!      !"#$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_`abcdefghijklmnopqrstuvwxyz{|}~  None6   concraft-plDisamb tiers configuration. concraft-plDisamb module configuration.     None6' concraft-pl0Tagging configuration (everything apart tagset).'()'()None"#N!/ concraft-pl A sentence.3 concraft-pl A sentence.4 concraft-plZNo space, space or newline. TODO: Perhaps we should use a bit more informative data type.8 concraft-pl!A morphosyntactic interpretation.< concraft-plA word.A concraft-plJA segment consists of a word and a set of morphosyntactic interpretations.D concraft-plDInterpretations of the token, each interpretation annotated with a disamb Boolean value (if 5, the interpretation is correct within the context).E concraft-pl2A textual representation of a morphosyntactic tag.F concraft-pl!Select one chosen interpretation.G concraft-pl>Select multiple interpretations and one chosen interpretation.H concraft-plSelect interpretations.I concraft-plRestore textual representation of a sentence. The function is not very accurate, it could be improved if we enrich representation of a space.J concraft-plUse I to translate 3 to a /. concraft-pl3Convert a segment to a segment from a core library.K concraft-pl3Convert a segment to a segment from a core library.L concraft-pl5Convert a sentence to a sentence from a core library.M concraft-pl5Convert a sentence to a sentence from a core library./0123465789:;<=?>@ABCDEFGHIJKLMEABCD<=?>@89:;4657FGH3/012IJKLM None"#+?  concraft-plPrinting configuration. concraft-plPrint weights instead of disamb tags. concraft-pl#Parse the text in the plain format. concraft-pl(Parse the paragraph in the plain format. concraft-pl5Parse paragraph already divided into sentence chunks. concraft-plIdentify empty chunks of text. concraft-pl'Parse the sentence in the plain format. concraft-plShow the plain data. concraft-plShow the paragraph. concraft-plShow the sentence. concraft-plAn infix synonym for . concraft-pl"Tag which indicates unknown words. None"#<8s c concraft-plA pool of Maca instances. concraft-plMaca communication channels. concraft-plOutput channel. concraft-plInput channel. concraft-plRun Maca instance. concraft-plpRun Maca server on given channels. TODO: Should check, if maca works. In particular, if morfeusz is available. concraft-plAnalyse paragraph with Maca. concraft-plNRestore info abouts spaces from a text and insert them to a parsed paragraph.d concraft-plRun Maca server.e concraft-plAnalyse paragraph with Maca. The function is thread-safe. As a pre-processing step, all non-printable characters are removed from the input (based on empirical observations, Maca behaves likewise). concraft-plA weight of a text. concraft-plA weight of a segment. concraft-plA weight of a sentence.d concraft-plNumber of Maca instancescdecdeNone"#N]Jf concraft-pl4A sentence with its original textual representation.j concraft-pl-A sentence. type Sent t = DAG Space (Edge t)k concraft-plA word.o concraft-pl!A morphosyntactic interpretation.q concraft-plThe base form (lemma)r concraft-plThe (morphosyntactic) tagv concraft-plUThe remaining four are ignored for the moment, but we plan to rely on them later on.w concraft-pl2A textual representation of a morphosyntactic tag.x concraft-pl.An almost empty interpretation, with only the r actually specified.y concraft-plSelect interpretations.z concraft-plSelect interpretations.{ concraft-plRestore textual representation of a sentence. The function is not very accurate, it could be improved if we enriched the representation of spaces.| concraft-plUse { to translate j to a f.} concraft-pl A version of {# which places a space on each node. concraft-pl\Pick any path from the given DAG. The result is a list of interleaved node and edge labels.