Q(hs dZddlZddlZddlZddlZddlZddlmZddl m Z ddl m Z ddl mZmZmZmZddl mZddlmZdd lmZmZmZmZd Zej8gd j:Zej8gd j:ZeejAZ!ed dZ"ed e"ed ded ded dzed ded dzeddzedded dzeddzgZ#e#Dcgc] }|e"k7s | c}Z$ejJjMde#dZ'dZ(ejJjMde$dZ)ejJjMde#dZ*ejJjMde#dZ+ejJjMde$dZ,ejJjMde$dZ-ejJjMde#dZ.ejJjMde#dZ/ejJjMde#d Z0d!Z1ejJjMde#ejJjMd"e!ejdejfdejh#gd$Z5d%Z6d&Z7ejJjMde#d'Z8d(Z9d)Z:ejJjMde$d*Z;d+Zd.Z?d/Z@ejJjMde#d0ZAd1ZBejJjMd2d3ejd4ieDd5feej ejf6d7d8eDd9fgd:ZFd;ZGd<ZHejJjMd=d>d?gejJjMd@gdAdBZIejJjMd=d>d?gejJjMd@gdAdCZJejJjMd@gdDejJjMdEdFdGgdHZKejJjMd@gdDdIZLdJZMGdKdLeZNdMZOycc}w)Nz'Testing for Gaussian process regressionN) approx_fprime)ConvergenceWarning)GaussianProcessRegressor)RBF DotProductExpSineSquared WhiteKernel)ConstantKernel) MiniSeqKernel)assert_allcloseassert_almost_equalassert_array_almost_equalassert_array_lessc2|tj|zS)N)npsin)xs e/var/www/skyplay_api_hub/venv/lib/python3.12/site-packages/sklearn/gaussian_process/tests/test_gpr.pyfr"s rvvay=)?g@g@g@g@g @)@g@g@g@g@rfixed length_scalelength_scale_boundsr)MbP?@@{Gz?Y@h㈵>)r#r"皙?kernelc.tjdkrtjdt |j t t}|jt d\}}t|tttj|dy)N#This test may fail on 32 bit Pythonr%T return_cov) sysmaxsizepytestxfailrfitXypredictr rdiag)r%gpry_predy_covs rtest_gpr_interpolationr98sf {{e :; #& 1 5 5a ;CKKdK3MFE",rc td}gd}tjgd}t|j ||}|j |d\}}t ||dd jd tjt|z jt ||t tj|d y) Nr)baseline_similarity_bounds)ABC)r)Tr* eval_gradientr?r,) r rarrayrr1r4r raveleyelenr5)r%r2r3r6r7r8s r!test_gpr_interpolation_structuredrHEs g >FA A "& 1 5 5a ;CKKdK3MFEq%a(..01rvvc!f~3E2L2L2N",rc"tjdkrtjdt |j t t}|j|jj|j|jkDsJy)Nr'r(r)) r-r.r/r0rr1r2r3log_marginal_likelihoodkernel_thetar%r6s rtest_lml_improvingrNTso {{e :; #& 1 5 5a ;C  & &s{{'8'8 9C  rc:t|jtt}t j |j jjtj}|j|dt|j j|dy)Nr)dtypeF) clone_kernel) rr1r2r3ronesrKrLshapefloat64rJr )r%r6 input_thetas rtest_lml_without_cloning_kernelr[ish #& 1 5 5a ;C''#++++11DK %@ ));:rct|jtt}|j |j j d\}}tjtj|dk|j j |j jdddfk(z|j j |j jdddfk(zsJy)Nr)T-C6?rr?) rr1r2r3rJrKrLrallabsbounds)r%r6lml lml_gradients rtest_converged_to_local_maximumrcss #& 1 5 5a ;C33CKK4E4EtLC 66   $ ;;   2 21a4 8 8 : ;;   2 21a4 8 8 :  rct|jtt}|jj }t j|jjjj}d}||t j|dddfdf<t|dddf|jj|zt|jj|dddf|zy)Nr)绽|=r?r) rr1r2r3rKr`rfinforLrTmaxisfiniter)r%r6r`max_tinys rtest_solution_inside_boundsrks #& 1 5 5a ;C [[  F 88CKK%%++ , 0 0D D,0FBKKq!t % %q ()fQTlCKK$5$5$<=ckk''1)<=rct|jttj |j d\}}t |j fdd}t||dy)Nr)Tc(j|dS)NF)rJ)rLr6s rz#test_lml_gradient..sC$?