src/Gargantext/Pipeline.hs

   1 {-|
   2 Module      : Gargantext.Pipeline
   3 Description : Server API
   4 Copyright   : (c) CNRS, 2017-Present
   5 License     : AGPL + CECILL v3
   6 Maintainer  : team@gargantext.org
   7 Stability   : experimental
   8 Portability : POSIX
   9
  10 -}
  11
  12 {-# OPTIONS_GHC -fno-warn-name-shadowing #-}
  13 {-# LANGUAGE NoImplicitPrelude           #-}
  14
  15 module Gargantext.Pipeline
  16   where
  17
  18 import Data.Text.IO (readFile)
  19 ----------------------------------------------
  20 import Gargantext.Core (Lang(FR))
  21 import Gargantext.Prelude
  22
  23 import Gargantext.Viz.Graph.Index (score)
  24 import Gargantext.Viz.Graph.Distances.Matrice (distributional)
  25 import Gargantext.Text.Metrics.Occurrences (cooc, removeApax)
  26 import Gargantext.Text.Terms (TermType(Multi), extractTerms)
  27 import Gargantext.Text.Context (splitBy, SplitContext(Sentences))
  28
  29
  30 pipeline path = do
  31   -- Text  <- IO Text <- FilePath
  32   text     <- readFile path
  33   let contexts = splitBy (Sentences 3) text
  34   myterms <- extractTerms Multi FR contexts
  35   -- TODO    filter (\t -> not . elem t stopList) myterms
  36   -- TODO    groupBy (Stem | GroupList)
  37   let myCooc = removeApax $ cooc myterms
  38   -- Cooc -> Matrix
  39   pure $ score distributional myCooc
  40   -- Matrix -> Clustering -> Graph -> JSON
  41