-- Hoogle documentation, generated by Haddock
-- See Hoogle, http://www.haskell.org/hoogle/


-- | Morphological disambiguation based on constrained CRFs
--   
--   A morphological disambiguation library based on constrained
--   conditional random fields.
@package concraft
@version 0.14.2


-- | Segmentation-level ambiguities. TODO: consider moving the module
--   contents to <a>DAG</a>.
module NLP.Concraft.DAG.Morphosyntax.Ambiguous

-- | Identify ambigouos segments (roughly, segments which can be by-passed)
--   in the given DAG. Such ambiguous edges are marked in the resulting DAG
--   with <a>True</a> values.
identifyAmbiguousSegments :: DAG a b -> DAG a Bool

module NLP.Concraft.Morphosyntax.WMap

-- | A set with a non-negative weight assigned to each of its elements.
data WMap a

-- | Create WMap from a map.
fromMap :: Map a Double -> WMap a

-- | Map function over weighted collection elements.
mapWMap :: Ord b => (a -> b) -> WMap a -> WMap b

-- | Make a weighted collection. Negative elements will be ignored.
mkWMap :: Ord a => [(a, Double)] -> WMap a

-- | Trim down the set of potential labels to <tt>k</tt> most probable
--   ones.
trim :: Ord a => Int -> WMap a -> WMap a
instance Data.Binary.Class.Binary a => Data.Binary.Class.Binary (NLP.Concraft.Morphosyntax.WMap.WMap a)
instance GHC.Classes.Ord a => GHC.Classes.Ord (NLP.Concraft.Morphosyntax.WMap.WMap a)
instance GHC.Classes.Eq a => GHC.Classes.Eq (NLP.Concraft.Morphosyntax.WMap.WMap a)
instance GHC.Show.Show a => GHC.Show.Show (NLP.Concraft.Morphosyntax.WMap.WMap a)


-- | Types and functions related to the morphosyntax data layer.
module NLP.Concraft.Morphosyntax

-- | A segment parametrized over a word type and a tag type.
data Seg w t
Seg :: w -> WMap t -> Seg w t

-- | A word represented by the segment. Typically it will be an instance of
--   the <a>Word</a> class.
[word] :: Seg w t -> w

-- | A set of interpretations. To each interpretation a weight of
--   appropriateness within the context is assigned.
[tags] :: Seg w t -> WMap t

-- | Map function over segment tags.
mapSeg :: Ord b => (a -> b) -> Seg w a -> Seg w b

-- | Interpretations of the segment.
interpsSet :: Seg w t -> Set t

-- | Interpretations of the segment.
interps :: Seg w t -> [t]
class Word a

-- | Orthographic form.
orth :: Word a => a -> Text

-- | Out-of-vocabulary (OOV) word.
oov :: Word a => a -> Bool

-- | A sentence.
type Sent w t = [Seg w t]

-- | Map function over sentence tags.
mapSent :: Ord b => (a -> b) -> Sent w a -> Sent w b

-- | A sentence with original, textual representation.
data SentO w t
SentO :: Sent w t -> Text -> SentO w t
[segs] :: SentO w t -> Sent w t
[orig] :: SentO w t -> Text

-- | Map function over sentence tags.
mapSentO :: Ord b => (a -> b) -> SentO w a -> SentO w b
instance (GHC.Show.Show w, GHC.Show.Show t) => GHC.Show.Show (NLP.Concraft.Morphosyntax.SentO w t)
instance (GHC.Show.Show w, GHC.Show.Show t) => GHC.Show.Show (NLP.Concraft.Morphosyntax.Seg w t)
instance NLP.Concraft.Morphosyntax.Word w => NLP.Concraft.Morphosyntax.Word (NLP.Concraft.Morphosyntax.Seg w t)
instance Data.Aeson.Types.ToJSON.ToJSON w => Data.Aeson.Types.ToJSON.ToJSON (NLP.Concraft.Morphosyntax.Seg w Data.Text.Internal.Text)
instance Data.Aeson.Types.FromJSON.FromJSON w => Data.Aeson.Types.FromJSON.FromJSON (NLP.Concraft.Morphosyntax.Seg w Data.Text.Internal.Text)


-- | Accuracy statistics.
module NLP.Concraft.Morphosyntax.Accuracy

-- | Statistics.
data Stats

-- | Number of segments in gold corpus
Stats :: Int -> Int -> Stats

-- | Number of correct tags
[good] :: Stats -> Int
[gold] :: Stats -> Int

-- | Accuracy given stats.
accuracy :: Stats -> Double

-- | Accuracy weak lower bound.
weakLB :: Word w => Tagset -> [Seg w Tag] -> [Seg w Tag] -> Stats

-- | Accuracy weak upper bound.
weakUB :: Word w => Tagset -> [Seg w Tag] -> [Seg w Tag] -> Stats

-- | Accuracy strong lower bound.
strongLB :: Word w => Tagset -> [Seg w Tag] -> [Seg w Tag] -> Stats

-- | Accuracy strong upper bound.
strongUB :: Word w => Tagset -> [Seg w Tag] -> [Seg w Tag] -> Stats


-- | Morphosyntactic analysis utilities.
--   
--   See <a>reAnaSent</a> function for a description of how reanalsis is
--   performed. At some point it would be nice to change the entire process
--   so that sentence-level segmentation is also taken from the reanalysed
--   data.
module NLP.Concraft.Analysis

-- | An analyser performs word-level segmentation and morphological
--   analysis.
type Analyse w t = Text -> IO (Sent w t)

