!"#$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_`abcdefghijklmnopqrstuvwxyz{|}~  Safe-Inferred8An atomic part of morphosyntactic tag with optional POS.A tier description.#Does it include the part of speech?Tier grammatical attributes.Select tier attributes.Split the positional tag.  None24BCA set with a non-negative weight assigned to each of its elements. 1A sentence with original, textual representation. A sentence.Orthographic form.Out-of-vocabulary (OOV) word.7A segment parametrized over a word type and a tag type.MA word represented by the segment. Typically it will be an instance of the  class.oA set of interpretations. To each interpretation a weight of appropriateness within the context is assigned.Map function over segment tags.Interpretations of the segment.Interpretations of the segment. Map function over sentence tags. Map function over sentence tags.?Make a weighted collection. Negative elements will be ignored.0Map function over weighted collection elements.        None NoneESynchronize two datasets, taking disamb tags from the first one and the rest of information form the second one. In case of differences in token-level segmentation, reference segmentation (token-level) is assumed. Otherwise, it would be difficult to choose correct disamb tags.If both arguments contain only one segment, insert disamb interpretations from the first segment into the second segment. Otherwise, the first list of segments will be returned unchanged.Align two lists of segments.CFind the shortest, length-matching prefixes in the two input lists.NoneHAn analyser performs word-level segmentation and morphological analysis.Reanalyse sentence./From the reference sentence the function takes:Word-level segmentationChosen interpretations (tags)0From the reanalysed sentence the function takes:Potential interpretationsReanalyse paragraph.None$ Configuration of the schema. All configuration elements specify the range over which a particular observation type should be taken on account. For example, the  [-1, 0, 2]_ range means that observations of particular type will be extracted with respect to previous (k - 1 ), current (k) and after the next (k + 2?) positions when identifying the observation set for position k in the input sentence."The 7 schema block.#The 8 schema block.$The 9G schema block. The first list of ints represents lengths of prefixes.%The :G schema block. The first list of ints represents lengths of suffixes.&The ; schema block.'The < schema block.(The = schema block.)The > schema block.* Maybe entry.+Body of configuration entry.-%Range argument for the schema block. .0When true, the entry is used only for oov words./*Additional arguments for the schema block.0A block is a chunk of the Ox computation performed within the context of the sentence and the list of absolute sentence positions.0Record structure of the basic observation types.1{A schema is a block of the Ox computation performed within the context of the sentence and the absolute sentence position.2BThe Ox monad specialized to word token type and text observations.3SAn observation consist of an index (of list type) and an actual observation value.4A dummy schema block.5GSequence the list of schemas (or blocks) and discard individual values.Construct the  structure given the sentence.6Transform a block to a schema depending on * A list of relative sentence positions, * A boolean value; if true, the block computation will be performed only on positions where an OOV word resides.7*Orthographic form at the current position.8*Orthographic form at the current position.9-List of lowercased prefixes of given lengths.:-List of lowercased suffixes of given lengths.;Shape of the word.<Shape of the word.=Packed shape of the word.>Packed shape of the word.? Entry with additional arguemnts.@)Plain entry with no additional arugments.A-Null configuration of the observation schema.B,Build the schema based on the configuration.C9Use the schema to extract observations from the sentence.- !"#$%&'()*+,-./0123456789:;<=>?@ABC$ !"#$%&'()*+,-./0123456789:;<=>?@ABC$32145C+,-./*@? !"#$%&'()AB06789:;<=> !"#$%&'()*+,-./0123456789:;<=>?@ABCNone+DTraining configuration.GSGD parameters.HStore SGD dataset on diskIR0 construction methodJ6Method of constructing the default set of labels (R0).KSee LSee MSee  NA guessing model.0Schematize the input sentence with according to schema rules.RDetermine the k most probable labels for each word in the sentence. TODO: Perhaps it would be better to use sets instead of lists as output?S`Insert guessing results into the sentence. Only interpretations of OOV words will be extended.TCombine R with S. UTrain guesser.Schematized dataset.DEFGHIJKLMNOPQRSTUTraining configuration Training dataEvaluation dataDEFGHIJKLMNOPQRSTUNOPQRSTDEFGHIJMLKU DEFGHIJMLKNOPQRSTUNone VTraining configuration.