~ concraft-pl7Convert a sentence to a sentence from the core library. concraft-plqConvert a sentence to a sentence from the core library. packSentO :: P.Tagset -> SentO Tag -> X.SentO Word P.Tag4657fgihjklmnopqvrstuwxyz{|}~wklmnopqvrstux4657zyjfgih{}|~None"#1$ concraft-plTraining configuration. concraft-plTagset. concraft-plSGD parameters. concraft-plStore SGD dataset on disk. concraft-pl$Numer of guessed tags for each word. concraft-pl parameter. concraft-pl parameter concraft-pl-Extract only visible features for the guesser concraft-plGlobal configuration concraft-plAnnotated sentence. concraft-pl`The sentence after guessing and segmentation (TODO: and annotated with marginal probabilities?) concraft-plDisambiguation markers concraft-pl<Marginal probabilities according to the disambiguation model concraft-pl;Maximal probabilities according to the disambiguation model concraft-plAnnotation config. concraft-pl:How many morphosyntactic tags should be kept for OOV words concraft-pl@Which path picking method should be used. The function takes the  concraft-pl:Default configuration for the guessing observation schema.  concraft-pl>Default configuration for the segmentation observation schema.  concraft-pl@Default configuration for the disambiguation observation schema.  concraft-pl@Default tiered tagging configuration for the segmentation model. concraft-pl6Tag the sentence with guessing marginal probabilities.  concraft-pl<Tag the sentence with disambiguation marginal probabilities. concraft-pl3Tag the sentence with disambiguation probabilities. concraft-pl/Perform guessing -> trimming -> disambiguation. concraft-pl7Tag with the help of a lower-level annotation function. concraft-pl<Annotate with the help of a lower-level annotation function. concraft-pl Tranform each tag into two tags, one with EOS(end-of-sentence)=True, one with EOS=False. The idea behind this transformation is that, at some point, we may want to tag with an EOS-aware model, with no EOS-related information coming from the previous tagging stages.NOTE: this procedure does not apply to OOV words. The motivation: it seems highly unlikely that any OOV word can mark a sentence end.NOTE: this procedure does not apply to segmentation-ambiguous words either. Allowing the tagger to cut on such words might cause losing DAG edges. concraft-pl2Decide if the word should be marked as eos or not. concraft-pl^Determine the weight of the most probable interpretation which satisfies the given predicate. concraft-plfTry to segment the sentence based on the EOS markers. segment :: X.Sent Word Tag -> [X.Sent Word Tag] concraft-pl.Annotate all possibly interesting information. concraft-plwTrain concraft model. TODO: It should be possible to supply the two training procedures with different SGD arguments. concraft-pl:Simplify the tag for the sake of the disambiguation model. concraft-plzSimplify the tag for the sake of the guessing model. TODO: it is also used in the evaluation script, which assumes that  simplifies to a positional tag. The name of the function should reflect this, perhaps, or there should be two separate functions: one dedicated to guesser, one dedicated to evaluation (and other more generic things). concraft-plCompute an average of the list. concraft-plTrimming parameter concraft-pl 5 means that the probability of the tag being EOS is twice as high as that of not being EOS. 1.0 means that EOS will be marked only if its 100% probable. concraft-pl Training data concraft-plEvaluation dataNone"#1]  concraft-plType of probabilities. concraft-pl%Marginals of the disambiguation model concraft-pl-Max probabilities of the disambiguation model concraft-plMarginals of the guessing model concraft-plPrinting configuration. concraft-pl7Which type of probabilities to show (unless suppressed) concraft-plFPrint disamb markers as numerical values instead of probability values concraft-plShow entire data. concraft-plShow the given sentence. concraft-pl!Parse the text in the DAG format. concraft-pl!Parse sentence in the DAG format. concraft-pl"Tag which indicates unknown words. concraft-pl Tail node concraft-pl Head node concraft-plWord None"#6bj concraft-plPortNone"# concraft-plTraining configuration. concraft-plTagset. concraft-plSGD parameters. concraft-plPerform