-- | Reanalyse sentence.
--   
--   From the reference sentence the function takes:
--   
--   <ul>
--   <li>Word-level segmentation</li>
--   <li>Chosen interpretations (tags)</li>
--   </ul>
--   
--   From the reanalysed sentence the function takes:
--   
--   <ul>
--   <li>Potential interpretations</li>
--   </ul>
reAnaSent :: Word w => Tagset -> Analyse w Tag -> SentO w Tag -> IO (Sent w Tag)

-- | Reanalyse paragraph.
reAnaPar :: Word w => Tagset -> Analyse w Tag -> [SentO w Tag] -> IO [Sent w Tag]


-- | Types and functions related to the morphosyntax data layer.
module NLP.Concraft.DAG.Morphosyntax

-- | A segment parametrized over a word type and a tag type.
data Seg w t
Seg :: w -> WMap t -> Seg w t

-- | A word represented by the segment. Typically it will be an instance of
--   the <a>Word</a> class.
[word] :: Seg w t -> w

-- | A set of interpretations. To each interpretation a weight of
--   appropriateness within the context is assigned.
[tags] :: Seg w t -> WMap t

-- | Map function over segment tags.
mapSeg :: Ord b => (a -> b) -> Seg w a -> Seg w b

-- | Interpretations of the segment.
interpsSet :: Seg w t -> Set t

-- | Interpretations of the segment.
interps :: Seg w t -> [t]
class Word a

-- | Orthographic form.
orth :: Word a => a -> Text

-- | Out-of-vocabulary (OOV) word.
oov :: Word a => a -> Bool

-- | A sentence. type Sent w t = [Seg w t]
type Sent w t = DAG () (Seg w t)

-- | Map function over sentence tags.
mapSent :: Ord b => (a -> b) -> Sent w a -> Sent w b

-- | A sentence with original, textual representation.
data SentO w t
SentO :: Sent w t -> Text -> SentO w t
[segs] :: SentO w t -> Sent w t
[orig] :: SentO w t -> Text

-- | Map function over sentence tags.
mapSentO :: Ord b => (a -> b) -> SentO w a -> SentO w b
instance (GHC.Show.Show w, GHC.Show.Show t) => GHC.Show.Show (NLP.Concraft.DAG.Morphosyntax.Seg w t)
instance NLP.Concraft.DAG.Morphosyntax.Word w => NLP.Concraft.DAG.Morphosyntax.Word (NLP.Concraft.DAG.Morphosyntax.Seg w t)
instance Data.Aeson.Types.ToJSON.ToJSON w => Data.Aeson.Types.ToJSON.ToJSON (NLP.Concraft.DAG.Morphosyntax.Seg w Data.Text.Internal.Text)
instance Data.Aeson.Types.FromJSON.FromJSON w => Data.Aeson.Types.FromJSON.FromJSON (NLP.Concraft.DAG.Morphosyntax.Seg w Data.Text.Internal.Text)


-- | Baseline word-segmentation functions.
module NLP.Concraft.DAG.Segmentation

-- | Which path type to search: shortest (<a>Min</a>) or longest
--   (<a>Max</a>)
data PathTyp
Min :: PathTyp
Max :: PathTyp
Freq :: FreqConf -> PathTyp

-- | Select the shortest-path (or longest, depending on <a>PathTyp</a>) in
--   the given DAG and remove all the edges which are not on this path.
pickPath :: Word b => PathTyp -> DAG a b -> DAG a b

-- | Retrieve the edges which belong to the shortest/longest (depending on
--   the argument function: <a>minimum</a> or <a>maximum</a>) path in the
--   given DAG.
findPath :: Word b => PathTyp -> DAG a b -> Set EdgeID

-- | Compute chosen/not-chosen counts of the individual orthographic forms
--   in the DAGs. Only the ambiguous segments are taken into account.
computeFreqs :: Word w => [Sent w t] -> Map Text (Int, Int)

-- | Configuration related to frequency-based path picking.
data FreqConf
FreqConf :: Map Text (Int, Int) -> Double -> FreqConf

-- | A map which assigns (chosen, not chosen) counts to the invidiaul
--   orthographic forms (see <a>computeFreqs</a>).
[pickFreqMap] :: FreqConf -> Map Text (Int, Int)

-- | A naive smoothing related parameter, which should be adddd to each
--   count in <a>pickFreqMap</a>. , orth :: DAG.EdgeID -&gt; T.Text -- ^
--   Orthographic form of a given edge
[smoothingParam] :: FreqConf -> Double

-- | Compute: * the number of tokens participating in ambiguities * the
--   total number of tokens
computeAmbiStats :: Word w => AmbiCfg -> [Sent w t] -> AmbiStats

-- | Numbers of tokens.
data AmbiCfg
AmbiCfg :: Bool -> AmbiCfg

-- | Only take the chosen tokens into account
[onlyChosen] :: AmbiCfg -> Bool

-- | Numbers of tokens.
data AmbiStats
AmbiStats :: !Int -> !Int -> AmbiStats

-- | Ambiguous tokens
[ambi] :: AmbiStats -> !Int

-- | All tokens
[total] :: AmbiStats -> !Int
instance GHC.Classes.Ord NLP.Concraft.DAG.Segmentation.AmbiStats
instance GHC.Classes.Eq NLP.Concraft.DAG.Segmentation.AmbiStats
instance GHC.Show.Show NLP.Concraft.DAG.Segmentation.AmbiStats
instance GHC.Classes.Ord NLP.Concraft.DAG.Segmentation.AmbiCfg
instance GHC.Classes.Eq NLP.Concraft.DAG.Segmentation.AmbiCfg
instance GHC.Show.Show NLP.Concraft.DAG.Segmentation.AmbiCfg