$?u$MrrerA)rr1r2r3rJrLrr )r%rarblml_gradient_approxr6s @rtest_lml_gradientrps] #& 1 5 5a ;C33FLL$GC' Mu &91=rcpt|}|jtd\}}t|ddt |j j dkDrAttj|tj|j ddyttj|ddy)Nr)Tr*rr?) rr4r2r rGr%rLrr5exp)r%r6y_meanr8s r test_priorrus #& 1CKKdK3MFE1% 3::  q BGGENBFF6<<?,CQGBGGENAq1rct|jtt}|j t d\}}|j t d}t|tj|ddttj|tj|jz tj|dtj|jz dy)Nr)Tr*ir?) rr1r2r3r4X2sample_yr rmeanr5rgvar)r%r6rtr8sampless rtest_sample_statisticsr|s #& 1 5 5a ;CKKtK4MFEll2v&G 3Q7 ++-- wRWWU^//11 rctd}t|djtt}t j |jjdk(sJy)Nrr% optimizer) rrr1r2r3rrsrKrLrMs rtest_no_optimizerrsE XF "&D A E Ea KC 66#++## $ ++ +rtargetrSchtjdkrtjdt |j t t}|jtd\}}|jtd\}}ttjtj||y)Nr'r(r)Tr* return_std)r-r.r/r0rr1r2r3r4rwr rsqrtr5)r%rr6rtr8y_stds rtest_predict_cov_vs_stdrs} {{e :; #& 1 5 5a ;CKKtK4MFEKKtK4MFE/7rctjjd}|jddd}|dddfd|dddfzz}t ddg}t |j ||}tj|jjdtj|jjddzkDsJy) Nrr?)2r@r$rr)rr) rrandom RandomStateuniformrrr1rsrKrL)rngr2r3r%r6s rtest_anisotropic_kernelrs ))   "C B7#A !Q$#!Q$-A #s_F "& 1 5 5a ;C 66#++##A& '"&&1B1B11E*F*J JJ Jrcd\}}tjjd}|j||dzdz }tj|j dtjd|zj dz|j d|z}td d td g|zd g|z ztd dz}tj }tdD]|}t||dj||}|j|jj } | |tj"tj$j&z kDsJ| }~y)N)r@rr@r?axisrAr$)scalesizerr )r]r"rr#)r#$@) noise_levelnoise_level_boundsrr)r%n_restarts_optimizer random_state)rrrrandnrsumnormalr>rr infrangerr1rJrKrLrffloat32eps) n_samples n_featuresrr2r3r%last_lmlrgpras rtest_random_startsrsV"Iz ))   "C )Z(1,q0A q 1  &&Q-    # $ **3Y* / 0 sK 3UZ'k]Z=W$EFFwH %a %!5  #a) (()9)9:X 4 8 88888rcTtjt}tjt}t|z |z }t |}|j t |t |d}|j t t|jtd\}}||z|z}||z}|jtd\}} t||t|| |jtd\} } | |dzz} |jtd\} } t| | y)a Test normalization of the target values in GP Fitting non-normalizing GP on normalized y and fitting normalizing GP on unnormalized y should yield identical results. Note that, here, 'normalized y' refers to y that has been made zero mean and unit variance. r)Tr% normalize_yrr*r@N) rryr3stdrr1r2r4rwr ) r%rtry_normr6gpr_normr7 y_pred_std y_pred_normy_pred_std_norm_r8 y_cov_norms rtest_y_normalizationrsWWQZF FF1IE&jE !F #& 1CGGAv(v4HH LLARD9FJ e^f $Fe#J#+#3#3B4#3#H K , O4{{2${/HAu E1H E$$RD$9MAzz*rcFdtz}ddi}tdi|}t|d}|jt||j t d\}}tjgd}tjgd}t||d d t||d d y )a Here we test that, when noramlize_y=True, our GP can produce a sensible fit to training data whose variance is significantly larger than unity. This test was made in response to issue #15612. GP predictions are verified against predictions that were made using GPy which, here, is treated as the 'gold standard'. Note that we only investigate the RBF kernel here, as that is what was used in the GPy implementation. The following code can be used to recreate the GPy data: -------------------------------------------------------------------------- import GPy kernel_gpy = GPy.kern.RBF(input_dim=1, lengthscale=1.) gpy = GPy.models.GPRegression(X, np.vstack(y_large), kernel_gpy) gpy.optimize() y_pred_gpy, y_var_gpy = gpy.predict(X2) y_pred_std_gpy = np.sqrt(y_var_gpy) -------------------------------------------------------------------------- rrTrr)gy(V.