/ii< dZddlZddlZddlmZmZddlZddlZddl Z ddl m Z ddl m Z ddlmZmZmZddlmZmZddlmZmZmZmZmZmZmZmZdd lmZdd l m!Z!dd l"m#Z#m$Z$dd l%m&Z&m'Z'm(Z(dd l)m*Z*m+Z+ddl,m-Z-ddl.m/Z/m0Z0ddl1m2Z2ddl3m4Z4m5Z5ddl6m7Z7m8Z8m9Z9m:Z:m;Z;ddlZ>ddl?m@Z@ddlAmBZBmCZCmDZDddlEmFZFmGZGe@dZHeZIeHJeIjKjLZMeIjNeMeI_NeIjKeMeI_KeZOeHJeOjKjLZMeOjNeMeO_NeOjKeMeO_KdZPe jQRdeeGeFzdddddddddddddd dddd!ggd"d#ZSd$ZTe jQRd%eGeFzd&ZUGd'd(e ZVd)ZWd*ZXd+ZYd,ZZd-Z[d.Z\d/Z]d0Z^e jQj_d1Z`d2Zae jQj_d3Zbd4Zcdcd6Zdd7Zed8Zfd9Zgd:Zhd;Zid<ZjGd=d>e ZkGd?d@e Zle jQRdAeege jQRdBddge jQRdCddge jQRdDdEdFgdGZmdHZndIZodJZpdKZqdLZrdMZsdNZtdOZudPZvdQZwdRZxdSZye jQRdTeedUVdfeedUVdfee#dfee5dfgdWZze dXe jQRdYeedUZdU[eedUZdU[gd\Z{e jQRd]e9d^d^fe8d_d`fe7d_d^fge dXdaZ|e jQRdYeedUZdU[eedUZdU[gdbZ}dS)dzE Testing for the bagging ensemble module (sklearn.ensemble.bagging). N)cycleproduct)config_context) BaseEstimator) load_diabetes load_irismake_hastie_10_2)DummyClassifierDummyRegressor)AdaBoostClassifierAdaBoostRegressorBaggingClassifierBaggingRegressorHistGradientBoostingClassifierHistGradientBoostingRegressorRandomForestClassifierRandomForestRegressor)_get_n_samples_bootstrap) SelectKBest)LogisticRegression Perceptron) GridSearchCV ParameterGridtrain_test_split)KNeighborsClassifierKNeighborsRegressor) make_pipeline)FunctionTransformerscale)SparseRandomProjection)SVCSVR)"ConsumingClassifierWithOnlyPredict)ConsumingClassifierWithoutPredictLogProba&ConsumingClassifierWithoutPredictProba _Registrycheck_recorded_metadata)DecisionTreeClassifierDecisionTreeRegressor)check_random_state)assert_allcloseassert_array_almost_equalassert_array_equal)CSC_CONTAINERSCSR_CONTAINERSctd}ttjtj|\}}}}t ddgddgddgddgd }dt td td ttg}t|t|D]<\}}td||d d||||=dS)Nr random_state??TF max_samples max_features bootstrapbootstrap_featuresmax_iter) max_depth) estimatorr2 n_estimators)r*ririsdatatargetrr rr(rr!ziprrfitpredict) rngX_trainX_testy_trainy_testgrid estimatorsparamsrAs i/var/www/html/bet.cuttalo.com/ml/venv/lib/python3.11/site-packages/sklearn/ensemble/tests/test_bagging.pytest_classificationrSHs* Q  C'7 4;S((($GVWf :F#'-     D B+++  J!uZ'8'89900       #gw   00z sparse_container, params, methodr3r?Tr7r4r6Fr9r:r;r8r:r;)rI predict_probapredict_log_probadecision_functioncGddt}td}tttjtj|\}}}}||} ||} td |dddd || |} t| || } td |dddd |||} t| ||}t| |t| d | j