src/Gargantext/Text/Terms/Stop.hs

   1 {-|
   2 Module      : Gargantext.Text.Terms.Stop
   3 Description : Mono Terms module
   4 Copyright   : (c) CNRS, 2017 - present
   5 License     : AGPL + CECILL v3
   6 Maintainer  : team@gargantext.org
   7 Stability   : experimental
   8 Portability : POSIX
   9
  10 Stop words and (how to learn it).
  11
  12 Main type here is String.
  13
  14 -}
  15
  16 {-# LANGUAGE NoImplicitPrelude #-}
  17
  18 module Gargantext.Text.Terms.Stop (detectLang, detectLangs, stopList)
  19   where
  20
  21 import GHC.Base (Functor)
  22 import Numeric.Probability.Distribution ((??))
  23 import qualified Numeric.Probability.Distribution as D
  24
  25 import Data.Char (toLower)
  26 import qualified Data.List as DL
  27
  28 import Data.Maybe (maybe)
  29 import Data.Map.Strict (Map, toList)
  30 import qualified Data.Map.Strict as DM
  31
  32 import Data.String (String)
  33
  34 import Data.Text (Text)
  35 import Data.Text (pack, unpack)
  36 import Data.Tuple.Extra (both)
  37
  38 import Gargantext.Prelude
  39 import Gargantext.Core (Lang(..), allLangs)
  40 import Gargantext.Text.Terms.Mono (words)
  41 import Gargantext.Text.Metrics.Count (occurrencesWith)
  42
  43 import qualified Gargantext.Text.Samples.FR as FR
  44 import qualified Gargantext.Text.Samples.EN as EN
  45 --import qualified Gargantext.Text.Samples.DE as DE
  46 --import qualified Gargantext.Text.Samples.SP as SP
  47 --import qualified Gargantext.Text.Samples.CH as CH
  48
  49 ------------------------------------------------------------------------
  50 data Candidate = Candidate { stop :: Double
  51                            , noStop :: Double
  52  } deriving (Show)
  53
  54 -- * String preparation
  55
  56 -- | String prepare
  57 blanks :: String -> String
  58 blanks [] = []
  59 blanks xs = [' '] <> xs <> [' ']
  60
  61 -- | Blocks increase the size of the word to ease computations
  62 -- some border and unexepected effects can happen, need to be tested
  63 blockOf :: Int -> String -> String
  64 blockOf n = DL.concat . DL.take n . DL.repeat
  65
  66 -- | Chunks is the same function as splitBy in Context but for Strings,
  67 -- not Text (without pack and unpack operations that are not needed).
  68 chunks :: Int -> Int -> String -> [String]
  69 chunks n m = DL.take m . filter (not . all (== ' '))
  70                        . chunkAlong (n+1) 1
  71                        . DL.concat
  72                        . DL.take 1000
  73                        . DL.repeat
  74                        . blanks
  75
  76 allChunks :: [Int] -> Int -> String -> [String]
  77 allChunks ns m st = DL.concat $ map (\n -> chunks n m st) ns
  78
  79 allChunks' :: [Int] -> Int -> String -> [[String]]
  80 allChunks' ns m st = map (\n -> chunks n m st) ns
  81
  82 ------------------------------------------------------------------------
  83 -- * Analyze candidate
  84 type StringSize = Int
  85 type TotalFreq  = Int
  86 type Freq       = Int
  87 type Word       = String
  88
  89 data LangWord = LangWord Lang Word
  90
  91 type LangProba = Map Lang Double
  92
  93 ------------------------------------------------------------------------
  94 detectLang :: Text -> Maybe Lang
  95 detectLang = head . map fst . detectLangs
  96
  97 detectLangs :: Text -> [(Lang, Double)]
  98 detectLangs = detectLangs' . unpack
  99
 100 detectLangs' :: String -> [(Lang, Double)]
 101 detectLangs' s =  DL.reverse $ DL.sortOn snd
 102                             $ toList
 103                             $ detect (wordsToBook [0..2] s) eventLang
 104
 105 part :: (Eq p, Fractional p) => p -> p -> p