^A disambiguation model.0Schematize the input sentence with according to schema rules.WUnsplit the complex tag (assuming, that it is one of the interpretations of the word).c)Perform context-sensitive disambiguation.d0Insert disambiguation results into the sentence.eCombine c with d. f'Tag labels with marginal probabilities.gpPrune disamb model: discard model features with absolute values (in log-domain) lower than the given threshold.hTrain disamb model.%Schematized data from the plain file.VWXYZ[\]^_`abcdefghTraining configuration Training dataEvaluation dataResultant modelVWXYZ[\]^_`abcdefgh^_`abfcdeVYWZ[\]X\]hg V YWZ[\]X\]^_`abcdefghNone i Statistics.j!Number of segments in gold corpuskNumber of correct tags Add stats,mAccuracy given stats.nAccuracy weak lower bound.oAccuracy strong lower bound.pAccuracy weak upper bound.qAccuracy strong upper bound.All tags are expanded here.Positive tags. ijklmnopq ijklmnopq ijklmnpoq ijklmnopqNone rConcraft data.x@Save model in a file. Data is compressed using the gzip format.yLoad model from a file.zTag sentence using the model. In your code you should probably use your analysis function, translate results into a container of ences, evaluate z` on each sentence and embed the tagging results into the morphosyntactic structure of your own.)The function returns guessing results as = elements of the output pairs and disambiguation results as & elements of the corresponding pairs.{Determine marginal probabilities corresponding to individual tags w.r.t. the disambiguation model. Since the guessing model is used first, the resulting weighted maps corresponding to OOV words may contain tags not present in the input sentence.| Train the r model after dataset reanalysis.The  and P instances are used to store processed input data in temporary files on a disk.} Train the r; model. No reanalysis of the input data will be performed.The  and P instances are used to store processed input data in temporary files on a disk.~xPrune disambiguation model: discard model features with absolute values (in log-domain) lower than the given threshold.Store dataset on a disk and run a handler on a list which is read lazily from the disk. A temporary file will be automatically deleted after the handler is done.rstuvwxyz{|"A morphosyntactic tagset to which Bs of the training and evaluation input data must correspond.BAnalysis function. It will be used to reanalyse input dataset.kHow many tags is the guessing model supposed to produce for a given OOV word? It will be used (see Te) on both training and evaluation input data prior to the training of the disambiguation model..Training configuration for the guessing model.7Training configuration for the disambiguation model.~Training dataset. This IO action will be executed a couple of times, so consider using lazy IO if your dataset is big. PEvaluation dataset IO action. Consider using lazy IO if your dataset is big.}"A morphosyntactic tagset to which Bs of the training and evaluation input data must correspond.kHow many tags is the guessing model supposed to produce for a given OOV word? It will be used (see Te) on both training and evaluation input data prior to the training of the disambiguation model..Training configuration for the guessing model.7Training configuration for the disambiguation model.~Training dataset. This IO action will be executed a couple of times, so consider using lazy IO if your dataset is big. PEvaluation dataset IO action. Consider using lazy IO if your dataset is big.~Directory to create the file in Template for  Input datasetHandler rstuvwxyz{|}~ rstuvwxyz{}|~ rstuvwxyz{|}~     !"#$%&''()*+,-./01123456789:;<=>?@ABCDEFGHIIJKLMNOPQRRSTUVWXIYZI[JKL\\]ST^V_`aXbbcdefghijjklm^nop`qXarstuvwx y z { | } ~  HHconcraft-0.9.4NLP.Concraft.DisambNLP.Concraft.MorphosyntaxNLP.Concraft.AnalysisNLP.Concraft.SchemaNLP.Concraft.Guess"NLP.Concraft.Morphosyntax.Accuracy NLP.ConcraftNLP.Concraft.Disamb.PositionalNLP.Concraft.Format.TempNLP.Concraft.Morphosyntax.AlignAtomposattsTierwithPoswithAttsWMapunWMapSentOsegsorigSentWordorthoovSegwordtagsmapSeg interpsSetinterpsmapSentmapSentOmkWMapmapWMapAnalyse reAnaSentreAnaPar SchemaConforthClowOrthC lowPrefixesC lowSuffixesCknownCshapeCpackedC begPackedCEntryBodyrangeoovOnlyargsBlockSchemaOxObvoid sequenceS_ fromBlockorthBlowOrthB lowPrefixesB lowSuffixesBknownBshapeBpackedB begPackedB entryWithentrynullConffromConf schematize TrainConf schemaConfTsgdArgsTonDiskTr0TR0T OovChosen AnyChosen AnyInterpsGuesser schemaConfcrfguessinclude guessSenttrain ReTrainConfinitDmbtiersTDisambtiersdisamb disambSent marginalspruneStatsgoodgoldaccuracyweakLBstrongLBweakUBstrongUBConcrafttagsetguessNumguesser saveModel loadModeltag reAnaTrainselectsplit $fBinaryAtom $fBinaryTier $fWordSeg $fFromJSONSeg $fToJSONSeg encodePar decodeParwriteParreadParsync moveDisambalignmatchBaseObmkBaseOblowOrthmkArg0mkArg1$fBinarySchemaConf $fBinaryBodycrf-chain1-constrained-0.3.1!Data.CRF.Chain1.Constrained.Train oovChosen anyChosen anyInterpsschemed$fBinaryGuesserunSplit$fBinaryDisamb.+.choicepositivebase Data.Tuplefstsnd aeson-0.8.0.2Data.Aeson.Types.ClassFromJSONToJSONwithTemp modelVersiontagset-positional-0.3.0Data.Tagset.PositionalTagtemporary-1.1.2.5System.IO.Temp withTempFile$fBinaryConcraft