o iE@sddlZddlZddlmZddlmZddlmZmZm Z m Z m Z m Z m Z mZmZmZmZddlZddlZddlZddlmZmZmZddlmZddlmZdd lmZdd lm Z dd l!m"Z"m#Z#m$Z$dd l%m&Z&dd l'm(Z(ddl)m*Z*ddl+m,Z,ddl-m.Z.m/Z/ddl0m1Z1ddlm2Z2m3Z3ddl4m5Z6ddl7m8Z8m9Z9m:Z:m;Z;mZ>m?Z?dZ@dZAdZBdZCdZDdZEdZFed?ZLd@edAedBdfdCdDZMdEe e.dFe eKdGedHeIdBeeKeff dIdJZNeddKe eKdLed)dBeKfdMdNZOedKe eeKePfdLeddBeKfdOdNZO )ddKee eKe eeKePffdLeIdBeKfdPdNZOddQeePeQfdReIdBeKfdSdTZR U Vdde e.dWeKdXedYdZe eKdBePf d[d\ZSdGed]eKdBe eKfd^d_ZTdGedBeeKe eKffd`daZUdbe dBeQfdcddZVdeeeKeQfdfedBeQfdgdhZWddieIdBefdjdkZXdledmedBeQfdndoZYdpe edKe eKdBe efdqdrZZdEe e.dseeKefdZeKdBeeKeffdtduZ[dveeKeffdwdxZ\eFfdyedBeePeQffdzd{Z]d|eePeQfd}ePdBeePeQffd~dZ^dS)N)Counter)Path) AnyDictIterableListOptionalSequenceSetTupleUnioncastoverload)MESSAGESPrintermsg)util)Literal)Language) Morphology) MorphologizerSpanCategorizer TrainablePipe) EditTrees)nonproj) DELIMITER)ConfigSchemaTraining)Exampleremove_bilu_prefix)get_sourced_components)registryresolve_dot_names)Mode)ArgOpt_format_numberapp debug_cli import_codeparse_config_overridesshow_validation_error2diZdataT)allow_extra_argsignore_unknown_options)context_settings debug-data)r4hidden.zPath to config file)helpexists allow_dashz --code-pathz--codez-czNPath to Python file with additional code (registered functions) to be imported)r7Fz--ignore-warningsz-IWz+Ignore warnings, only show stats and errorsz --verbosez-Vz-Print additional information and explanationsz --no-formatz-NFzDon't pretty-print the resultsctx config_path code_pathignore_warningsverbose no_formatcCs@|jjdkr tdt|j}t|t|||||dddS)a Analyze, debug and validate your training and development data. Outputs useful stats, and can help you find problems like invalid entity annotations, cyclic dependencies, low data labels and more. DOCS: https://spacy.io/api/cli#debug-data r5zThe debug-data command is now available via the 'debug data' subcommand (without the hyphen). You can run python -m spacy debug --help for an overview of the other available debugging commands.Fconfig_overridesr=r>r?silentN)commandnamerwarnr+argsr* debug_data)r:r;r<r=r>r? overridesrIH/home/ubuntu/.local/lib/python3.10/site-packages/spacy/cli/debug_data.pydebug_data_cli?s   rKr@rArBcIst|| |d}t|"tj||d}t|j}tj|dt d} Wdn1s1wYt |} | dfdd| D} j } fddj D} | d | d | d g}t ||\}fd d |dt}t|}|dt|| ddt|| dd}t|| dd}d}|d}| d| d|dj|dd| | r|dd| r|dd|t|d|t|dt|s|dt||}|r||dn|d| sFt|tkrFdt|d }t|tkr3||n|||jd!td"td |d#| d$d%}||d&td'd(d)d*krqd)}||d+|d)d*kr|d)}||d,d'd-}|jd.t|dd/|d#tjjr jjj t!j"kr|d0tjjd1jj#d2jjj$d3jjj%d4 nG|tjjd5jjj&d6jj#d7t'd8(}|d9)|d:|d%|jd;)td8d-dd/|d#n|d<d=| vsd>| vrt*}d}d}| d?|j+|d@dAgddB|jdC|d#d=,D]\}} |jdD|dt| ,dd/|d#q=fdEdF|-D}!|!,D]=\}} | ,D]d\}"}#||-v}$|$r|"||vr|dG|"dH|dI|#t.kr|dJ|"dK|dL|#d d}|/dMt0||"d=|}%Wdn 1swY|%d*kr|dN|"dOd}qn|/dPt1||}&Wdn 1swY|dQ|dO|dRt2|&t3dS|}'t4|'t5dT}(|dUt5dVt6|(-dW|&dXdY|&dZd[t7|(d\ |jd]t7|'|d#|&d^t8krG|d_n|d`|&da(})t'|)t9}*dbd|*d-D}+|jdc)t|+|d#|&ddt:kr{|den|df|&dg(},t'|,t9}-dhd|-d-D}.