| /** |
| * Hivemall: Hive scalable Machine Learning Library |
| */ |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS hivemall_version") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION hivemall_version AS 'hivemall.HivemallVersionUDF'") |
| |
| /** |
| * binary classification |
| */ |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS train_perceptron") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION train_perceptron AS 'hivemall.classifier.PerceptronUDTF'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS train_pa") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION train_pa AS 'hivemall.classifier.PassiveAggressiveUDTF'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS train_pa1") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION train_pa1 AS 'hivemall.classifier.PassiveAggressiveUDTF$PA1'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS train_pa2") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION train_pa2 AS 'hivemall.classifier.PassiveAggressiveUDTF$PA2'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS train_cw") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION train_cw AS 'hivemall.classifier.ConfidenceWeightedUDTF'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS train_arow") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION train_arow AS 'hivemall.classifier.AROWClassifierUDTF'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS train_arowh") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION train_arowh AS 'hivemall.classifier.AROWClassifierUDTF$AROWh'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS train_adagrad_rda") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION train_arowh AS 'hivemall.classifier.AdaGradRDAUDTF'") |
| |
| /** |
| * Multiclass classification |
| */ |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS train_multiclass_perceptron") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION train_multiclass_perceptron AS 'hivemall.classifier.multiclass.MulticlassPerceptronUDTF'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS train_multiclass_pa") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION train_multiclass_pa AS 'hivemall.classifier.multiclass.MulticlassPassiveAggressiveUDTF'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS train_multiclass_pa1") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION train_multiclass_pa1 AS 'hivemall.classifier.multiclass.MulticlassPassiveAggressiveUDTF$PA1'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS train_multiclass_pa2") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION train_multiclass_pa2 AS 'hivemall.classifier.multiclass.MulticlassPassiveAggressiveUDTF$PA2'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS train_multiclass_cw") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION train_multiclass_cw AS 'hivemall.classifier.multiclass.MulticlassConfidenceWeightedUDTF'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS train_multiclass_scw") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION train_multiclass_scw AS 'hivemall.classifier.multiclass.MulticlassSoftConfidenceWeightedUDTF$SCW1'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS train_multiclass_scw2") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION train_multiclass_scw2 AS 'hivemall.classifier.multiclass.MulticlassSoftConfidenceWeightedUDTF$SCW2'") |
| |
| /** |
| * similarity functions |
| */ |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS cosine_sim") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION cosine_similarity AS 'hivemall.knn.similarity.CosineSimilarityUDF'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS jaccard") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION jaccard_similarity AS 'hivemall.knn.similarity.JaccardIndexUDF'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS angular_similarity") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION angular_similarity AS 'hivemall.knn.similarity.AngularSimilarityUDF'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS euclid_similarity") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION euclid_similarity AS 'hivemall.knn.similarity.EuclidSimilarity'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS distance2similarity") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION distance2similarity AS 'hivemall.knn.similarity.Distance2SimilarityUDF'") |
| |
| /** |
| * 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 euclid_distance") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION euclid_distance AS 'hivemall.knn.distance.EuclidDistanceUDF'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS cosine_distance") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION cosine_distance AS 'hivemall.knn.distance.CosineDistanceUDF'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS angular_distance") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION angular_distance AS 'hivemall.knn.distance.AngularDistanceUDF'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS jaccard_distance") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION jaccard_distance AS 'hivemall.knn.distance.JaccardDistanceUDF'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS manhattan_distance") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION manhattan_distance AS 'hivemall.knn.distance.ManhattanDistanceUDF'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS minkowski_distance") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION minkowski_distance AS 'hivemall.knn.distance.MinkowskiDistanceUDF'") |
| |
| /** |
| * LSH functions |
| */ |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS minhashes") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION minhashes AS 'hivemall.knn.lsh.MinHashesUDFWrapper'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS minhash") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION minhash AS 'hivemall.knn.lsh.MinHashUDTF'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS bbit_minhash") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION bbit_minhash AS 'hivemall.knn.lsh.bBitMinHashUDF'") |
| |
| /** |
| * voting functions |
| */ |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS voted_avg") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION voted_avg AS 'hivemall.ensemble.bagging.VotedAvgUDAF'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS weight_voted_avg") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION weight_voted_avg AS 'hivemall.ensemble.bagging.WeightVotedAvgUDAF'") |
| |
| /** |
| * misc functions |
| */ |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS max_label") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION max_label AS 'hivemall.ensemble.MaxValueLabelUDAF'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS maxrow") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION maxrow AS 'hivemall.ensemble.MaxRowUDAF'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS argmin_kld") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION argmin_kld AS 'hivemall.ensemble.ArgminKLDistanceUDAF'") |