D}tfd |DsJdS) Nc"eZdZdZfdZxZS)-test_sparse_classification..CustomSVC7SVC variant that records the nature of the training setctt||t||_|SNsuperrHtype data_type_selfXy __class__s rRrHz1test_sparse_classification..CustomSVC.fit- GGKK1   "1ggDOKrT__name__ __module__ __qualname____doc__rH __classcell__rhs@rR CustomSVCr\>EE         rTrqrr1linearovr)kerneldecision_function_shaper5rAr2cg|] }|j SrCrc.0is rR z.test_sparse_classification..s A A AaQ\ A A ArTcg|]}|k SrCrCr{t sparse_types rRr}z.test_sparse_classification..s000Q[ 000rTrC)r!r*rrrDrErFrrHgetattrr,rb estimators_all)sparse_containerrQmethodrqrJrKrLrMrNX_train_sparse X_test_sparsesparse_classifiersparse_resultsdense_classifier dense_resultstypesrs @rRtest_sparse_classificationris4C Q  C'7 di$+C((($GVWf&%g..N$$V,,M))8UKKK  c.'""  8W.77 FFN))8UKKK  c'7  6G,f55f==Mnm<<<~&&K A A#4#@ A A AE 0000%000 1 111111rTctd}ttjddtjdd|\}}}}t ddgddgddgddgd}dt tttfD]=}|D]8}td ||d | || |9>dS) Nr2r1r3r4TFr7rwrC) r*rdiabetesrErFrr r)rr"rrHrI)rJrKrLrMrNrOrArQs rRtest_regressionrs Q  C'7 crcHOCRC0s((($GVWf : #J#'-     D        F  Mys M Mf M M Q Q  gfoooo   rTrctd}ttjddtjdd|\}}}}Gddt }ddddd d d ddd dd dd ddd dg}||}||} |D]} t d|dd| ||} | | } t d|dd| |||} t|d| j D}t| | tfd|DsJt| | dS)Nrrr1c"eZdZdZfdZxZS))test_sparse_regression..CustomSVRr]ctt||t||_|Sr_r`rds rRrHz-test_sparse_regression..CustomSVR.fitrirTrjrps@rR CustomSVRrrrrTrr3r?Tr7r4r6FrUrVr5rwcg|] }|j SrCryrzs rRr}z*test_sparse_regression..sEEE!EEErTcg|]}|k SrCrCrs rRr}z*test_sparse_regression..s444A$444rTrC) r*rrrErFr"rrHrIrbrr,r)rrJrKrLrMrNrparameter_setsrrrQrrrrrs @rRtest_sparse_regressionrs  Q  C'7 crcHOCRC0s((($GVWfC"&   "&   dKK$eLLN"&%g..N$$V,,M AA, ikk  5;  #ng & & +22=AA  Myy{{ M Mf M M S' " " WV__  >** EE'8'DEEE!.-@@@4444e44455555!.-@@@@'AArTceZdZdZdZdS)DummySizeEstimatorc\|jd|_tj||_dSNr)shapetraining_size_joblibhashtraining_hash_rerfrgs rRrHzDummySizeEstimator.fits&gaj$k!nnrTc@tj|jdSr)nponesrrerfs rRrIzDummySizeEstimator.predictswqwqz"""rTN)rkrlrmrHrIrCrTrRrrs2---#####rTrctd}ttjtj|\}}}}t ||}tt dd|||}||||||ksJtt dd|||}||||||ksJttd||}g}|j D]4}|j |j dksJ| |j5tt!