|jdi)t|.|d#qe|r|jdjt.dk|d#n|dl|r|jdm|d#n|dndo| vrt;dpdqdoD}/do}0t<do}1d}d}d}2d}3| dr|t|1ds|0d3}4||4dt|/D]}"t|"d*kr|duqdvd|0D}5t|5dd/}5|jdw|5|d#|1|/}6|6rA|dxt|6dydzrR|dzd{d}2|/D]F}"|0|"t.kr|dJ|"dL|0|"d d}|/dMt0||"do}%Wdn 1swY|%d*kr|dN|"dOd}qTd|r|d|d}d}3|s|dl|s|dn|2s|d~|3s|d|r|jdt.dk|d#|r|jd|d#|2r|dd| vr| dt<d}/|dt|/ds|jdt|/|d#|/t;d}6|6r(|dxt|6dyt;dt;|dkrI|dtddt|ddyt|/dkrU|ddd*ksc|dd*krh|ddd*krt|d|dd*kr|dd| vr| dt<d}/|dt|/ds|jdt|/|d#|/t;d}6|6r|dxt|6dyt;dt;|dkr|dtddt|ddydd*ks|dd*kr|ddd*kr|dd*kr|dn|d|dd*kr|dd| vr| dt=d,\}7}8|t|7dt>?|8}9|9|9'}9|9 t>@|9't>@t|7}:||:dt<d}1t;|7}/|1|/}6|6rr|dxt|6dytddd/}5|j|5|d#d| vr| ddddD}7t<d}1|t|7dt;|7}/|1|/}6|6r|dxt|6dytddd/}5|j|5|d#d| vrSd}| d|dddd%ddddtd};|;dkr|d|;dddddD}|dd*kr5|d}?|d|?d|dd*krI|d}?|d|?d|t|=d|t|<dt|ddd/}5|j|5|d#|dD]}"|d|"tAkr|dJ|"dL|d|"d d}qrg}@dD]}"d|"tAkrtB|"vr|@C|"dd|"qt|@d*kr|dt|@d|jdd|@|d#d}t;|<t;|>r|jddt;|<t;|>|d#t;|>t;|<r |jddt;|>t;|<|d#|r|jdtAd|d#t|ddkr/|dd|dddd*krA|ddddd*krS|dddd| v r,| d¡d}A|d}B|t|Adĝ|t|Bdŝ|B|A}Ct|Cd*krt|Ct|B}D|t|Cd|Dd:ddǝn|dȡdd*krd}E||Edʝ|dd*kr|d}E||Ed˝dd*kr܈d}E||Ed͝n|dΡ|dd*kr|d}E||Edϝn|dСdd*k rd}E||Edҝn|dӡ|dd*k r'|d}E||Edԝn|dա| d֡|jDtEjF}F|jDtEjG}G|jDtEjH}H|F rX||Fd|Fdk rRdndٛdڝ|G rl||Gd|Gdk rgdndܛ|H r||Hd|Hdk r{dndޛtIJddSdS)N)no_printprettyr=)rHtraining)schemafrozen_componentscsg|]}|vr|qSrIrI).0p)rPrIrJ zdebug_data..csg|]}|jqSrI get_pipe_metafactory)rQpipenlprIrJrSrTzData file validation train_corpus dev_corpuscsSNrIrI)rZr[rIrJszdebug_data..z%Pipeline can be initialized with datazCorpus is loadableT) make_projFtextszTraining statsz Language: zTraining pipeline: , z!Components from other pipelines: zFrozen components: z training docsz evaluation docszNo evaluation docsz* training examples also in evaluation dataz/No overlap between training and evaluation dataz0Low number of examples to train a new pipeline ()z!It's recommended to use at least z examples (minimum )showzVocab & Vectorsn_wordsz total word(s) in the data (wordsz unique)n_misaligned_wordsrz' misaligned tokens in the training dataz" misaligned tokens in the dev data z10 most common words: )countszfloret vectors with z vectors, z dimensions, -z char n-gram subwordsz vectors (z unique keys, z dimensions)words_missing_vectorsz3{} words in training data without vectors ({:.0f}%)r/z(10 most common words without vectors: {}z&No word vectors present in the packagespancatspancat_singlelabelzSpan Categorizationz Spans KeyLabels)headerdividerzLabel counts in train data: zKey: csi|] }|d|qS)rkrI)rQ spans_key)gold_train_datarIrJ s zdebug_data..zLabel 'z-' is not present in the model labels of key 'z*'. Performance may degrade after training.z"Low number of examples for label 'z ' in key '' (zAnalyzing label distribution...z)No examples for texts WITHOUT new label ''z!Obtaining span characteristics...z$Span characteristics for spans_key 'z8SD = Span Distinctiveness, BD = Boundary Distinctiveness spans_length) thresholdzOver z % of spans have lengths of 1 -- z (min= min_lengthz, max= max_lengthz%). The most common span lengths are: z. If you are using the n-gram suggester, note that omitting infrequent n-gram lengths can greatly improve speed and memory usage.z#Full distribution of span lengths: avg_sdz5Spans may not be distinct from the rest of the corpusz.Spans are distinct from the rest of the corpusp_spanscSg|]\}}|qSrIrIrQw_rIrIrJrS=z10 most common span tokens: {}avg_bdzjszdebug_data..zNamed Entity Recognitionz label(s)z) missing value(s) (tokens with '-' label)zEmpty label found in train datacSs g|] \}}|dkr||fqS)rirI)rQrcountrIrIrJrS{s zLabels in train data: z|Some model labels are not present in the train data. The model performance may be degraded for these labels after training: .ws_entsz invalid whitespace entity spansboundary_cross_entsz, entity span(s) crossing sentence boundarieszTo train a new entity type, your data should include at least zuTraining data should always include examples of entities in context, as well as examples without a given entity type.z`Entity spans consisting of or starting/ending with whitespace characters are considered invalid.textcatz'Text Classification (Exclusive Classes)zText Classification: zLabels: catszWPotential train/dev mismatch: the train and dev labels are not the same. Train labels: z. Dev labels: rzThe model does not have enough labels. 'textcat' requires at least two labels due to mutually-exclusive classes, e.g. LABEL/NOT_LABEL or POSITIVE/NEGATIVE for a binary classification task.n_cats_bad_valueszMUnsupported values for cats: the supported values are 1.0/True and 0.0/False.n_cats_multilabelzThe train data contains instances without mutually-exclusive classes. Use the component 'textcat_multilabel' instead of 'textcat'.zThe dev data contains instances without mutually-exclusive classes. Use the component 'textcat_multilabel' instead of 'textcat'.textcat_multilabelz Text Classification (Multilabel)zPotential train/dev mismatch: the train data contains instances without mutually-exclusive classes while the dev data contains only instances with mutually-exclusive classes.zThe train data contains only instances with mutually-exclusive classes. You can potentially use the component 'textcat' instead of 'textcat_multilabel'.zTrain/dev mismatch: the dev data contains instances without mutually-exclusive classes while the train data contains only instances with mutually-exclusive classes.taggerzPart-of-speech Taggingtagsz label(s) in train dataz is the normalised label entropy morphologizerzMorphologizer (POS+Morph)cSg|]}|qSrIrIrrIrIrJrS5morphsparserzDependency ParsingzFound n_sentsz' sentence(s) with an average length of z.1fz words.g?zThe training data contains z.2fz sentences per document. When there are very few documents containing more than one sentence, the parser will not learn how to segment longer texts into sentences.cSrrIrIrrIrIrJrSZrdepscSrrIrIrrIrIrJrS[scSrrIrIrrIrIrJrS^r