 106 part 0 _ = 0
 107 part _ 0 = 0
 108 part x y = x / y
 109
 110 toProba :: (Eq b, Fractional b, Functor t, Foldable t) =>
 111                  t (a, b) -> t (a, b)
 112 toProba xs = map (\(a,b) -> (a, part b total)) xs
 113   where
 114     total = sum $ map snd xs
 115
 116 textSample :: Lang -> String
 117 textSample EN = EN.textSample
 118 textSample FR = FR.textSample
 119 --textSample DE = DE.textSample
 120 --textSample SP = SP.textSample
 121 --textSample CH = CH.textSample
 122
 123 langWord :: Lang -> LangWord
 124 langWord l = LangWord l (textSample l)
 125
 126 eventLang :: EventLang
 127 eventLang = toEventLangs [0..2] [ langWord l | l <- allLangs ]
 128
 129 detect :: EventBook -> EventLang -> LangProba
 130 detect (EventBook mapFreq _) el =
 131   DM.unionsWith (+)
 132   $ map DM.fromList
 133   $ map (\(s,n) -> map (\(l,f) -> (l, (fromIntegral n) * f)) $ toPrior s el)
 134   $ filter (\x -> fst x /= "  ")
 135   $ DM.toList mapFreq
 136
 137 ------------------------------------------------------------------------
 138 -- | TODO: monoids
 139 type EventLang = Map Lang EventBook
 140
 141 toEventLangs :: [Int] -> [LangWord] -> EventLang
 142 toEventLangs ns = foldl' (opLang (+)) (emptyEventLang ns) . map (toLang ns)
 143
 144 emptyEventLang :: [Int] -> EventLang
 145 emptyEventLang ns = toLang ns (LangWord FR "")
 146
 147 toLang :: [Int] -> LangWord -> EventLang
 148 toLang ns (LangWord l txt) = DM.fromList [(l, wordsToBook ns txt)]
 149
 150 opLang :: (Freq -> Freq -> Freq) -> EventLang -> EventLang -> EventLang
 151 opLang f = DM.unionWith (op f)
 152
 153 ------------------------------------------------------------------------
 154 -- | TODO: monoids (but proba >= 0)
 155
 156 peb :: String -> EventBook -> Double
 157 peb st (EventBook mapFreq mapN) = (fromIntegral a) / (fromIntegral b)
 158   where
 159     a = maybe 0 identity $ DM.lookup st mapFreq
 160     b = maybe 1 identity $ DM.lookup (length st) mapN
 161
 162 peb' :: String -> EventBook -> (Freq, TotalFreq)
 163 peb' st (EventBook mapFreq mapN) = (fromIntegral a, fromIntegral b)
 164   where
 165     a = maybe 0 identity $ DM.lookup st mapFreq
 166     b = maybe 1 identity $ DM.lookup (length st) mapN
 167
 168 ------------------------------------------------------------------------
 169 toPrior :: String -> EventLang -> [(Lang, Double)]
 170 toPrior s el = prior $ pebLang s el
 171
 172 pebLang :: String -> EventLang -> [(Lang, (Freq,TotalFreq))]
 173 pebLang st = map (\(l,eb) -> (l, peb' st eb)) .  DM.toList
 174 ------------------------------------------------------------------------
 175 prior :: [(Lang, (Freq, TotalFreq))] -> [(Lang, Double)]
 176 prior ps = zip ls $ zipWith (\x y -> x^(99::Int) * y) (map (\(a,_) -> part a (sum $ map fst ps')) ps')
 177                                 (map (\(a,b) -> a / b) ps')
 178   where
 179
 180     (ls, ps'') = DL.unzip ps
 181     ps' = map (both fromIntegral) ps''
 182 ------------------------------------------------------------------------
 183 data EventBook = EventBook { events_freq :: Map String     Freq
 184                            , events_n    :: Map StringSize TotalFreq
 185                            }
 186                              deriving (Show)
 187
 188 emptyEventBook :: [Int] -> EventBook
 189 emptyEventBook ns = wordToBook ns " "
 190
 191 wordsToBook :: [Int] -> String -> EventBook
 192 wordsToBook ns txt = foldl' (op (+)) (emptyEventBook ns) eventsBook
 193   where
 194     ws = map unpack $ words $ pack txt
 195     eventsBook = map (wordToBook ns) ws
 196
 197 wordToBook :: [Int] -> Word -> EventBook
 198 wordToBook ns txt = EventBook ef en