@gV],;g6}~CgHW -@gƠKQ@)g ؼG'@gsr@g{u>5?g ?g.[ ?gQ?r)rtolatolg333333?N) r3rrr1r2r4rwrrDr )y_large RBF_paramsr%r6r7r y_pred_gpyy_pred_std_gpys rtest_large_variance_yr"s21fG!#&J  : F "&d CCGGAwRD9FJKJ XXDNFJT: JTBrctjttdzfj}t d}t |dd}|j ttt |dd}|j t||jtd\}}|jtd\}}|jtd\}} |jtd\}} t||ddd ft||ddd fdz t|jd D]$} t||d | ft| | d | f&|jtd } |jtd } | jdk(sJ| jdk(sJt| | ddd ddftD]}t |d}|j ttt |d}|j ttjttfjt|jj |jj dy)Nr@rrF)r%rrTrr*rr?.rr)rrr)rrr@rr)rvstackr3Trrr1r2r4rwr rrXrxkernelsrKrL)y_2dr%r6gpr_2d y_pred_1dy_std_1d y_pred_2dy_std_2dry_cov_1dy_cov_2dr y_sample_1d y_sample_2ds rtest_y_multioutputrYs 99aQZ " "Dc "F "&De TCGGAqM %VtQV WF JJq$++bT+:Ix ...=Ix++bT+2KAx...5KAx 9QT?3 9QT?Q#67 1 &=HhsF{&;<HhsF{&;<=,,rR,0K//"/3K    '' '    ** * [Aq%9:H&f$G 1 )TJ 1biiA'))*CKK--v~~/C/CQGHrcd}t||}|jtt|j |j j |j |jj kDsJy)Nc ^tjjd}|||d}}tdD]q}tj|j tj d|dddftjd|dddf}||d}||ksn||}}s||fS)NrFrBrr?)rrrr atleast_1drmaximumminimum) obj_func initial_thetar`r theta_optfunc_minrrLrs rrz(test_custom_optimizer..optimizersii##A&+X . 8 r /AMM BJJr6!Q$<8"**QqRSt :UVEe4A8|&+Q8  /(""rr~)rr1r2r3rJrKrLr%)r%rr6s rtest_custom_optimizerrsc # #&I FCGGAqM  & &s{{'8'8 9Ccft|d}t|d}tjttdf}tjt t ddzf}|j ||tjttddzf}tjt t ddzf}|j ||tjddddddf}|j|d \}}|j|d \}} t||t|| y) Nr!rrr?gV瞯$>vRV$>$W!NM n5 ]3rcftddtddz}t}|jtd\}}|jtd\}}t|}|jtd\}}|jtd\}}t ||t ||y) Nrr)constant_value_boundsrTrr*r))r>rrr4r2r)default_kernelgpr1ry_std1y_cov1gpr2y_std2y_cov2s rtest_no_fit_default_predictrss':S >N $ %D Q4 0IAv Q4 0IAv #> :D Q4 0IAv Q4 0IAvff-ff-rc8tddg}t|}d}tjt|5|j t tdddtddgtdd gz}t|}tjd 5}tjd |j t tt|d k(sJt|djtsJ|djj ddk(sJt|djtsJ|djj ddk(sJ dddt#j$t d }tddgddg}t|}tjd 5}tjd |j |tt|d k(sJt|djtsJ|djj ddk(sJt|djtsJ|djj ddk(sJ dddy#1swY%xYw#1swY*xYw#1swYyxYw)Nr#rrr)zThe optimal value found for dimension 0 of parameter length_scale is close to the specified upper bound 0.001. Increasing the bound and calling fit again may find a better value.rrrgj@T)recordalwaysr@rzThe optimal value found for dimension 0 of parameter k1__noise_level is close to the specified upper bound 0.001. Increasing the bound and calling fit again may find a better value.r?zThe optimal value found for dimension 0 of parameter k2__length_scale is close to the specified lower bound 1000.0. Decreasing the bound and calling fit again may find a better value.rrrr"rzThe optimal value found for dimension 0 of parameter length_scale is close to the specified lower bound 10.0. Decreasing the