n_nonprojz nonprojective train sentence(s)z nonprojective dev sentence(s)z% label(s) in projectivized train dataz: zLow number of examples for z label(s) in the projectivized dependency trees used for training. You may want to projectivize labels such as punct before training in order to improve parser performance.z3Projectivized labels with low numbers of examples: z7The following labels were found only in the train data:z5The following labels were found only in the dev data:z5To train a parser, your data should include at least z instances of each label.rootsr$zMultiple root labels (zq) found in training data. spaCy's parser uses a single root label ROOT so this distinction will not be available.z. nonprojective projectivized train sentence(s)n_cyclesz, projectivized train sentence(s) with cyclestrainable_lemmatizerzTrainable Lemmatizerlemmatizer_treesz. lemmatizer trees generated from training dataz) lemmatizer trees generated from dev dataz lemmatizer trees (z7% of dev trees) were found exclusively in the dev data.z/All trees in dev data present in training data.n_low_cardinality_lemmasz) training docs with 0 or 1 unique lemmas.z$ dev docs with 0 or 1 unique lemmas.no_lemma_annotationsz) training docs with no lemma annotations.z)All training docs have lemma annotations.z$ dev docs with no lemma annotations.z$All dev docs have lemma annotations.partial_lemma_annotationsz. training docs with partial lemma annotations.z2All training docs have complete lemma annotations.z) dev docs with partial lemma annotations.z-All dev docs have complete lemma annotations.Summary checkchecksz passedwarningwarningserrorerrors)Krr,r load_configload_model_from_configconfig interpolater!resolverr pipe_namesror" initializegoodlist _compile_goldtextlangjoinlenfail intersectionrEBLANK_MODEL_THRESHOLDBLANK_MODEL_MIN_THRESHOLDinfo most_common_format_labelsvocabvectorsmode VectorsModefloretvectors_lengthminnmaxnn_keyssumvaluesformat_get_labels_from_spancattableitemskeysNEW_LABEL_THRESHOLDloading_get_examples_without_label_get_span_characteristics_print_span_characteristics_get_spans_length_freq_dist_filter_spans_length_freq_dist SPAN_LENGTH_THRESHOLD_PERCENTAGEmax _format_freqsSPAN_DISTINCT_THRESHOLDrBOUNDARY_DISTINCT_THRESHOLDset_get_labels_from_modelzipnumpyarraylog2DEP_LABEL_THRESHOLDrappendrhrGOODWARNFAILsysexit)Ir;rAr=r>r?rBrcfgrTsourced_componentsresume_componentspipeline factory_names dot_namesr\ train_dataset dev_datasetgold_train_unpreprocessed_data gold_dev_data train_texts dev_textsoverlaprrd n_misalignedmost_common_wordsn_missing_vectorsmodel_labels_spancathas_low_data_warninghas_no_neg_warningrp data_labelsdata_labels_in_componentrrspans_key_in_modelneg_docsspan_characteristics _span_freqs_filtered_span_freqsrzall_span_tokensmost_common_spansrall_span_bound_tokensmost_common_boundslabels label_counts model_labelshas_ws_ents_errorhas_boundary_cross_ents_warningmissing_valueslabels_with_countsmissing_labels