src/Gargantext/Text/Flow.hs

   1 {-|
   2 Module      : Gargantext.Text.Flow
   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 From text to viz, all the flow of texts in Gargantext.
  11
  12 -}
  13
  14 {-# LANGUAGE NoImplicitPrelude           #-}
  15 {-# LANGUAGE OverloadedStrings           #-}
  16
  17 module Gargantext.Text.Flow
  18   where
  19
  20 import qualified Data.Text as T
  21 --import Data.Text.IO (readFile)
  22 import Database.PostgreSQL.Simple (Connection)
  23 import GHC.IO (FilePath)
  24 --import Gargantext.Core (Lang)
  25 import Gargantext.Core.Types (CorpusId)
  26
  27 {-
  28   ____                             _            _
  29  / ___| __ _ _ __ __ _  __ _ _ __ | |_ _____  _| |_
  30 | |  _ / _` | '__/ _` |/ _` | '_ \| __/ _ \ \/ / __|
  31 | |_| | (_| | | | (_| | (_| | | | | ||  __/>  <| |_
  32  \____|\__,_|_|  \__, |\__,_|_| |_|\__\___/_/\_\\__|
  33                  |___/
  34 -}
  35
  36
  37 contextText :: [T.Text]
  38 contextText = ["The dog is an animal."
  39               ,"The bird is an animal."
  40               ,"The dog is an animal."
  41               ,"The animal is a bird or a dog ?"
  42               ,"The table is an object."
  43               ,"The pen is an object."
  44               ,"The object is a pen or a table ?"
  45               ,"The girl is a human."
  46               ,"The boy  is a human."
  47               ,"The boy or the girl are human."
  48               ]
  49
  50
  51 -- | Control the flow of text
  52 data TextFlow = CSV FilePath
  53               | FullText FilePath
  54               | Contexts [T.Text]
  55               | DBV3 Connection CorpusId
  56               | Query T.Text
  57
  58 {-
  59 textFlow :: TermType Lang -> TextFlow -> IO Graph
  60 textFlow termType workType = do
  61   contexts <- case workType of
  62                 FullText path -> splitBy (Sentences 5) <$> readFile path
  63                 CSV      path -> readCsvOn [csv_title, csv_abstract] path
  64                 Contexts ctxt -> pure ctxt
  65                 DBV3 con corpusId -> catMaybes <$> map (\n -> hyperdataDocumentV3_title (_node_hyperdata n)  <> hyperdataDocumentV3_abstract (_node_hyperdata n))<$> runReaderT (getDocumentsV3WithParentId corpusId) con
  66                 _             -> undefined -- TODO Query not supported
  67
  68   textFlow' termType contexts
  69
  70
  71 textFlow' :: TermType Lang -> [T.Text] -> IO Graph
  72 textFlow' termType contexts = do
  73   -- Context :: Text -> [Text]
  74   -- Contexts = Paragraphs n | Sentences n | Chars n
  75
  76   myterms <- extractTerms termType contexts
  77   -- TermsType = Mono | Multi | MonoMulti
  78   -- myterms # filter (\t -> not . elem t stopList)
  79   --         # groupBy (Stem|GroupList|Ontology)
  80   --printDebug "terms" myterms
  81   --printDebug "myterms" (sum $ map length myterms)
  82
  83   -- Bulding the map list
  84   -- compute copresences of terms, i.e. cooccurrences of terms in same context of text
  85   -- Cooc = Map (Term, Term) Int
  86   let myCooc1 = coocOn (_terms_label) myterms
  87   --printDebug "myCooc1 size" (M.size myCooc1)
  88
  89   -- Remove Apax: appears one time only => lighting the matrix
  90   let myCooc2 = Map.filter (>0) myCooc1
  91   --printDebug "myCooc2 size" (M.size myCooc2)
  92   --printDebug "myCooc2" myCooc2
  93   g <- cooc2graph myCooc2
  94   pure g
  95 -}
  96