 199   where
 200     chks = allChunks' ns 10 txt
 201     en = DM.fromList $ map (\(n,ns') -> (n, length ns')) $ zip ns chks
 202     ef = foldl' DM.union DM.empty $ map (occurrencesWith identity) chks
 203
 204 op :: (Freq -> Freq -> Freq) -> EventBook -> EventBook -> EventBook
 205 op f (EventBook ef1 en1)
 206      (EventBook ef2 en2) = EventBook (DM.unionWith f ef1 ef2)
 207                                      (DM.unionWith f en1 en2)
 208
 209 ------------------------------------------------------------------------
 210 ------------------------------------------------------------------------
 211 -- * Make the distributions
 212 makeDist :: [String] -> D.T Double String
 213 makeDist = D.uniform . DL.concat . map (allChunks [0,2] 10)
 214
 215 stopDist :: D.T Double String
 216 stopDist = makeDist stopList
 217
 218 candDist :: D.T Double String
 219 candDist = makeDist candList
 220
 221 ------------------------------------------------------------------------
 222 sumProba :: Num a => D.T a String -> [Char] -> a
 223 sumProba ds x = sum $ map ((~?) ds) $ allChunks [0,2] 10 $ map toLower x
 224
 225 -- | Get probability according a distribution
 226 (~?) :: (Num prob, Eq a) => D.T prob a -> a -> prob
 227 (~?) ds x = (==x) ?? ds
 228
 229 ------------------------------------------------------------------------
 230 candidate :: [Char] -> Candidate
 231 candidate x = Candidate (sumProba stopDist x) (sumProba candDist x)
 232
 233 ------------------------------------------------------------------------
 234 candList :: [String]
 235 candList = [ "france", "alexandre", "mael", "constitution"
 236            , "etats-unis", "associes", "car", "train", "spam"]
 237
 238
 239 stopList :: [String]
 240 stopList = map show ([0..9]::[Int]) <> [
 241     "a","a's","able","about","above","apply","according","accordingly",
 242     "across","actually","after","afterwards","again","against",
 243     "ain't","all","allow","allows","almost","alone","along",
 244     "involves", "already","also","although","always","am","among","amongst",
 245     "an","and","another","any","anybody","anyhow","anyone","anything",
 246     "anyway","anyways","anywhere","analyze","apart","appear","appreciate","appropriate",
 247     "are","aren't","around","as","aside","ask","asking","associated","at",
 248     "available","away","awfully","based", "b","be","became","because","become",
 249     "becomes","becoming","been","before","beforehand","behind","being",
 250     "believe","below","beside","besides","best","better","between","beyond",
 251     "both","brief","but","by","c","c'mon","c's","came","can","can't","cannot",
 252     "cant","cause","causes","certain","certainly","changes","clearly","co",
 253     "com","come","comes","common","concerning","consequently","consider","considering",
 254     "contain","containing","contains","corresponding","could","couldn't","course",
 255     "currently","d","definitely","described","detects","detecting","despite","did","didn't","different",
 256     "do","does","doesn't","doing","don't","done","down","downwards","during","e",
 257     "each","edu","eg","eight","either","else","elsewhere","enough","entirely",
 258     "especially","et","etc","even","ever","every","everybody","everyone",
 259     "everything","everywhere","ex","exactly","example","except","f","far",
 260     "few","find","fifth","first","five","followed","following","follows","for",
 261     "former","formerly","forth","four","from","further","furthermore","g",
 262     "get","gets","getting","given","gives","go","goes","going","gone","got",
 263     "gotten","greetings","h","had","hadn't","happens","hardly","has","hasn't",
