqlContext.sql("CREATE TEMPORARY FUNCTION train_mf_adagrad AS 'hivemall.mf.MatrixFactorizationAdaGradUDTF'") +sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS train_bprmf") +sqlContext.sql("CREATE TEMPORARY FUNCTION train_bprmf AS 'hivemall.mf.BPRMatrixFactorizationUDTF'") + +sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS bprmf_predict") +sqlContext.sql("CREATE TEMPORARY FUNCTION bprmf_predict AS 'hivemall.mf.BPRMFPredictionUDF'") + /** * Factorization Machine */ @@ -480,6 +565,12 @@ sqlContext.sql("CREATE TEMPORARY FUNCTION fm_predict AS 'hivemall.fm.FMPredictGe sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS train_fm") sqlContext.sql("CREATE TEMPORARY FUNCTION train_fm AS 'hivemall.fm.FactorizationMachineUDTF'") +sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS train_ffm") +sqlContext.sql("CREATE TEMPORARY FUNCTION train_ffm AS 'hivemall.fm.FieldAwareFactorizationMachineUDTF'") + +sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS ffm_predict") +sqlContext.sql("CREATE TEMPORARY FUNCTION ffm_predict AS 'hivemall.fm.FFMPredictUDF'") + /** * Anomaly Detection */ @@ -497,23 +588,20 @@ sqlContext.sql("CREATE TEMPORARY FUNCTION sst AS 'hivemall.anomaly.SingularSpect sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS train_randomforest_classifier") sqlContext.sql("CREATE TEMPORARY FUNCTION train_randomforest_classifier AS 'hivemall.smile.classification.RandomForestClassifierUDTF'") +sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS train_randomforest_regressor") +sqlContext.sql("CREATE TEMPORARY FUNCTION train_randomforest_regressor AS 'hivemall.smile.regression.RandomForestRegressionUDTF'") + sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS train_randomforest_regr") sqlContext.sql("CREATE TEMPORARY FUNCTION train_randomforest_regr AS 'hivemall.smile.regression.RandomForestRegressionUDTF'") sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS tree_predict") sqlContext.sql("CREATE TEMPORARY FUNCTION tree_predict AS 'hivemall.smile.tools.TreePredictUDF'") -sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS vm_tree_predict") -sqlContext.sql("CREATE TEMPORARY FUNCTION vm_tree_predict AS 'hivemall.smile.tools.TreePredictByStackMachineUDF'") - -sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS js_tree_predict") -sqlContext.sql("CREATE TEMPORARY FUNCTION js_tree_predict AS 'hivemall.smile.tools.TreePredictByJavascriptUDF'") - sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS rf_ensemble") sqlContext.sql("CREATE TEMPORARY FUNCTION rf_ensemble AS 'hivemall.smile.tools.RandomForestEnsembleUDAF'") -sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS train_gradient_tree_boosting_classifier") -sqlContext.sql("CREATE TEMPORARY FUNCTION train_gradient_tree_boosting_classifier AS 'hivemall.smile.classification.GradientTreeBoostingClassifierUDTF'") - sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS guess_attribute_types") sqlContext.sql("CREATE TEMPORARY FUNCTION guess_attribute_types AS 'hivemall.smile.tools.GuessAttributesUDF'") + +sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS train_gradient_tree_boosting_classifier") +sqlContext.sql("CREATE TEMPORARY FUNCTION train_gradient_tree_boosting_classifier AS 'hivemall.smile.classification.GradientTreeBoostingClassifierUDTF'") http://git-wip-us.apache.org/repos/asf/incubator-hivemall/blob/e158f58a/resources/ddl/import-packages.spark ---------------------------------------------------------------------- diff --git a/resources/ddl/import-packages.spark b/resources/ddl/import-packages.spark index d1685a8..2015cd8 100644 --- a/resources/ddl/import-packages.spark +++ b/resources/ddl/import-packages.spark @@ -2,14 +2,12 @@ * An initialization script for DataFrame use */ -import org.apache.spark.ml.feature.HmLabeledPoint -import org.apache.spark.ml.feature.HmFeature import org.apache.spark.sql._ import org.apache.spark.sql.functions._ import org.apache.spark.sql.types._ import org.apache.spark.sql.hive.HivemallOps._ import org.apache.spark.sql.hive.HivemallUtils -import org.apache.spark.sql.hive.XGBoostOptions +import hivemall.xgboost.XGBoostOptions // Needed for implicit conversions import org.apache.spark.sql.hive.HivemallUtils._