|t|ksJdS)Nrr1r4F)rAr8r:r2T)rAr:)r*rrrErFr)rHrscorerrrrappendrlenset)rJrKrLrMrNrAensemble training_hashs rRtest_bootstrap_samplesrs Q  C'7 xS((($GVWf&''++GW==I '))   c'7   ??7G , ,w0P0P P P P P '))   c'7   ??7G , ,x~~gw/O/O O O O O  *<*>*>$OOOSSHM)77 '7=+;;;;;Y56666 s=!! " "c-&8&8 8 8 8 8 8 8rTcVtd}ttjtj|\}}}}t t dd|||}|jD]<}tjj dtj |j dksJ=t t dd|||}|jD]<}tjj dtj |j dksJ=dS)Nrr1r4F)rAr9r;r2r5T) r*rrrErFrr)rHestimators_features_rrunique)rJrKrLrMrNrfeaturess rRtest_bootstrap_featuresr1sE Q  C'7 xS((($GVWf '))    c'7  1FF}"1%8)<)<)B1)EEEEEE'))   c'7  1EE}"1% ((;(;(A!(DDDDDDEErTc td}ttjtj|\}}}}t jdd5tt| ||}tt j | |dt j t|t| |t j||tt#|d  ||}tt j | |dt j t|t| |t j||ddddS#1swxYwYdS) Nrr1ignore)divideinvalidrwr5)axis)rAr2r8)r*rrDrErFrerrstaterr(rHr,sumrWrrexprXrrJrKrLrMrNrs rRtest_probabilityrMs" Q  C'7 4;S((($GVWf Hh 7 7 7  $,..S   #gw    " F8))&11 : : :BGCKK>HHc ,td}ttjtj|\}}}}t t fD]}t|ddd|||}| ||}t||j z dksJd}tj t|5t|d dd|}|||dddn #1swxYwYdS) Nrr1dTrArBr: oob_scorer2皙?{Some inputs do not have OOB scores. This probably means too few estimators were used to compute any reliable oob estimates.matchr5)r*rrDrErFr(r!rrHrabs oob_score_pytestwarns UserWarning) rJrKrLrMrNrAclf test_scorewarn_msgs rRtest_oob_score_classificationrps Q  C'7 4;S((($GVWf-..6&&     #gw   YYvv.. :.//#5555 J \+X 6 6 6 & &## C GGGW % % % & & & & & & & & & & & & & & &%&&s+DD D c td}ttjtj|\}}}}t t ddd|||}|||}t||j z dksJd}tj t|5t t d dd|}|||ddddS#1swxYwYdS) Nrr1rTrrrrr5)r*rrrErFrr)rHrrrrrr) rJrKrLrMrNrrrregrs rRtest_oob_score_regressionrsr Q  C'7 xS((($GVWf '))      c'7 66**J zCN* + +c 1 1 1 1 F  k 2 2 2##+--     '"""##################s?7DD Dctd}ttjtj|\}}}}t t ddd|||}t ||}t| || |dS)Nrr1r5F)rArBr:r;r2) r*rrrErFrrrHr,rI)rJrKrLrMrNclf1clf2s rRtest_single_estimatorrs Q  C'7 xS((($GVWf %''       c'7    $ $Wg 6 6Ddll622DLL4H4HIIIIIrTctjtj}}t}t t |||drJdS)NrY)rDrErFr(hasattrrrH)rfrgbases rR test_errorrsP 9dkqA ! # #D(..221a88:MNNNNNNNrTcttjtjd\}}}}t t dd||}||}|d||}t||t t dd||}||}t||t tddd||}| |}|d| |} t|| t tddd||}| |} t|| dS) Nrr1n_jobsr2r5rrt)rv) rrDrErFrr(rHrW set_paramsr,r!rY) rKrLrMrNry1y2y3 decisions1 decisions2 decisions3s rRtest_parallel_classificationrs'7 4;Q((($GVWf!   c'7    ' 'B q!!!    ' 'Bb"%%%    c'7     ' 'Bb"%%%! E***11 c'7 ++F33J q!!!++F33Jj*555  E***11 c'7 ++F33Jj*55555rTcXtd}ttjtj|\}}}}t t dd||}|d| |}|d| |}t||t t dd||}| |}t||dS)Nrr1rrr5rr?) r*rrrErFrr)rHrrIr,) rJrKrLrMrNrrrrs rRtest_parallel_regressionrs. Q  C'7 xS((($GVWf 5 7 7PQRRRVVH q!!!   & ! !B q!!!   & ! !Bb"%%% 5 7 7PQRRRVVH   & ! !Bb"%%%%%rTctjtj}}d||dk<ddd}tt t |d||dS)Nr5r?)r5r?)rB estimator__Croc_auc)scoring)rDrErFrrr!rH)rfrg parameterss rRtest_gridsearchrsf 9dkqAAa1fI#)&AAJ"355)):yIIIMMaQRSSSSSrTcHtd}ttjtj|\}}}}t ddd||}t|jtsJt tdd||}t|jtsJt tdd||}t|jtsJttjtj|\}}}}tddd||}t|jtsJttdd||}t|jtsJttdd||}t|jtsJdS)Nrr1rr)r*rrDrErFrrH isinstance estimator_r(rrrr)r"rs rRtest_estimatorr(s Q  C(8 4;S((($GVWf!aa@@@DDWgVVH h)+A B BBBB    c'7  h)+A B BBBB aaHHHLLH h): 6 6666(8 xS((($GVWf QQ???CCGWUUH h)+@ A AAAA 5 7 7PQRRRVVH h)+@ A AAAAaa@@@DDWgVVH h)3 / //////rTc2tttdtd}|t jt jt|dj ddj tsJdS)Nr5)kr?)r9r) rrrr(rHrDrErFrstepsr2intrAs rRtest_bagging_with_pipelinerUs!kA&&&(>(@(@AAPQIMM$)T[))) il(,Q/.rs555d!555rTcg|] }|j SrCr1rs rRr}z#test_warm_start..ss111t 111rT)r rrrHrr)r2rfrgclf_wsrB clf_no_wss rRtest_warm_startr ]s bq 9 9 9DAq F++ >&) QUFF   <  8 8 8 1a6{{l*****!luIMM!Q 55f555 6 6#11y111;;      rTc8tdd\}}tdd}||||dt jt 5|||ddddS#1swxYwYdS) Nr<r5rrT)rBrr6r)r rrHrrraises ValueErrorrfrgrs rR$test_warm_start_smaller_n_estimatorsrws bq 9 9 9DAq t < < D; ?EEEc eZdZdZddZdZdS)EstimatorAcceptingSampleWeightz&Fake estimator accepting sample_weightNc0||_||_||_dSzRecord values passed during fitN)X_y_sample_weight_)rerfrgrs rRrHz"EstimatorAcceptingSampleWeight.fits+rTcdSr_rCrs rRrIz&EstimatorAcceptingSampleWeight.predict rTr_rkrlrmrnrHrIrCrTrRr'r's=00,,,,      rTr'ceZdZdZdZdZdS)EstimatorRejectingSampleWeightz&Fake estimator rejecting sample_weightc"||_||_dSr))r*r+rs rRrHz"EstimatorRejectingSampleWeight.fitsrTcdSr_rCrs rRrIz&EstimatorRejectingSampleWeight.predictr.rTNr/rCrTrRr1r1s800      rTr1 bagging_classaccept_sample_weightmetadata_routingr8rg?ctjddd}tjddgd}tjd}d|d<d|d<|rt }nt }|j\}} t|tr%t|| z} n|} t| 5|r|r| d }|||d } | ||| t| j| jD]8\} } tj| | }t|t'| cxkr| ksnJtj| ddgsJ|rg| jj|| fksJ| jj|fksJt1| j|t1| j|t1| j|| jj| | fksJ| jj| fksJt1| j|| t1| j|| : ddddS#1swxYwYdS)Nrrr5rrr6r?renable_metadata_routingTr)r8rB) minlength)rarangereshaperepeatzerosr'r1rrfloatrrrset_fit_requestrHrGrestimators_samples_bincountrisinrr*r+r+r,)r4r5r6r8rfrgrbase_estimatorr n_featuresexpected_integer_max_samplesbaggingrAsamplescountss rR%test_draw_indices_using_sample_weightrJs #r1%%A 1a&"AHSMMMM!M!:799799GIz+u%%3(+;9J9J9L9L+L'M'M$$'2$ 0@ A A A::  P 4 P+;;$;OON-KVWXXX Aq 666"%g&97;V"W"W : : Iw[I>>>Fv;;#g,,NNNN2NNNNNNN77QF++//11 1 1 1# : |)i-DDDDD |)i\9999 a000 a000 8&AAAA!