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 {-# OPTIONS_GHC -fno-warn-name-shadowing #-}
  12 {-# LANGUAGE NoImplicitPrelude           #-}
  13
  14 module Gargantext.Pipeline
  15   where
  16
  17 import Data.Text.IO (readFile)
  18
  19 import Control.Arrow ((***))
  20 import Data.Map.Strict (Map)
  21 import qualified Data.Map.Strict as M
  22 import qualified Data.List       as L
  23 import Data.Tuple.Extra (both)
  24 ----------------------------------------------
  25 import Gargantext.Core (Lang(FR))
  26 import Gargantext.Prelude
  27
  28 import Gargantext.Viz.Graph.Index (score, createIndices, toIndex, fromIndex, cooc2mat, mat2map)
  29 import Gargantext.Viz.Graph.Distances.Matrice (conditional', conditional)
  30 import Gargantext.Viz.Graph.Index (Index)
  31 import Gargantext.Viz.Graph (Graph)
  32 import Gargantext.Text.Metrics.Count (cooc, removeApax)
  33 import Gargantext.Text.Metrics
  34 import Gargantext.Text.Terms (TermType(Multi, Mono), extractTerms)
  35 import Gargantext.Text.Context (splitBy, SplitContext(Sentences))
  36
  37 import Data.Graph.Clustering.Louvain.CplusPlus (cLouvain, LouvainNode)
  38
  39
  40 {-
  41   ____                             _            _
  42  / ___| __ _ _ __ __ _  __ _ _ __ | |_ _____  _| |_
  43 | |  _ / _` | '__/ _` |/ _` | '_ \| __/ _ \ \/ / __|
  44 | |_| | (_| | | | (_| | (_| | | | | ||  __/>  <| |_
  45  \____|\__,_|_|  \__, |\__,_|_| |_|\__\___/_/\_\\__|
  46                  |___/
  47
  48 -}
  49
  50 -----------------------------------------------------------
  51 data2graph :: Map (Int, Int) Int -> Map (Int, Int) Double -> [LouvainNode] -> Graph
  52 data2graph = undefined
  53 -----------------------------------------------------------
  54
  55
  56 workflow lang path = do
  57   -- Text  <- IO Text <- FilePath
  58   text     <- readFile path
  59   let contexts = splitBy (Sentences 5) text
  60   myterms <- extractTerms Multi lang contexts
  61
  62   -- TODO    filter (\t -> not . elem t stopList) myterms
  63   -- TODO    groupBy (Stem | GroupList)
  64
  65   -- @np FIXME optimization issue of filterCooc (too much memory consumed)
  66   let myCooc = filterCooc $ removeApax $ cooc myterms
  67   --pure myCooc
  68   -- Cooc -> Matrix
  69   let (ti, _) = createIndices myCooc
  70   -- Matrix -> Clustering
  71   let distance   = score conditional $ toIndex ti myCooc
  72   partitions <- cLouvain distance
  73   --pure partitions
  74 ---- | Building : -> Graph -> JSON
  75   pure partitions
  76   --pure $ data2graph myCooc distance partitions
  77