-- | Observation schema blocks for Concraft.
module NLP.Concraft.DAG.Schema

-- | An observation consist of an index (of list type) and an actual
--   observation value.
type Ob = ([Int], Text)

-- | The Ox monad specialized to word token type and text observations.
type Ox a = Ox Text a

-- | A schema is a block of the Ox computation performed within the context
--   of the sentence and the absolute sentence position.
type Schema w t a = Sent w t -> EdgeID -> Ox a

-- | A dummy schema block. TODO: is it a monad, an applicative?
void :: a -> Schema w t a

-- | Sequence the list of schemas (or blocks) and discard individual
--   values.
sequenceS_ :: [Sent w t -> a -> Ox b] -> Sent w t -> a -> Ox ()

-- | Use the schema to extract observations from the sentence.
schematize :: Schema w t a -> Sent w t -> DAG () [Ob]

-- | Body of configuration entry.
data Body a
Body :: [Int] -> Bool -> a -> Body a

-- | Range argument for the schema block.
[range] :: Body a -> [Int]

-- | When true, the entry is used only for oov words.
[oovOnly] :: Body a -> Bool

-- | Additional arguments for the schema block.
[args] :: Body a -> a

-- | Maybe entry.
type Entry a = Maybe (Body a)

-- | Plain entry with no additional arugments.
entry :: [Int] -> Entry ()

-- | Entry with additional arguemnts.
entryWith :: a -> [Int] -> Entry a

-- | Configuration of the schema. All configuration elements specify the
--   range over which a particular observation type should be taken on
--   account. For example, the <tt>[-1, 0, 2]</tt> range means that
--   observations of particular type will be extracted with respect to
--   previous (<tt>k - 1</tt>), current (<tt>k</tt>) and after the next
--   (<tt>k + 2</tt>) positions when identifying the observation set for
--   position <tt>k</tt> in the input sentence.
data SchemaConf
SchemaConf :: Entry () -> Entry () -> Entry [Int] -> Entry [Int] -> Entry () -> Entry () -> Entry () -> Entry () -> SchemaConf

-- | The <a>orthB</a> schema block.
[orthC] :: SchemaConf -> Entry ()

-- | The <a>lowOrthB</a> schema block.
[lowOrthC] :: SchemaConf -> Entry ()

-- | The <a>lowPrefixesB</a> schema block. The first list of ints
--   represents lengths of prefixes.
[lowPrefixesC] :: SchemaConf -> Entry [Int]

-- | The <a>lowSuffixesB</a> schema block. The first list of ints
--   represents lengths of suffixes.
[lowSuffixesC] :: SchemaConf -> Entry [Int]

-- | The <a>knownB</a> schema block.
[knownC] :: SchemaConf -> Entry ()

-- | The <a>shapeB</a> schema block.
[shapeC] :: SchemaConf -> Entry ()

-- | The <a>packedB</a> schema block.
[packedC] :: SchemaConf -> Entry ()

-- | The <a>begPackedB</a> schema block.
[begPackedC] :: SchemaConf -> Entry ()

-- | Null configuration of the observation schema.
nullConf :: SchemaConf

-- | Build the schema based on the configuration.
fromConf :: Word w => SchemaConf -> Schema w t ()

-- | A block is a chunk of the Ox computation performed within the context
--   of the sentence and the list of absolute sentence positions.
type Block w t a = Sent w t -> [EdgeID] -> Ox a

-- | Transform 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.
fromBlock :: Word w => Block w t a -> [Int] -> Bool -> Schema w t a

-- | Orthographic form at the current position.
orthB :: Word w => Block w t ()

-- | Orthographic form at the current position.
lowOrthB :: Word w => Block w t ()

-- | List of lowercased prefixes of given lengths.
lowPrefixesB :: Word w => [Int] -> Block w t ()

-- | List of lowercased suffixes of given lengths.
lowSuffixesB :: Word w => [Int] -> Block w t ()

-- | Shape of the word.
knownB :: Word w => Block w t ()

-- | Shape of the word.
shapeB :: Word w => Block w t ()

-- | Packed shape of the word.
packedB :: Word w => Block w t ()

-- | Packed shape of the word.
begPackedB :: Word w => Block w t ()
instance GHC.Show.Show NLP.Concraft.DAG.Schema.SchemaConf
instance GHC.Show.Show a => GHC.Show.Show (NLP.Concraft.DAG.Schema.Body a)
instance Data.Binary.Class.Binary NLP.Concraft.DAG.Schema.SchemaConf
instance Data.Binary.Class.Binary a => Data.Binary.Class.Binary (NLP.Concraft.DAG.Schema.Body a)


-- | Accuracy statistics.
module NLP.Concraft.DAG.Morphosyntax.Accuracy

-- | True positives, false positives, etc.
data Stats
Stats :: !Int -> !Int -> !Int -> !Int -> !Int -> Stats

-- | True positive
[tp] :: Stats -> !Int

-- | False positive
[fp] :: Stats -> !Int

-- | True negative
[tn] :: Stats -> !Int

-- | False negative
[fn] :: Stats -> !Int

-- | Consistency error (number of edges for which both <a>fp</a> and
--   <a>fn</a> hold)
[ce] :: Stats -> !Int

-- | Configuration of accuracy computation.
data AccCfg x
AccCfg :: Bool -> Bool -> Set x -> Tagset -> Bool -> Bool -> Bool -> Bool -> AccCfg x

-- | Limit calculations to OOV words
[onlyOov] :: AccCfg x -> Bool

-- | Limit calculations to segmentation-ambiguous words
[onlyAmb] :: AccCfg x -> Bool

-- | Limit calculations to segments marked with one of the given labels; if
--   empty, the option has no effect
[onlyMarkedWith] :: AccCfg x -> Set x