 264     "have","haven't","having","he","he's","hello","help","hence","her","here",
 265     "here's","hereafter","hereby","herein","hereupon","hers","herself","hi",
 266     "him","himself","his","hither","hopefully","how","howbeit","however","i",
 267     "i'd","identify","i'll","i'm","i've","ie","if","ignored","immediate","in","inasmuch",
 268     "inc","indeed","indicate","indicated","indicates","inner","insofar",
 269     "instead","into","inward","is","isn't","it","it'd","it'll","it's","its",
 270     "itself","j","just","k","keep","keeps","kept","know","known","knows","l",
 271     "last","lately","later","latter","latterly","least","less","lest","let",
 272     "let's","like","liked","likely","little","look","looking","looks","ltd",
 273     "m","mainly","many","may","maybe","me","mean","meanwhile","merely","might",
 274     "more","moreover","most","mostly","much","must","my","myself","n",
 275     "name","namely","nd","near","nearly","necessary","need","needs","neither",
 276     "never","nevertheless","new","next","nine","no","nobody","non","none",
 277     "noone","nor","normally","not","nothing","novel","now","nowhere","o",
 278     "obviously","of","off","often","oh","ok","okay","old","on","once","one",
 279     "ones","only","onto","or","other","others","otherwise","ought","our",
 280     "ours","ourselves","out","outside","over","overall","own","p","particular",
 281     "particularly","per","perhaps","placed","please","plus","possible",
 282     "presents","presumably","probably","provides","q","que","quite","qv","r","rather",
 283     "rd","re","really","reasonably","regarding","regardless","regards",
 284     "relatively","respectively","right","s","said","same","saw","say",
 285     "saying","says","second","secondly","see","seeing","seem","seemed",
 286     "seeming","seems","seen","self","selves","sensible","sent","serious",
 287     "seriously","seven","several","shall","she","should","shouldn't","since",
 288     "six","so","some","somebody","somehow","someone","something","sometime",
 289     "sometimes","somewhat","somewhere","soon","sorry","specified","specify",
 290     "specifying","still","sub","such","sup","sure","t","t's","take","taken",
 291     "tell","tends","th","than","thank","thanks","thanx","that","that's",
 292     "thats","the","their","theirs","them","themselves","then","thence","there",
 293     "there's","thereafter","thereby","therefore","therein","theres",
 294     "thereupon","these","they","they'd","they'll","they're","they've",
 295     "think","third","this","thorough","thoroughly","those","though","three",
 296     "through","throughout","thru","thus","to","together","too","took","toward",
 297     "towards","tried","tries","truly","try","trying","twice","two","u","un",
 298     "under","unfortunately","unless","unlikely","until","unto","up","upon",
 299     "us","use","used","useful","uses","using","usually","uucp","v","value",
 300     "various","very","via","viz","vs","w","want","wants","was","wasn't","way",
 301     "we","we'd","we'll","we're","we've","welcome","well","went","were",
 302     "weren't","what","what's","whatever","when","whence","whenever","where",
 303     "where's","whereafter","whereas","whereby","wherein","whereupon",
 304     "wherever","whether","which","while","whither","who","who's","whoever",
 305     "whole","whom","whose","why","will","willing","wish","with","within",
 306     "without","won't","wonder","would","wouldn't","x","y","yes","yet","you",
 307     "you'd","you'll","you're","you've","your","yours","yourself","yourselves",
 308     "z","zero"]
 309
 310
 311
 312