| |
| /** |
| * hashing functions |
| */ |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS mhash") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION mhash AS 'hivemall.ftvec.hashing.MurmurHash3UDF'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS sha1") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION sha1 AS 'hivemall.ftvec.hashing.Sha1UDF'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS array_hash_values") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION array_hash_values AS 'hivemall.ftvec.hashing.ArrayHashValuesUDF'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS prefixed_hash_values") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION prefixed_hash_values AS 'hivemall.ftvec.hashing.ArrayPrefixedHashValuesUDF'") |
| |
| /** |
| * pairing functions |
| */ |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS polynomial_features") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION polynominal_features AS 'hivemall.ftvec.pairing.PolynomialFeaturesUDF'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS powered_features") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION powered_features AS 'hivemall.ftvec.pairing.PoweredFeaturesUDF'") |
| |
| /** |
| * scaling functions |
| */ |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS rescale") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION rescale AS 'hivemall.ftvec.scaling.RescaleUDF'") |
| |
| 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'") |
| |
| /** |
| * misc functions |
| */ |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS amplify") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION amplify AS 'hivemall.ftvec.amplify.AmplifierUDTF'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS rand_amplify") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION rand_amplify AS 'hivemall.ftvec.amplify.RandomAmplifierUDTF'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS add_bias") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION add_bias AS 'hivemall.ftvec.AddBiasUDFWrapper'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS sort_by_feature") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION sort_by_feature AS 'hivemall.ftvec.SortByFeatureUDF'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS extract_feature") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION extract_feature AS 'hivemall.ftvec.ExtractFeatureUDFWrapper'") |
| |
| 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'") |
| |
| /** |
| * feature converting functions |
| */ |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS conv2dense") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION conv2dense AS 'hivemall.ftvec.conv.ConvertToDenseModelUDAF'") |
| |
| 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_sparse_features") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION to_sparse_features AS 'hivemall.ftvec.conv.ToSparseFeaturesUDF'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS quantify") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION quantify AS 'hivemall.ftvec.conv.QuantifyColumnsUDTF'") |
| |
| /** |
| * feature transformers |
| */ |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS vectorize_features") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION vectorize_features AS 'hivemall.ftvec.trans.VectorizeFeaturesUDF'") |
| |
| 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 indexed_features") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION indexed_features AS 'hivemall.ftvec.trans.IndexedFeatures'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS quantified_features") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION quantified_features AS 'hivemall.ftvec.trans.QuantifiedFeaturesUDTF'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS quantitative_features") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION quantitative_features AS 'hivemall.ftvec.trans.QuantitativeFeaturesUDF'") |
| |
| /** |
| * ftvec/text functions |
| */ |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS tf") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION tf AS 'hivemall.ftvec.text.TermFrequencyUDAF'") |
| |
| /** |
| * fegression 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 train_pa1_regr") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION train_pa1_regr AS 'hivemall.regression.PassiveAggressiveRegressionUDTF'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS train_pa1a_regr") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION train_pa1a_regr AS 'hivemall.regression.PassiveAggressiveRegressionUDTF$PA1a'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS train_pa2_regr") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION train_pa2_regr AS 'hivemall.regression.PassiveAggressiveRegressionUDTF$PA2'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS train_pa2a_regr") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION train_pa2a_regr AS 'hivemall.regression.PassiveAggressiveRegressionUDTF$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$AROWe'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS train_arowe2_regr") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION train_arow_regr AS 'hivemall.regression.AROWRegressionUDTF$AROWe2'") |
| |
| /** |
| * array functions |
| */ |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS float_array") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION float_array AS 'hivemall.tools.array.AllocFloatArrayUDF'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS array_remove") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION array_remove AS 'hivemall.tools.array.ArrayRemoveUDF'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS sort_and_uniq_array") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION sort_and_uniq_array AS 'hivemall.tools.array.SortAndUniqArrayUDF'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS subarray_endwith") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION subarray_endwith AS 'hivemall.tools.array.SubarrayEndWithUDF'") |
| |
| 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 subarray") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION subarray AS 'hivemall.tools.array.SubarrayUDF'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS array_avg") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION array_avg AS 'hivemall.tools.array.ArrayAvgGenericUDAF'") |
| |
| /** |
| * compression functions |
| */ |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS inflate") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION inflate AS 'hivemall.tools.compress.InflateUDF'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS deflate") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION deflate AS 'hivemall.tools.compress.DeflateUDF'") |
| |
| /** |
| * map functions |