-- | The underlying tagset
[accTagset] :: AccCfg x -> Tagset

-- | Should the tags be expanded?
[expandTag] :: AccCfg x -> Bool

-- | Compute segmentation-level accurracy. The actually chosen tags are
--   ignored, only information about the chosen DAG edges is relevant.
[ignoreTag] :: AccCfg x -> Bool

-- | If weak, there has to be an overlap in the tags assigned to a given
--   segment in both datasets. Otherwise, the two sets of tags have to be
--   identical. , discardProb0 :: Bool -- ^ Whether sentences with near 0
--   probability should be discarded from -- evaluation.
[weakAcc] :: AccCfg x -> Bool

-- | Print information about compared elements
[verbose] :: AccCfg x -> Bool

-- | Compute the accuracy of the model with respect to the labeled dataset.
--   To each <a>Tag</a> an additional information <tt>x</tt> can be
--   assigned, which will be taken into account when computing statistics.
collect :: (Word w, Ord x, Show x) => AccCfg x -> [Sent w (Tag, x)] -> [Sent w (Tag, x)] -> Stats
precision :: Stats -> Double
recall :: Stats -> Double
accuracy :: Stats -> Double
instance GHC.Classes.Ord NLP.Concraft.DAG.Morphosyntax.Accuracy.Stats
instance GHC.Classes.Eq NLP.Concraft.DAG.Morphosyntax.Accuracy.Stats
instance GHC.Show.Show NLP.Concraft.DAG.Morphosyntax.Accuracy.Stats

module NLP.Concraft.DAG.Guess

-- | A guessing model.
data Guesser t s
Guesser :: SchemaConf -> CRF Ob s -> s -> Set t -> (t -> s) -> (s -> t) -> Guesser t s
[schemaConf] :: Guesser t s -> SchemaConf
[crf] :: Guesser t s -> CRF Ob s
[zeroProbLab] :: Guesser t s -> s

-- | The tagset considered for the unknown words (TODO: a solution parallel
--   and not 100% consistent with what is implemented in the CRF library)
--   TODO: with <a>complexify</a>, <a>unkTagSet</a> is not needed anymore!
[unkTagSet] :: Guesser t s -> Set t

-- | A tag simplification function
[simplify] :: Guesser t s -> t -> s

-- | NEW: instead of an <a>unkTagSet</a>, a function which makes a complex
--   tag out of a simple tag.
--   
--   WARNING: we assume, that this function does not conflate simplified
--   tags, i.e., tag to tags of type <tt>s</tt> cannot lead to one and the
--   same complex tag of type <tt>t</tt>.
[complexify] :: Guesser t s -> s -> t

-- | Store the entire guessing model apart from the simplification
--   function.
putGuesser :: (Binary t, Binary s, Ord s) => Guesser t s -> Put

-- | Get the disambiguation model, provided the simplification function.
--   getGuesser :: (M.Map t T.Tag) -&gt; Get (Guesser t)
getGuesser :: (Binary t, Binary s, Ord s, Ord t) => (t -> s) -> (s -> t) -> Get (Guesser t s)

-- | Determine the marginal probabilities of the individual labels in the
--   sentence.
marginals :: (Word w, Ord t, Ord s) => Config s -> Guesser t s -> Sent w t -> DAG () (WMap t)

-- | Replace the probabilities of the sentence labels with the marginal
--   probabilities stemming from the model.
marginalsSent :: (Word w, Ord t, Ord s) => Config s -> Guesser t s -> Sent w t -> Sent w t

-- | Training configuration.
data TrainConf t s
TrainConf :: SchemaConf -> SgdArgs -> Bool -> R0T -> t -> (t -> s) -> (s -> t) -> (t -> t) -> Bool -> TrainConf t s
[schemaConfT] :: TrainConf t s -> SchemaConf

-- | SGD parameters.
[sgdArgsT] :: TrainConf t s -> SgdArgs

-- | Store SGD dataset on disk
[onDiskT] :: TrainConf t s -> Bool

-- | R0 construction method
[r0T] :: TrainConf t s -> R0T

-- | Zero probability label
[zeroProbLabel] :: TrainConf t s -> t

-- | Label simplification function
[simplifyLabel] :: TrainConf t s -> t -> s

-- | Label complexification function
[complexifyLabel] :: TrainConf t s -> s -> t

-- | Strip the label from irrelevant information. Used to determine the set
--   of possible tags for unknown words. TODO: we don't need this with
--   <a>complexify</a> anymore!?
[stripLabel] :: TrainConf t s -> t -> t

-- | Guess only visible features
[onlyVisible] :: TrainConf t s -> Bool

-- | Method of constructing the default set of labels (R0).
data R0T

-- | See <a>anyInterps</a>
AnyInterps :: R0T

-- | See <a>anyChosen</a>
AnyChosen :: R0T

-- | See <a>oovChosen</a>
OovChosen :: R0T

-- | Train guesser.
train :: (Word w, Ord t, Ord s) => TrainConf t s -> IO [Sent w t] -> IO [Sent w t] -> IO (Guesser t s)

-- | Schematized dataset.
schemed :: (Word w, Ord t, Ord s) => (t -> s) -> Schema w s a -> [Sent w t] -> [SentL Ob s]
instance Data.Data.Data NLP.Concraft.DAG.Guess.R0T
instance GHC.Enum.Enum NLP.Concraft.DAG.Guess.R0T
instance GHC.Classes.Ord NLP.Concraft.DAG.Guess.R0T
instance GHC.Classes.Eq NLP.Concraft.DAG.Guess.R0T
instance GHC.Show.Show NLP.Concraft.DAG.Guess.R0T

module NLP.Concraft.DAG.Disamb

-- | A disambiguation model.
data Disamb t
Disamb :: [Tier] -> SchemaConf -> CRF Ob Atom -> (t -> Tag) -> Disamb t
[tiers] :: Disamb t -> [Tier]
[schemaConf] :: Disamb t -> SchemaConf
[crf] :: Disamb t -> CRF Ob Atom