label_listrhrR norm_entropy sents_per_doc labels_trainlabels_train_unpreprocessed labels_devrrare_projectivized_labels trees_train trees_dev dev_not_trainpctn good_counts warn_counts fail_countsrI)rPrqrZr[rJrGjs,                                                                   "                            $""rG file_pathrreturncCs|jd}|jdkr2|d|d t|}Wdn1s#wY|d||S|jdkr_|d|d t|}Wdn1sPwY|d||S|jd|jdd d dS) Nz.jsonzLoading z...zLoaded z.jsonlzCan't load file extension zExpected .json or .jsonlr$)exits)partssuffixrsrsly read_jsonr read_jsonlr)rr file_namer1rIrIrJ _load_files$       r(examplesrrZr_c$ Csidtdtdtdtdtdtdtdtd td td td d dd dd dd dtdd d d d d ttd d d d }d|vrYt|jj}|D]p}|j}|j}dd|D} |d| |dt | 7<|j } |D]} | j rq| j j| jdkr|dd7<q|d|jt |jjrdd|DD]} |jj| |jjvr|d| gqd|vr'|} t|D]R\}}|durq|dr||jr|d d7<|drt|}|d|d7<| |r|dr|dd7<q|dkr&|ddd7<qd|vs1d|vr^t|jjD]#}||dvrJt|d|<t|jj|D]\}}|jdur^qR|d||jd7<qR||d vr{t|d |<|j|D],}|jdurq|j|d |vrg|d ||j<|d ||jt |q||d vrt|d |<|j|D]"}|j|d |vrg|d ||j<|d ||j|qd}||d vrt|d |<|j|D]c}|j|d |vrggd|d ||j<t|D]E}|j|d}|d kr6|d ||jd |||d|j |d}|t |krX|d ||jd!||d|qqq9d"|vshd#|vr|d|j!t"d$d%|j!#Dr|d&d7<t|j!#$ddkr|d'd7<d(|vr|j%d)d*d+}|dd,d|Dd-|vr |j%d.d*d+}|j%d/d*d+}t&||D]?\}}|dus|durڐq|d0kr|d0krqt'(|}|r||t)j*<|jjj|jjj+|}|d|gqd1|vrc|j,|d2\}}|dd3d|Dtt&||D]\}\}} | |krE|d|g|dd7<q*t-.|rU|d4d7<t-/|rc|d5d7<d|vrt0d6d%|Dr{|d7d7<q[t"d8d%|Dr|d9d7<t}!|D]#} | j1d kr|!| j1|| j| j2}"|3|"}#|d:|#qt |!d;krt |dkr|d<d7<q[|S)=Nrrrrrrerrkruspans_per_type sb_per_typerrrrdrfrjr) rrrrr`rrrrrcSg|]}|jqSrIrrQxrIrIrJrS6z!_compile_gold..r$r`cSr,rIr-)rQtrIrIrJrSAr0)B-U-L-)r2r3)zI-r4rirl)startendr5r6rrcss|]}|dvVqdS))rr$NrI)rQvalrIrIrJrz _compile_gold..rrrTAGT) as_stringcSg|]}|dur|qSr]rIr.rIrIrJrSrTrPOSMORPHr) projectivizecSr;r]rIr.rIrIrJrSrTrrcs|]}|jdkVqdSrNlemmarQtokenrIrIrJrrcsr@rArBrDrIrIrJrrFrrrr)4rdictrrrstrings reference predictedupdater alignmentorth_isspacex2ylengthsiaddrrget_aligned_sent_starts enumerateget_aligned_ner startswithis_spacerrspansrlabel_rranger5r6ranyrr get_alignedrr feats_to_dictrPOS_FEAT morphologyget_aligned_parseris_nonproj_treecontains_cycleallrClemma_ tree_to_str)$r)rrZr_r1treeseggolddoc valid_wordsalignrEword sent_startsrQrcombined_labelrpspan window_sizeoffset sb_start_idx sb_end_idxrpos_tagsrposmorph label_dict aligned_heads aligned_depsdephead lemma_settree_idtree_strrIrIrJrsf                             rrrhcCdSr]rIrrhrIrIrJrsrcCrr]rIrrIrIrJrscCsH|rdddtttttf|DSdddttt|DS)NracSs"g|] \}}d|d|dqS)rtrsrbrIrQlcrIrIrJrSs"z"_format_labels..cSsg|]}d|dqS)rtrIrQrrIrIrJrSrT)rr rr strintrrIrIrJrs freqssortcCsL|r tt|}dd|D}dddtttttf|DS)NcSsg|] \}}t||fqSrI)rrQkvrIrIrJrSz!_format_freqs..racSs g|] \}}|d|dqS)z (z%)rIrrIrIrJrSs ) rGsortedrrr rr rfloat)rr_freqsrIrIrJrs rrscr component)rrkrpcCsjd}|D].