|).JJ-WWWWW |).J-LLLLL aj999 aj9999# : ::::::::::::::::::sFI66I:=I:cd\}}}t||||ksJd\}}}t||||ksJd\}}}t|||dksJd\}}}d}tjt|5t|||t ||zksJ dddn #1swxYwYd\}}}tjt|5t|||dksJ dddn #1swxYwYd}t jd }d d |d }}}tjt|5t|||t || zksJ dddn #1swxYwYt j d }tj 5tj dd\}}}t||||ksJdd|d }}}t|||t || zksJ ddddS#1swxYwYdS)N)rNnot_used)rrrL)rh㈵>Nr5)rgQ?Nz?.+the number of samples.+low number.+max_samples.+as an integerrzI.+the total sum of sample weights.+low number.+max_samples.+as an integerri@BrM)sizererror)rNrr3)rrrrrrrandom default_rnguniformrrwarningscatch_warnings simplefilter)rr8r warning_msgwarning_msg_with_weightsrJs rRtest_get_n_samples_bootstraprY3s,@)I{M #I{M J Ji W W W W,=)I{M KGG;VVVV-;)I{M #I{M J Ja O O O O,:)I{MSK k 5 5 5  ' ; NNRU ) #S S                     -;)I{M k 5 5 5TT' ; NNRSSSSSSTTTTTTTTTTTTTTT T )   " "C,5tS[[i[=X=XM{I k)A B B B  ' ; NNRU -++-- -S S                     GAJJM  " "   g&&&0=- ; $Y ] K K    14S#++3+:O:O ; ' ; NNRU -++-- -S S                         sI;(B00B47B4DDD":F))F-0F-BI..I25I2c\tdd\}}tdd}||||ddd |||t jt 5t|d ddddS#1swxYwYdS) Nrr5rrT)rBrFr)rrrBr)r rrHrrr AttributeErrorrrs rR$test_oob_score_removed_on_warm_startr\ds c : : :DAq d ; ; ;CGGAqMMMNNde"NEEEGGAqMMM ~ & &##\"""##################sB!!B%(B%ctdd\}}ttdddd}|||j|||jksJdS)Nr5rr3T)r8r9rr2)r rrrHrrfrgrGs rRtest_oob_score_consistencyr`qs~ c : : :DAq G ;;q!   '7;;q!+<+<+G G G G G G GrTc^tdd\}}ttdddd}||||j}|j}|j}t|t|ksJt|dt|dzksJ|djj d ksJd}||}||}||} ||dd|f} ||} | j } | | | | j } t| | dS) Nr^r5rr3F)r8r9r2r:rr?r|) r rrrHrArrrdtypekindcoef_r,)rfrgrGestimators_samplesestimators_featuresrPestimator_indexestimator_samplesestimator_featuresrArKrM orig_coefs new_coefss rRtest_estimators_samplesrls` c : : :DAq G KK1!4!6$J ! " "c*oo 5 5 5 5 !!$ % %Q1 4 4 4 4 a & +s 2 2 2 2O*?;,_=?+I"#QQQ(:%:;G!"GJ MM'7###Ij)44444rTcFt}|j|j}}tt dt }t |dd}||||jdj ddj }|jd}|j d}|j d}||dd|f} ||} || | t|j ddj |dS)Nr?) n_componentsr3r)rAr8r2rr5)rrErFrr rrrHrrrdcopyrArr-) rDrfrg base_pipelinerpipeline_estimator_coefrAestimator_sampleestimator_featurerKrMs rR%test_estimators_samples_deterministicrts ;;D 9dkqA!A...0B0D0DM mST U U UCGGAqMMM!oa06r:1=CHHJJ"I.q103!"AAA'8$89G !G MM'7###yr*1-35LMMMMMrTcd}td|zd\}}tt|dd}||||j|ksJdS)Nrr?r5rr3)r8r9r2)r rrrH _max_samples)r8rfrgrGs