bound and calling fit again may find a better value.zThe optimal value found for dimension 1 of parameter length_scale is close to the specified lower bound 10.0. Decreasing the bound and calling fit again may find a better value.)rrr/warnsrr1r2r3r warningscatch_warnings simplefilterrG issubclasscategoryrargsrtile) r%r6warning_message kernel_sumgpr_sumrX_tile kernel_dimsgpr_dimss rtest_warning_boundsrs dD\ 2F "& 1C  ( @ 1 t = #JAJ'j9G   - h' Aq6{a&),,.@AAA 1I   " "1 %*1 1  1&),,.@AAA 1I   " "1 %*1 1  1% 4WWQ]FC:C:NK'{;H   - h' VQ6{a&),,.@AAA 1I   " "1 %*1 1  1&),,.@AAA 1I   " "1 %*1 1  1%  K  <  s%I6CJ0B)rr@r?)r)rX fill_valuerN).r?)rrWr2rXrYrr1 _y_train_stdr/rPr4r r5rr concatenaterfull) r% y_constantr6r7r8r n_targetsrr3Y_predY_covs rtest_constant_targetr$,sn2::6J "&d CCGGAz   v}}S1 11 1KKdK3MFEFJ'BGGENCd3771:qyI ))   "C  JJYNJ + GG9a.Q 7   AGGAqMKKdK3MFEF1a4L!$BGGE&M*Cd; <rr rrrDr1r4r rdiagonal) r%r6X_trainy_trainrpred1rpred2covs r2test_gpr_consistency_std_cov_non_invertible_kernelr,UsK}m ,s  & 0%&F #&T JChh #J  % #  +   $ * %   Ghh     O     O   GGGGW XX + &  % * %  $  FV5JE3V5JE3CS!12>rzparams, TypeError, err_msgrrzCalpha must be a scalar or an array with same number of entries as yrr@)r%rz#requires that all bounds are finitectdi|}tj||5|jtt dddy#1swYyxYw)z0Check that expected error are raised during fit.rNr)rr/rr1r2r3)params TypeErrorerr_msgr6s rtest_gpr_fit_errorr1sA( # ,V ,C y 0 1 s AActtjtt}d}t j t|5|jddddy#1swYyxYw)z7Check that we raise the proper error in the LML method.r)z.Gradient can only be evaluated for theta!=NonerTrBN) rrr1r2r3r/r ValueErrorrJr6r0s rtest_gpr_lml_errorr5sT "#% 0 4 4Q :C>G z 18 ##$#7888s A&&A/cttjtt}d}t j t|5|jtdddddy#1swYyxYw)z4Check that we raise the proper error during predict.r)z9At most one of return_std or return_cov can be requested.rT)r+rN) rrr1r2r3r/r RuntimeErrorr4r4s rtest_gpr_predict_errorr8sU "#% 0 4 4Q :CIG |7 39 A$4 8999s A,,A5rTFr!)Nr?rctjjd}d\}}}|f}|||fz}|f}| |dkDr||fz}|j||}|j||} |j|} t |} | j || | j | d\} } | j | d\}}| j|k(sJ| j|k(sJ|j|f|zk(sJy) aCheck the shapes of y_mean, y_std, and y_cov in single-output (n_targets=None) and multi-output settings, including the edge case when n_targets=1, where the sklearn convention is to squeeze the predictions. Non-regression test for: https://github.com/scikit-learn/scikit-learn/issues/17394 https://github.com/scikit-learn/scikit-learn/issues/18065 https://github.com/scikit-learn/scikit-learn/issues/22174 )r rVNr?rTrr*)rrrrrr1r4rX)rr!rrn_samples_trainn_samples_test y_train_shape y_test_shaper'rr(modelr7rrr8s rtest_predict_shapesrBs ))   %C29/J$&M% 