| */ |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS map_get_sum") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION map_get_sum AS 'hivemall.tools.map.MapGetSumUDF'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS map_tail_n") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION map_tail_n AS 'hivemall.tools.map.MapTailNUDF'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS to_map") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION to_map AS 'hivemall.tools.map.UDAFToMap'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS to_ordered_map") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION to_ordered_map AS 'hivemall.tools.map.UDAFToOrderedMap'") |
| |
| /** |
| * List functions |
| */ |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS to_ordered_list") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION to_ordered_list AS 'hivemall.tools.list.UDAFToOrderedList'") |
| |
| /** |
| * Math functions |
| */ |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS sigmoid") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION sigmoid AS 'hivemall.tools.math.SigmoidGenericUDF'") |
| |
| /** |
| * mapred functions |
| */ |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS rowid") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION rowid AS 'hivemall.tools.mapred.RowIdUDFWrapper'") |
| |
| /** |
| * misc functions |
| */ |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS generate_series") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION generate_series AS 'hivemall.tools.GenerateSeriesUDTF'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS convert_label") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION convert_label AS 'hivemall.tools.ConvertLabelUDF'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS x_rank") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION x_rank AS 'hivemall.tools.RankSequenceUDF'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS each_top_k") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION each_top_k AS 'hivemall.tools.EachTopKUDTF'") |
| |
| /** |
| * Text processing functions |
| */ |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS tokenize") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION tokenize AS 'hivemall.tools.text.TokenizeUDF'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS is_stopword") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION is_stopword AS 'hivemall.tools.text.StopwordUDF'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS singularize") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION singularize AS 'hivemall.tools.text.SingularizeUDF'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS split_words") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION split_words AS 'hivemall.tools.text.SplitWordsUDF'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS normalize_unicode") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION normalize_unicode AS 'hivemall.tools.text.NormalizeUnicodeUDF'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS base91") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION base91 AS 'hivemall.tools.text.Base91UDF'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS unbase91") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION unbase91 AS 'hivemall.tools.text.Unbase91UDF'") |
| |
| /** |
| * Dataset generator functions |
| */ |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS lr_datagen") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION lr_datagen AS 'hivemall.dataset.LogisticRegressionDataGeneratorUDTFWrapper'") |
| |
| /** |
| * Evaluating functions |
| */ |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS f1score") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION f1score AS 'hivemall.evaluation.FMeasureUDAF'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS mae") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION mae AS 'hivemall.evaluation.MeanAbsoluteErrorUDAF'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS mse") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION mse AS 'hivemall.evaluation.MeanSquaredErrorUDAF'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS rmse") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION rmse AS 'hivemall.evaluation.RootMeanSquaredErrorUDAF'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS auc") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION auc AS 'hivemall.evaluation.AUCUDAF'") |
| |
| /** |
| * 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 train_mf_sgd") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION train_mf_sgd AS 'hivemall.mf.MatrixFactorizationSGDUDTF'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS train_mf_adagrad") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION train_mf_adagrad AS 'hivemall.mf.MatrixFactorizationAdaGradUDTF'") |
| |
| /** |
| * Matrix factorization functions |
| */ |
| |
| 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 train_mf_sgd") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION train_mf_sgd AS 'hivemall.mf.MatrixFactorizationSGDUDTF'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS train_mf_adagrad") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION train_mf_adagrad AS 'hivemall.mf.MatrixFactorizationAdaGradUDTF'") |
| |
| /** |
| * Factorization Machine |
| */ |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS fm_predict") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION fm_predict AS 'hivemall.fm.FMPredictGenericUDAF'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS train_fm") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION train_fm AS 'hivemall.fm.FactorizationMachineUDTF'") |
| |
| /** |
| * Geo Spatial Functions |
| */ |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS tile") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION tile AS 'hivemall.geospatial.TileUDF'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS map_url") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION map_url AS 'hivemall.geospatial.MapURLUDF'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS lat2tiley") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION lat2tiley AS 'hivemall.geospatial.Lat2TileYUDF'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS lon2tilex") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION lon2tilex AS 'hivemall.geospatial.Lon2TileXUDF'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS tilex2lon") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION tilex2lon AS 'hivemall.geospatial.TileX2LonUDF'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS tiley2lat") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION tiley2lat AS 'hivemall.geospatial.TileY2LatUDF'") |
| |
| sqlContext.sql("DROP TEMPORARY FUNCTION IF EXISTS haversine_distance") |
| sqlContext.sql("CREATE TEMPORARY FUNCTION haversine_distance AS 'hivemall.geospatial.HaversineDistanceUDF'") |
| |
| /** |
| * Smile related features |
| */ |
| |
| 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_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'") |