-- | A map which simplifies the tags of generic type <tt>t</tt> to
--   simplified positional tags. The motivation behind this is that tags
--   can have a richer structure.
--   
--   NOTE: it can happen in real situations that a tag is encountered which
--   is not known by the model. It would be nice to be able to treat it as
--   the closest tag that can be handled. Then, one have to define the
--   notion of the similarilty between tags, though... But probably it
--   should be done at a different level (where more information about the
--   structure of <tt>t</tt> is known)
[simplify] :: Disamb t -> t -> Tag

-- | Store the entire disambiguation model apart from the simplification
--   function.
putDisamb :: Disamb t -> Put

-- | Get the disambiguation model, provided the simplification function.
--   getDisamb :: (M.Map t T.Tag) -&gt; Get (Disamb t)
getDisamb :: (t -> Tag) -> Get (Disamb t)

-- | A tier description.
data Tier
Tier :: Bool -> Bool -> Set Attr -> Tier

-- | Does it include the part of speech?
[withPos] :: Tier -> Bool

-- | End-of-sentence marker.
[withEos] :: Tier -> Bool

-- | Tier grammatical attributes.
[withAtts] :: Tier -> Set Attr

-- | An atomic part of morphosyntactic tag with optional POS.
data Atom
Atom :: Maybe POS -> Map Attr Text -> Maybe Bool -> Atom
[pos] :: Atom -> Maybe POS
[atts] :: Atom -> Map Attr Text

-- | NOTE: could be simplified to Bool, but this way it's more readable
[eos] :: Atom -> Maybe Bool

-- | Type of resulting probabilities.
data ProbType

-- | Marginal probabilities
Marginals :: ProbType

-- | TODO
MaxProbs :: ProbType

-- | Determine the marginal probabilities of to individual labels in the
--   sentence. marginalsSent :: (X.Word w, Ord t) =&gt; Disamb t -&gt;
--   X.Sent w t -&gt; DAG () (X.WMap [P.Atom])
probsSent :: (Word w, Ord t) => ProbType -> Disamb t -> Sent w t -> Sent w t

-- | Determine the marginal probabilities of to individual labels in the
--   sentence.
probs :: (Word w, Ord t) => ProbType -> Disamb t -> Sent w t -> DAG () (WMap t)

-- | Training configuration.
data TrainConf t
TrainConf :: [Tier] -> SchemaConf -> SgdArgs -> Bool -> (t -> Tag) -> TrainConf t
[tiersT] :: TrainConf t -> [Tier]
[schemaConfT] :: TrainConf t -> SchemaConf
[sgdArgsT] :: TrainConf t -> SgdArgs
[onDiskT] :: TrainConf t -> Bool

-- | Label simplification function
[simplifyLabel] :: TrainConf t -> t -> Tag

-- | Train disambiguation module.
train :: (Word w, Ord t) => TrainConf t -> IO [Sent w t] -> IO [Sent w t] -> IO (Disamb t)

-- | Prune disamb model: discard model features with absolute values (in
--   log-domain) lower than the given threshold.
prune :: Double -> Disamb t -> Disamb t

-- | Schematize the input sentence according to <tt>schema</tt> rules.
schematize :: Schema w [t] a -> Sent w [t] -> Sent Ob t


-- | A version of the disambigation model adapted to perform sentence
--   segmentation as well.
module NLP.Concraft.DAG.DisambSeg

-- | The internal tag type.
data Tag
Tag :: Tag -> Bool -> Tag

-- | Positional tag
[posiTag] :: Tag -> Tag

-- | End-of-sentence marker
[hasEos] :: Tag -> Bool

-- | A disambiguation model.
data Disamb t
Disamb :: [Tier] -> SchemaConf -> CRF Ob Atom -> (t -> Tag) -> Disamb t
[tiers] :: Disamb t -> [Tier]
[schemaConf] :: Disamb t -> SchemaConf
[crf] :: Disamb t -> CRF Ob Atom

-- | A function which simplifies the tags of the generic type <tt>t</tt> to
--   (i) the corresponding positional tags and (ii) information if the
--   segment represents sentence end.
--   
--   NOTE: it can happen in real situations that a tag is encountered which
--   is not known by the model. It would be nice to be able to treat it as
--   the closest tag that can be handled. Then, one have to define the
--   notion of the similarilty between tags, though... But probably it
--   should be done at a different level (where more information about the
--   structure of <tt>t</tt> is known)
[simplify] :: Disamb t -> t -> Tag

-- | Store the entire disambiguation model apart from the simplification
--   function.
putDisamb :: Disamb t -> Put

-- | Get the disambiguation model, provided the simplification function.
--   getDisamb :: (M.Map t T.Tag) -&gt; Get (Disamb t)
getDisamb :: (t -> Tag) -> Get (Disamb t)

-- | A tier description.
data Tier
Tier :: Bool -> Bool -> Set Attr -> Tier

-- | Does it include the part of speech?
[withPos] :: Tier -> Bool

-- | End-of-sentence marker.
[withEos] :: Tier -> Bool

-- | Tier grammatical attributes.
[withAtts] :: Tier -> Set Attr

-- | An atomic part of morphosyntactic tag with optional POS.
data Atom
Atom :: Maybe POS -> Map Attr Text -> Maybe Bool -> Atom
[pos] :: Atom -> Maybe POS
[atts] :: Atom -> Map Attr Text