4 #$LQ#yl2 ii4G YY~z 2Fcii'G $ =E IIgwMM&TM:MFE}}V}5HAu <<< '' ' ;;, && & ;;>+l: :: :rcvtjjd}d\}}d}d}|f}|||fz}| |dkDr|||f}n||f}|j||} |j||} |j|} t |} | j | | | j | |} | j|k(sJy) a&Check the shapes of y_samples in single-output (n_targets=0) and multi-output settings, including the edge case when n_targets=1, where the sklearn convention is to squeeze the predictions. Non-regression test for: https://github.com/scikit-learn/scikit-learn/issues/22175 r:)rr;rVrrNr?r<r)rrrrrr1rxrX)rr!rrr=n_samples_X_testn_samples_y_testr?r@r'rr(rA y_sampless rtest_sample_y_shapesrGs ))   %C"&J$&M% 4 Q()5EF (*:; ii4G YY' 4Fcii'G $ =E IIgwv1ABI ??l ** *r)Nr?r@rArr?rrcPtjjd}|jdd}|jd||nd}t |}|j ||j }|j|||j ||j }||k(sJy)zJCheck the output shape of `sample_y` is consistent before and after `fit`.rrANr?r!r)rrrrrrxrXr1)r!rrr2r3rAshape_before_fitshape_after_fits rtest_sample_y_shape_with_priorrMs ))   %C "aA "9#8ia@A $y 9E~~a9~=CC IIaOnnQ)n<BBO  .. .rc$tjjd}d}|j|d}|j|||nd}t |}|j |d\}}|j |d \}} |j |||j |d\} } |j |d \}} |j| jk(sJ|j| jk(sJ| j| jk(sJy) zIAFA $y 9E!MM!M=J ==t=4LAy IIaO--d-;Ix--d-3KAx   y .. . ??hnn ,, , ??hnn ,, ,rc&tjjd}|jdd}|jdd}t d}t j td5|j||d d d y #1swYy xYw) zmCheck that an error is raised when the number of targets seen at fit is inconsistent with n_targets. rrrAr@r?rJz!The number of targets seen in `y`rN) rrrrrr/rr3r1)rr2r3rAs rtest_n_targets_errorrX+ss ))   "C "aA "aA $q 1E z)L M !Qs +BBceZdZdZdZy) CustomKernelz A custom kernel that has a diag method that returns the first column of the input matrix X. This is a helper for the test to check that the input matrix X is not mutated. c|dddfS)Nrr)selfr2s rr5zCustomKernel.diag?sAwrN)__name__ __module__ __qualname____doc__r5rrrrZrZ8s  rrZcttjtt}t j t}|jtd\}}tt|y)z Check that the input X is not modified by the predict method of the GaussianProcessRegressor when setting return_std=True. Non-regression test for: https://github.com/scikit-learn/scikit-learn/issues/24340 r)TrN) rrZr1r2r3rcopyrwr4r )r6X2_copyrs r#test_gpr_predict_input_not_modifiedrdCsJ #,. 9 = =a CCggbkG ;;rd; +DAqB r)Pr`rr-rnumpyrr/scipy.optimizersklearn.exceptionsrsklearn.gaussian_processr sklearn.gaussian_process.kernelsrrrr r r>4sklearn.gaussian_process.tests._mini_sequence_kernelr sklearn.utils._testingr r rrr atleast_2drr2rwrEr3 fixed_kernelrnon_fixed_kernelsmark parametrizer9rHrNrQr[rcrkrprur|rrWrXrYrrrrrrrrrrrrr$r,zerosr3rr1r5r8rBrGrMrVrXrZrdr)s0rrrs6-   (1= OBMM0133R]],-//aDJJLA SSk:c;#3KPPc;#3KPPkc;#3KPPk +2LV|5KVL7+ -, - -#4567+,7+;,;#45 6 #45 >6 >7+ >, >7+ 2, 27+,",7+Awrwwqwwqz'L#MN8O,8 K67+$+,$+N4Cn+H\#4562 7+4,4*."F R 67+%<,%