rRtest_max_samples_consistencyrwsyK a+oA F F FDAq G  KK1  ; . . . . . .rTc^d}dgdgdggdz}gddz}gddz}gddz}td| ||j}td| ||j}td| ||j}||g||gksJdS) Nrrrr5)ABC)rrr5)rr5r?T)rr2)rrHr)r2rfY1Y2Y3x1x2x3s rR!test_set_oob_score_label_encodingrsL sQC1A 1 B aB QBD|DDD Q  D|DDD Q  D|DDD Q  8Bx      rTcd|dd}d|tj|<|S)Nr?T)ror)astyperisfinite)rfs rRreplacers0 t$$AAr{1~~o HrTc tjgdgddtjdgdtjdgdtj dgg}tjgdtjgdgdgdgdgdgg}|D]`}t }t t t|}||| |t|}||| |}|j |j ksJt }t |}tj t5|||dddn #1swxYwYt|}tj t5|||dddn #1swxYwYbdS)Nr5rrr?Nr?r)r?rrrr)r?r5 )rr)rarraynaninfr)rrrrHrIrrrr r)rfy_valuesrg regressorpipelinebagging_regressory_hats rR*test_bagging_regressor_with_missing_inputsrs]  II LL N N O    A !!          H(()++  !4W!=!=yII Q""1%%%,X66!%%a++33A66w%+%%%%*++  ++ ]: & &   LLA                  ,X66 ]: & & ( (  ! !!Q ' ' ' ( ( ( ( ( ( ( ( ( ( ( ( ( ( (((s$FF F 3GG G c tjgdgddtjdgdtjdgdtj dgg}tjgd}t }t t t|}||| |t|}|||| |}|j |j ksJ| || |t }t |}tjt 5|||dddn #1swxYwYt|}tjt 5|||ddddS#1swxYwYdS)Nrrr?r)rrrrr)rrrrr(rrrrHrIrrrXrWrr r)rfrg classifierrbagging_classifierrs rR+test_bagging_classifier_with_missing_inputsrsB  II LL N N O    A !!A'))J099:FFH LLAq!!!*8441a    & &q ) )E 7ek ! ! ! !((+++$$Q'''())JZ((H z " " Q*844 z " "%%q!$$$%%%%%%%%%%%%%%%%%%s$!FF F7GG"Gctjddgddgg}tjddg}ttdd}|||dS)Nr5r?rr6rg333333?)r9r2)rrrrrHr_s rRtest_bagging_small_max_featuresr:si 1a&1a&!""A !QA 2 4 43UVWWWG KK1rTcttj|}|dd}tjd}Gddt }t |dd}|||t|j dj |j ddS)N r6ceZdZdZdZdS)8test_bagging_get_estimators_indices..MyEstimatorz7An estimator which stores y indices information at fit.c||_dSr_)_sample_indicesrs rRrHz.MyEstimator.fitPs#$D rTN)rkrlrmrnrHrCrTrR MyEstimatorrMs)EE % % % % %rTrr5r)rArBr2) rrQ RandomStaterandnr;r)rrHr-rrrA)global_random_seedrJrfrgrrs rR#test_bagging_get_estimators_indicesrDs )   2 3 3C "aA " A%%%%%+%%% [[]]QR S S SCGGAqMMMsq)93;RST;UVVVVVrTzbagging, expected_allow_nanr5r=cN|jj|ksJdS)z*Check that bagging inherits allow_nan tag.N)__sklearn_tags__ input_tags allow_nan)rGexpected_allow_nans rRtest_bagging_allow_nan_tagrYs.  # # % % 0 :>P P P P P P PrTr8modelr)rArBcZ|tjtjdS)zAMake sure that metadata routing works with non-default estimator.NrHrDrErFrs rR"test_bagging_with_metadata_routingrks" IIdi%%%%%rTzsub_estimator, caller, calleerIrXrWctjddgddgddgg}gd}dgd}}t}||}d |zd z} t|| d d t | } | ||t| |tjddgddgddgg||t |sJ|D]}t|||||dS)aTest that metadata routing works in `BaggingClassifier` with dynamic selection of the sub-estimator's methods. Here we test only specific test cases, where sub-estimator methods are not present and are not tested with `ConsumingClassifier` (which possesses all the methods) in sklearn/tests/test_metaestimators_metadata_routing.py: `BaggingClassifier.predict()` dynamically routes to `predict` if the sub-estimator doesn't have `predict_proba` and `BaggingClassifier.predict_log_proba()` dynamically routes to `predict_proba` if the sub-estimator doesn't have `predict_log_proba`, or to `predict`, if it doesn't have it. rr?r5r6r)r5r?ra)registryset__requestT)rmetadatarr)rfrr)objrparentrrN)rrr&rrrHrr') sub_estimatorcallercalleerfrgrrrrAset_callee_requestrGs rR3test_metadata_routing_with_dynamic_method_selectionr|sI0 1a&1a&1a&)**A A c38M{{H x000I&:5*GI)**MMMM)444G KK1GGV (QFQFQF+ , ,# x==  '        rTcZ|tjtjdS)z^Make sure that we still can use an estimator that does not implement the metadata routing.Nrrs rR-test_bagging_without_support_metadata_routingrs" IIdi%%%%%rT)r)~rnr"rT itertoolsrrrnumpyrrsklearnr sklearn.basersklearn.datasetsrrr sklearn.dummyr r sklearn.ensembler r rrrrrrsklearn.ensemble._baggingrsklearn.feature_selectionrsklearn.linear_modelrrsklearn.model_selectionrrrsklearn.neighborsrrsklearn.pipelinersklearn.preprocessingrrsklearn.random_projectionr sklearn.svmr!r"%sklearn.tests.metadata_routing_commonr#r$r%r&r' sklearn.treer(r) sklearn.utilsr*sklearn.utils._testingr+r,r-sklearn.utils.fixesr.r/rJrD permutationrFrNpermrErrSmark parametrizerrrrrrrrrrrr thread_unsaferrrrr rrrrr r%r'r1rJrYr\r`rlrtrwrrrrrrrrrrrCrTrRrs  $$$$$$$$  """"""&&&&&&GGGGGGGGGG99999999                    ?>>>>>111111????????QQQQQQQQQQGGGGGGGG******<<<<<<<<<<<<<< GFFFFFFF,,,,,, ?>>>>>>>y{{ t{'(( IdO k$  =?? x+,, d# /$'000B& G' # !!&*    # !!&*    U$ O Od% P P ONN%.&2&2/.&2R8+^n-LMM5A5ANM5Ap########'9'9'9TEEE8    F"&"&"&J!#!#!#HJJJ(OOO&6&6&6V&&&4 T T T)0)0)0XCCC4444$&&&&333,777@      ]          ]    +;=N*OPP/%??+eT];;S 22/:/:32<;@?QP/:d. . . b # # # H H H%5%5%5PNNN6 / / /   4   &(&(&(R%%%@WWW*!  991EEE F FM  77CCC D DdK  --// 0 0%8  ##%% %( QQQ--- ,,!<<<1    ++;;;!     &&  .-& # /IF 5   ,-@)L  ---# # .-  # T ((a888    #4#4!#D#D#DSTUUU   &&  &&&rT