-- | NOTE: could be simplified to Bool, but this way it's more readable
[eos] :: Atom -> Maybe Bool

-- | Perform disambiguation.
disamb :: (Word w, Ord t) => Disamb t -> Sent w t -> DAG () (Map t Bool)

-- | Type of resulting probabilities.
data ProbType

-- | Marginal probabilities
Marginals :: ProbType

-- | TODO
MaxProbs :: ProbType

-- | Determine the marginal probabilities of to individual labels in the
--   sentence.
probsSent :: (Word w, Ord t) => ProbType -> Disamb t -> Sent w t -> Sent w t

-- | Determine the marginal probabilities of to individual labels in the
--   sentence.
probs :: (Word w, Ord t) => ProbType -> Disamb t -> Sent w t -> DAG () (WMap t)

-- | Training configuration.
data TrainConf t
TrainConf :: [Tier] -> SchemaConf -> SgdArgs -> Bool -> (t -> Tag) -> TrainConf t
[tiersT] :: TrainConf t -> [Tier]
[schemaConfT] :: TrainConf t -> SchemaConf
[sgdArgsT] :: TrainConf t -> SgdArgs
[onDiskT] :: TrainConf t -> Bool

-- | Label simplification function
[simplifyLabel] :: TrainConf t -> t -> Tag

-- | Train disambiguation module.
train :: (Word w, Ord t) => TrainConf t -> IO [Sent w t] -> IO [Sent w t] -> IO (Disamb t)

-- | Prune disamb model: discard model features with absolute values (in
--   log-domain) lower than the given threshold.
prune :: Double -> Disamb t -> Disamb t
instance GHC.Classes.Ord NLP.Concraft.DAG.DisambSeg.Tag
instance GHC.Classes.Eq NLP.Concraft.DAG.DisambSeg.Tag
instance GHC.Show.Show NLP.Concraft.DAG.DisambSeg.Tag


-- | Top-level module adated to DAGs, guessing and disambiguation.
module NLP.Concraft.DAGSeg

-- | Concraft data.
data Concraft t
Concraft :: Tagset -> Int -> Guesser t Tag -> Disamb t -> Disamb t -> Concraft t
[tagset] :: Concraft t -> Tagset
[guessNum] :: Concraft t -> Int
[guesser] :: Concraft t -> Guesser t Tag
[segmenter] :: Concraft t -> Disamb t
[disamber] :: Concraft t -> Disamb t

-- | Save model in a file. Data is compressed using the gzip format.
saveModel :: (Ord t, Binary t) => FilePath -> Concraft t -> IO ()

-- | Load model from a file.
loadModel :: (Ord t, Binary t) => (Tagset -> t -> Tag) -> (Tagset -> Tag -> t) -> (Tagset -> t -> Tag) -> FilePath -> IO (Concraft t)

-- | DAG annotation, assignes <tt>b</tt> values to <tt>a</tt> labels for
--   each edge in the graph.
type Anno a b = DAG () (Map a b)

-- | Determine max probabilities corresponding to individual tags w.r.t.
--   the disambiguation model.
disamb :: (Word w, Ord t) => Disamb t -> Sent w t -> Anno t Bool

-- | Find all optimal paths in the given annotation. Optimal paths are
--   those which go through tags with the assigned probability 1. For a
--   given chosen edge, all the tags with probability 1 are selected.
findOptimalPaths :: Ord t => Anno t Double -> [[(EdgeID, Set t)]]

-- | Make the given path with disamb markers in the given annotation and
--   produce a new disamb annotation.
disambPath :: Ord t => [(EdgeID, Set t)] -> Anno t Double -> Anno t Bool

-- | Determine marginal probabilities corresponding to individual tags
--   w.r.t. the guessing model.
guessMarginals :: (Word w, Ord t) => Config Tag -> Guesser t Tag -> Sent w t -> Anno t Double

-- | Determine marginal probabilities corresponding to individual tags
--   w.r.t. the disambiguation model.
disambMarginals :: (Word w, Ord t) => Disamb t -> Sent w t -> Anno t Double

-- | Determine max probabilities corresponding to individual tags w.r.t.
--   the disambiguation model.
disambProbs :: (Word w, Ord t) => ProbType -> Disamb t -> Sent w t -> Anno t Double

-- | Extend the OOV words with new, guessed interpretations.
--   
--   Determine marginal probabilities corresponding to individual tags
--   w.r.t. the guessing model and, afterwards, trim the sentence to keep
--   only the <tt>k</tt> most probably labels for each OOV edge. Note that,
--   for OOV words, the entire set of default tags is considered.
guessSent :: (Word w, Ord t) => Int -> Config Tag -> Guesser t Tag -> Sent w t -> Sent w t

-- | Perform guessing, trimming, and finally determine marginal
--   probabilities corresponding to individual tags w.r.t. the guessing
--   model.
guess :: (Word w, Ord t) => Int -> Config Tag -> Guesser t Tag -> Sent w t -> Anno t Double

-- | Perform guessing, trimming, and finally determine marginal
--   probabilities corresponding to individual tags w.r.t. the
--   disambiguation model.
tag :: (Word w, Ord t) => Int -> Config Tag -> Concraft t -> Sent w t -> Anno t Double

-- | Prune the disambiguation model: discard model features with absolute
--   values (in log-domain) lower than the given threshold.
prune :: Double -> Concraft t -> Concraft t


-- | Observation schema blocks for Concraft.
module NLP.Concraft.Schema

-- | An observation consist of an index (of list type) and an actual
--   observation value.
type Ob = ([Int], Text)

-- | The Ox monad specialized to word token type and text observations.
type Ox a = Ox Text a