}|dkrdd|D}|dkr*||jjvr(dd|jj|Dng}||vr2|d7}q|S)NrrcSsg|] }|dvrt|qS)r)rrrIrIrJrSs z/_get_examples_without_label..rkcSr,rI)rY)rQrorIrIrJrSr0r$)rUrIrX)r1rrrprrgrrIrIrJrs r factory_namecsNfddjD}t}|D]}|}t|tsJ||jq|S)Ncs g|] }|jkr|qSrIrUrQ pipe_namerrZrIrJrS z*_get_labels_from_model..)rrget_pipe isinstancerrKr)rZrrrrrXrIrrJrs  rcsffddjD}i}|D]"}|}t|tsJ|j|vr't||j<||j|jq|S)Ncs g|] }|jdvr|qS))rkrlrUrrYrIrJrSrz,_get_labels_from_spancat..)rrrrkeyrrKr)rZrrrrXrIrYrJrs    rrcCs"ttdd|Dt|S)z Compute geometric mean of a listcss|]}t|VqdSr])mathlog)rQrQrIrIrJr%rFz_gmean..)rexpfsumr)rrIrIrJ_gmean#s"rmetric frequenciescs*tfdd|D}|tS)Nc3s |] \}}||VqdSr]rI)rQ span_typevaluerrIrJr)sz_wgt_average..)rrr)rrtotalrIrrJ _wgt_average(sr normalizecstt}|D]}|D]}|jdddd}||d7<q q|r8t|dtfdd|D}|S)z2Get the frequency distribution given a set of Docsz``"z''r$csi|] \}}||qSrIrIrrrIrJrr7sz%_get_distribution..)rrlowerreplacerrr)docsr word_countsrirEr1rIrrJ_get_distribution-srrRqcCs4d}|D]\}}||t|||7}q|S)zHCompute the Kullback-Leibler divergence from two frequency distributionsr)rrr)rRrrrlp_wordrIrIrJ_get_kl_divergence;sr span_datacsfdd|D}t|S)z*Compile into one list for easier reportingcs*i|]gtfddDqS)c3s|] }t|VqdSr])r')rQdrrIrJrFsz._format_span_row...)r)rQrrrJrrEsz$_format_span_row..)rr)rrrrIrrJ_format_span_rowCs  r compiled_goldc s*|d|}dd|d|D}dd|d|D}dd|d|D}d d|d|D}td d|Dd d d d|d|D}dd|d|D} fdd|D} fdd| D} | | ||t|t|t| |t| |t||t||| d S)zObtain all span characteristicsrkcSi|] \}}|t|qSrI)r)rQrrrIrIrJrrRz-_get_span_characteristics..rucSrrIrrQrrXrIrIrJrrVrr*cSg|]}t|qSrI)minrrIrIrJrSZrz-_get_span_characteristics..cSrrI)rrrIrIrJrS[rcSr,rI)rI)rQrgrIrIrJrS^r0TrcSsi|] \}}|t|ddqS)TrrrrIrIrJrr_s cSs*i|]\}}|t|d|dddqS)r5r6Trr)rQrsbrIrIrJrrcsr+ci|] \}}|t|qSrIrrQr freq_distp_corpusrIrJrri crrIrrrrIrJrrmr) sdbdr*rPrwrxryr avg_lengthrrzr)rrrrrrrr) r)rrpr span_lengthr* min_lengths max_lengthsrzrspan_distinctivenesssb_distinctivenessrIrrJrLsF    rrc Csd}tdtdd|dD}|d|d|d|d g}t||dd }|d |d |d g}dgdd|Ddg}tj|||ddgdgt|d|ddS)z+Print all span characteristics into a table)z Span TypeLengthSDBDNcss|]}t|VqdSr]rrrIrIrJrr8z._print_span_characteristics..rrPrrr*)rrrryrz Wgt. AveragecSsg|] }dt|dqS)z{:.2f}r)rround)rQfrIrIrJrSrz/_print_span_characteristics..riTrrr$)footerrnroalignsmax_colN)rrrrr)rheadersr table_datar footer_datarrIrIrJrs0 r length_dictc Csg}|D] \}}||qt}|D]}||r%||d7<qd||<qi}|D]\}} | t|d} t| d} | ||<q0|S)zDGet frequency distribution of spans length under a certain thresholdr$gY@r)rextendrgetrrr) rrvall_span_lengthsr~rPrrQfreq_dist_percentagerr percentagerIrIrJrs     rrrvcCs<d}i}|D]\}}||kr|S|||<||7}q|S)zFilter frequency distribution with respect to a threshold We're going to filter all the span lengths that fall around a percentage threshold when summed. r)r)rrvrfiltered_freq_distrdistrIrIrJrs r)F)T)rr)_rr collectionsrpathlibrtypingrrrrrr r r r r rrr$typerwasabirrrr>rcompatrlanguagerr_rrrrr(pipeline._edit_tree_internals.edit_treesrpipeline._parser_internalsr"pipeline._parser_internals.nonprojrschemasrrNrrtraining.initializer r!r"rr#r_utilr%r&r'r(r)r*r+r,rrrrrrrrCContextboolrKrrGr(rrrrrrrrrrrrrrrrrrIrIrIrJs4  4          (  &    <"    "    6$