reanalysis. concraft-plStore SGD dataset on disk. concraft-pl$Numer of guessed tags for each word. concraft-pl parameter. concraft-pl:Default configuration for the guessing observation schema. concraft-pl:Default configuration for the guessing observation schema. concraft-pl%Default tiered tagging configuration. concraft-plTag the analysed sentence. concraft-pl-Tag the sentence with marginal probabilities. concraft-plwTrain concraft model. TODO: It should be possible to supply the two training procedures with different SGD arguments. concraft-plUAnalyse the given sentence with Maca. anaSent :: MacaPool -> L.Text -> IO (Sent Tag) concraft-pl Training data concraft-plEvaluation dataee None"#K  concraft-plA request to parse a long text. concraft-plA short request. concraft-plTagging configuration. concraft-pl Tag with marginal probabilities. concraft-plA request with configuration. concraft-plThe actuall request. concraft-plRequest configuration. concraft-plProcess the short request. concraft-plEProcess the long request given the processor for the short request.  None"# concraft-pl&Run a Concraft server on a given port. concraft-pl0Read and process short requests from the socket. concraft-plSubmit the given request.        !"#$%&'()*+,-./011234567889:;<=<>??@ABBCDEFFGHIJKLMNOPQRSTUVWXYZ[\]^_`abcdefghi88j:;BBCE??@AklmnIopLMNqPQRWXVY[\]abcrrstuvwxyz{{|}~I         rrstuvwA~       1 1     @  xA  (concraft-pl-2.0.2-2S7E2W8GW3xLLaC7gaQNZ0NLP.Concraft.PolishNLP.Concraft.Polish.DAGSeg%NLP.Concraft.Polish.DAG.Config.DisambNLP.Concraft.Polish.DAG.Config NLP.Concraft.Polish.MorphosyntaxNLP.Concraft.Polish.Maca$NLP.Concraft.Polish.DAG.Morphosyntax#NLP.Concraft.Polish.DAG.Format.BaseNLP.Concraft.Polish.DAG.ServerNLP.Concraft.Polish.RequestNLP.Concraft.Polish.Server NLP.Concraft.Polish.Format.Plain&concraft-0.11.0-6WbkFDChtFWASuLA5qZixi NLP.Concraftprune loadModel saveModelConcraftNLP.Concraft.DAGSegSetunSetTierCfgwithPoswithEoswithAtts DisambCfgtiersCfg $fToJSONSet $fFromJSONSet$fInterpretSet$fToJSONTierCfg$fFromJSONTierCfg$fInterpretTierCfg$fToJSONDisambCfg$fFromJSONDisambCfg$fInterpretDisambCfg $fGenericSet $fShowSet$fEqSet$fOrdSet$fGenericTierCfg $fShowTierCfg $fEqTierCfg $fOrdTierCfg$fGenericDisambCfg$fShowDisambCfg $fEqDisambCfg$fOrdDisambCfgConfig disambCfg$fToJSONConfig$fFromJSONConfig$fInterpretConfig$fGenericConfig $fShowConfigSentOsegsorigSentSpaceNoneNewLineInterpbasetagWordorthspaceknownSegwordinterpsTagselectselect' selectWMaprestorewithOrigpackSegpackSent packSentO$fBinaryInterp$fFromJSONSpace $fToJSONSpace $fBinarySpace $fBinaryWord$fFromJSONWord $fToJSONWord $fWordWord $fBinarySeg $fShowInterp $fEqInterp $fOrdInterp $fShowSpace $fEqSpace $fOrdSpace $fShowWord$fEqWord $fOrdWord $fShowSeg$fEqSeg$fOrdSegMacaPool newMacaPoolmacaParsent commonness qualifiermetaInfoeos voidInterp selectAnnorestore' TrainConftagsetsgdArgsonDiskguessNumr0 zeroProbLabelguessOnlyVisible globalConfigAnnoSent guessSentdisambs marginalsmaxProbsAnnoConf trimParampickPathguessannoAlltrain simplify4dmb simplify4gsrProbType MarginalsMaxProbsGuessedMarginalsShowCfgprobType numericDisambshowDatashowSent parseData parseSent$fShowProbType $fEqProbType $fOrdProbType$fEnumProbType$fDataProbType ClientCfg$sel:serverAddr:ClientCfg ServerCfg$sel:concraft:ServerCfg$sel:annoCfg:ServerCfg$sel:showCfg:ServerCfgAnswer$sel:dag:AnswerRequest$sel:dag:Request runServer sendRequest$fToJSONRequest$fFromJSONRequest$fFromJSONAnswer$fToJSONAnswer$fGenericRequest$fGenericAnswerreanaLongDocShortPartagProbsrqBodyrqConfshortlong$fBinaryRequest $fBinaryShort $fBinaryLong$fBinaryConfigrunConcraftServersubmitghc-prim GHC.TypesTruepackSeg_ showWsCfg parsePlain parsePara parsePara'isEmpty showPlainshowPara<>GHC.BasemappendignMacaOutInnewMaca runMacaOn doMacaPar restoreSpaces textWeight segWeight sentWeightNLP.Concraft.DAG.Guessr0TguessSchemaDefaultsegmentSchemaDefaultdisambSchemaDefault tiersSegmentdisambdisamb'tagWithannoWith addEosMarkers resolveEOS maxWeightWithsegmentaverage buildInterpNLP.Concraft.Guess tiersDefaultanaSent sockHandler