-- | A schema is a block of the Ox computation performed within the context
--   of the sentence and the absolute sentence position.
type Schema w t a = Vector (Seg w t) -> Int -> Ox a

-- | A dummy schema block.
void :: a -> Schema w t a

-- | Sequence the list of schemas (or blocks) and discard individual
--   values.
sequenceS_ :: [Vector (Seg w t) -> a -> Ox b] -> Vector (Seg w t) -> a -> Ox ()

-- | Use the schema to extract observations from the sentence.
schematize :: Schema w t a -> Sent w t -> [[Ob]]

-- | Body of configuration entry.
data Body a
Body :: [Int] -> Bool -> a -> Body a

-- | Range argument for the schema block.
[range] :: Body a -> [Int]

-- | When true, the entry is used only for oov words.
[oovOnly] :: Body a -> Bool

-- | Additional arguments for the schema block.
[args] :: Body a -> a

-- | Maybe entry.
type Entry a = Maybe (Body a)

-- | Plain entry with no additional arugments.
entry :: [Int] -> Entry ()

-- | Entry with additional arguemnts.
entryWith :: a -> [Int] -> Entry a

-- | Configuration of the schema. All configuration elements specify the
--   range over which a particular observation type should be taken on
--   account. For example, the <tt>[-1, 0, 2]</tt> range means that
--   observations of particular type will be extracted with respect to
--   previous (<tt>k - 1</tt>), current (<tt>k</tt>) and after the next
--   (<tt>k + 2</tt>) positions when identifying the observation set for
--   position <tt>k</tt> in the input sentence.
data SchemaConf
SchemaConf :: Entry () -> Entry () -> Entry [Int] -> Entry [Int] -> Entry () -> Entry () -> Entry () -> Entry () -> SchemaConf

-- | The <a>orthB</a> schema block.
[orthC] :: SchemaConf -> Entry ()

-- | The <a>lowOrthB</a> schema block.
[lowOrthC] :: SchemaConf -> Entry ()

-- | The <a>lowPrefixesB</a> schema block. The first list of ints
--   represents lengths of prefixes.
[lowPrefixesC] :: SchemaConf -> Entry [Int]

-- | The <a>lowSuffixesB</a> schema block. The first list of ints
--   represents lengths of suffixes.
[lowSuffixesC] :: SchemaConf -> Entry [Int]

-- | The <a>knownB</a> schema block.
[knownC] :: SchemaConf -> Entry ()

-- | The <a>shapeB</a> schema block.
[shapeC] :: SchemaConf -> Entry ()

-- | The <a>packedB</a> schema block.
[packedC] :: SchemaConf -> Entry ()

-- | The <a>begPackedB</a> schema block.
[begPackedC] :: SchemaConf -> Entry ()

-- | Null configuration of the observation schema.
nullConf :: SchemaConf

-- | Build the schema based on the configuration.
fromConf :: Word w => SchemaConf -> Schema w t ()

-- | A block is a chunk of the Ox computation performed within the context
--   of the sentence and the list of absolute sentence positions.
type Block w t a = Vector (Seg w t) -> [Int] -> Ox a

-- | Transform 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.
fromBlock :: Word w => Block w t a -> [Int] -> Bool -> Schema w t a

-- | Orthographic form at the current position.
orthB :: Word w => Block w t ()

-- | Orthographic form at the current position.
lowOrthB :: Word w => Block w t ()

-- | List of lowercased prefixes of given lengths.
lowPrefixesB :: Word w => [Int] -> Block w t ()

-- | List of lowercased suffixes of given lengths.
lowSuffixesB :: Word w => [Int] -> Block w t ()

-- | Shape of the word.
knownB :: Word w => Block w t ()

-- | Shape of the word.
shapeB :: Word w => Block w t ()

-- | Packed shape of the word.
packedB :: Word w => Block w t ()

-- | Packed shape of the word.
begPackedB :: Word w => Block w t ()
instance GHC.Show.Show NLP.Concraft.Schema.SchemaConf
instance GHC.Show.Show a => GHC.Show.Show (NLP.Concraft.Schema.Body a)
instance Data.Binary.Class.Binary NLP.Concraft.Schema.SchemaConf
instance Data.Binary.Class.Binary a => Data.Binary.Class.Binary (NLP.Concraft.Schema.Body a)

module NLP.Concraft.Guess

-- | A guessing model.
data Guesser t
Guesser :: SchemaConf -> CRF Ob t -> Guesser t
[schemaConf] :: Guesser t -> SchemaConf
[crf] :: Guesser t -> CRF Ob t

-- | Determine the <tt>k</tt> most probable labels for each word in the
--   sentence. TODO: Perhaps it would be better to use sets instead of
--   lists as output?
guess :: (Word w, Ord t) => Int -> Guesser t -> Sent w t -> [[t]]

-- | Insert guessing results into the sentence. Only interpretations of OOV
--   words will be extended.
include :: (Word w, Ord t) => [[t]] -> Sent w t -> Sent w t

-- | Combine <a>guess</a> with <a>include</a>.
guessSent :: (Word w, Ord t) => Int -> Guesser t -> Sent w t -> Sent w t

-- | Training configuration.
data TrainConf
TrainConf :: SchemaConf -> SgdArgs -> Bool -> R0T -> TrainConf
[schemaConfT] :: TrainConf -> SchemaConf

