src/Gargantext/Core/Text/Terms/Multi/Lang/Fr.hs

   1 {-|
   2 Module      : Gargantext.Core.Text.Terms.Multi.Lang.Fr
   3 Description : French Grammar rules to group postag tokens.
   4 Copyright   : (c) CNRS, 2017-Present
   5 License     : AGPL + CECILL v3
   6 Maintainer  : team@gargantext.org
   7 Stability   : experimental
   8 Portability : POSIX
   9
  10 This @group@ function groups horizontally ngrams in their context of
  11 sentence according to grammars specific of each language. In english, JJ
  12 is ADJectiv in french.
  13
  14 -}
  15
  16
  17 module Gargantext.Core.Text.Terms.Multi.Lang.Fr (groupTokens)
  18   where
  19
  20 import Gargantext.Prelude
  21 import Gargantext.Core.Types
  22 import Gargantext.Core.Text.Terms.Multi.Group (group2)
  23
  24 groupTokens :: [TokenTag] -> [TokenTag]
  25 groupTokens [] = []
  26 groupTokens ntags = group2 NP NP
  27             $ group2 NP VB
  28             -- group2 NP IN
  29             -- group2 IN DT
  30             $ group2 VB NP
  31             $ group2 JJ NP
  32             $ group2 NP JJ
  33             $ group2 JJ JJ
  34             -- group2 JJ CC
  35             $ ntags
  36
  37 ------------------------------------------------------------------------
  38 -- TODO
  39 --groupNgrams ((x,_,"PERSON"):(y,yy,"PERSON"):xs)             = groupNgrams ((x <> " " <> y,yy,"PERSON"):xs)
  40 --groupNgrams ((x,_,"ORGANIZATION"):(y,yy,"ORGANIZATION"):xs) = groupNgrams ((x <> " " <> y,yy,"ORGANIZATION"):xs)