-- | SGD parameters.
[sgdArgsT] :: TrainConf -> SgdArgs

-- | Store SGD dataset on disk
[onDiskT] :: TrainConf -> Bool

-- | R0 construction method
[r0T] :: TrainConf -> R0T

-- | Method of constructing the default set of labels (R0).
data R0T

-- | See <a>anyInterps</a>
AnyInterps :: R0T

-- | See <a>anyChosen</a>
AnyChosen :: R0T

-- | See <a>oovChosen</a>
OovChosen :: R0T

-- | Train guesser.
train :: (Word w, Ord t) => TrainConf -> IO [Sent w t] -> IO [Sent w t] -> IO (Guesser t)
instance Data.Data.Data NLP.Concraft.Guess.R0T
instance GHC.Enum.Enum NLP.Concraft.Guess.R0T
instance GHC.Classes.Ord NLP.Concraft.Guess.R0T
instance GHC.Classes.Eq NLP.Concraft.Guess.R0T
instance GHC.Show.Show NLP.Concraft.Guess.R0T
instance (GHC.Classes.Ord t, Data.Binary.Class.Binary t) => Data.Binary.Class.Binary (NLP.Concraft.Guess.Guesser t)

module NLP.Concraft.Disamb

-- | A disambiguation model.
data Disamb
Disamb :: [Tier] -> SchemaConf -> CRF Ob Atom -> Disamb
[tiers] :: Disamb -> [Tier]
[schemaConf] :: Disamb -> SchemaConf
[crf] :: Disamb -> CRF Ob Atom

-- | A tier description.
data Tier
Tier :: Bool -> Bool -> Set Attr -> Tier

-- | Does it include the part of speech?
[withPos] :: Tier -> Bool

-- | End-of-sentence marker.
[withEos] :: Tier -> Bool

-- | Tier grammatical attributes.
[withAtts] :: Tier -> Set Attr

-- | An atomic part of morphosyntactic tag with optional POS.
data Atom
Atom :: Maybe POS -> Map Attr Text -> Maybe Bool -> Atom
[pos] :: Atom -> Maybe POS
[atts] :: Atom -> Map Attr Text

-- | NOTE: could be simplified to Bool, but this way it's more readable
[eos] :: Atom -> Maybe Bool

-- | Tag labels with marginal probabilities.
marginals :: Word w => Disamb -> Sent w Tag -> [WMap Tag]

-- | Perform context-sensitive disambiguation.
disamb :: Word w => Disamb -> Sent w Tag -> [Tag]

-- | Insert disambiguation results into the sentence.
include :: (Sent w Tag -> [Tag]) -> Sent w Tag -> Sent w Tag

-- | Combine <a>disamb</a> with <a>include</a>.
disambSent :: Word w => Disamb -> Sent w Tag -> Sent w Tag

-- | Training configuration.
data TrainConf
TrainConf :: [Tier] -> SchemaConf -> SgdArgs -> Bool -> TrainConf
[tiersT] :: TrainConf -> [Tier]
[schemaConfT] :: TrainConf -> SchemaConf
[sgdArgsT] :: TrainConf -> SgdArgs
[onDiskT] :: TrainConf -> Bool
ReTrainConf :: Disamb -> SgdArgs -> Bool -> TrainConf
[initDmb] :: TrainConf -> Disamb
[sgdArgsT] :: TrainConf -> SgdArgs
[onDiskT] :: TrainConf -> Bool

-- | Train disamb model.
train :: Word w => TrainConf -> IO [Sent w Tag] -> IO [Sent w Tag] -> IO Disamb

-- | Prune disamb model: discard model features with absolute values (in
--   log-domain) lower than the given threshold.
prune :: Double -> Disamb -> Disamb
instance Data.Binary.Class.Binary NLP.Concraft.Disamb.Disamb

module NLP.Concraft

-- | Concraft data.
data Concraft
Concraft :: Tagset -> Int -> Guesser Tag -> Disamb -> Concraft
[tagset] :: Concraft -> Tagset
[guessNum] :: Concraft -> Int
[guesser] :: Concraft -> Guesser Tag
[disamb] :: Concraft -> Disamb

-- | Save model in a file. Data is compressed using the gzip format.
saveModel :: FilePath -> Concraft -> IO ()

-- | Load model from a file.
loadModel :: FilePath -> IO Concraft

-- | Tag sentence using the model. In your code you should probably use
--   your analysis function, translate results into a container of
--   <a>Sent</a>ences, evaluate <a>tag</a> on each sentence and embed the
--   tagging results into the morphosyntactic structure of your own.
--   
--   The function returns guessing results as <a>fst</a> elements of the
--   output pairs and disambiguation results as <a>snd</a> elements of the
--   corresponding pairs.
tag :: Word w => Concraft -> Sent w Tag -> [(Set Tag, Tag)]

-- | 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.
marginals :: Word w => Concraft -> Sent w Tag -> [WMap Tag]

-- | Train the <a>Concraft</a> model. No reanalysis of the input data will
--   be performed.
--   
--   The <a>FromJSON</a> and <a>ToJSON</a> instances are used to store
--   processed input data in temporary files on a disk.
train :: (Word w, FromJSON w, ToJSON w) => Tagset -> Int -> TrainConf -> TrainConf -> IO [Sent w Tag] -> IO [Sent w Tag] -> IO Concraft

-- | Train the <a>Concraft</a> model after dataset reanalysis.
--   
--   The <a>FromJSON</a> and <a>ToJSON</a> instances are used to store
--   processed input data in temporary files on a disk.
reAnaTrain :: (Word w, FromJSON w, ToJSON w) => Tagset -> Analyse w Tag -> Int -> TrainConf -> TrainConf -> IO [SentO w Tag] -> IO [SentO w Tag] -> IO Concraft

-- | Prune disambiguation model: discard model features with absolute
--   values (in log-domain) lower than the given threshold.
prune :: Double -> Concraft -> Concraft
instance Data.Binary.Class.Binary NLP.Concraft.Concraft