src-test/Core/Text/Flow.hs

   1 {-|
   2 Module      : Gargantext.Core.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
  15 module Core.Text.Flow
  16   where
  17
  18 {-
  19 import qualified Data.Text as T
  20 --import Data.Text.IO (readFile)
  21 import Database.PostgreSQL.Simple (Connection)
  22 import GHC.IO (FilePath)
  23 --import Gargantext.Core (Lang)
  24 import Gargantext.Core.Types (CorpusId)
  25 -}
  26
  27 {-
  28   ____                           _____         _
  29  / ___| __ _ _ __ __ _  __ _ _ _|_   _|____  _| |_
  30 | |  _ / _` | '__/ _` |/ _` | '_ \| |/ _ \ \/ / __|
  31 | |_| | (_| | | | (_| | (_| | | | | |  __/>  <| |_
  32  \____|\__,_|_|  \__, |\__,_|_| |_|_|\___/_/\_\\__|
  33                  |___/
  34 -}
  35
  36
  37 {-
  38 contextText :: [T.Text]
  39 contextText = ["The dog is an animal."
  40               ,"The bird is an animal."
  41               ,"The dog is an animal."
  42               ,"The animal is a bird or a dog ?"
  43               ,"The table is an object."
  44               ,"The pen is an object."
  45               ,"The object is a pen or a table ?"
  46               ,"The girl is a human."
  47               ,"The boy  is a human."
  48               ,"The boy or the girl are human."
  49               ]
  50
  51
  52 -- | Control the flow of text
  53 data TextFlow = CSV FilePath
  54               | FullText FilePath
  55               | Contexts [T.Text]
  56               | DBV3 Connection CorpusId
  57               | Query T.Text
  58
  59 -}
  60 {-
  61 textFlow :: TermType Lang -> TextFlow -> IO Graph
  62 textFlow termType workType = do
  63   contexts <- case workType of
  64                 FullText path -> splitBy (Sentences 5) <$> readFile path
  65                 CSV      path -> readCsvOn [csv_title, csv_abstract] path
  66                 Contexts ctxt -> pure ctxt
  67                 DBV3 con corpusId -> catMaybes <$> map (\n -> hyperdataDocumentV3_title (_node_hyperdata n)  <> hyperdataDocumentV3_abstract (_node_hyperdata n))<$> runReaderT (getDocumentsV3WithParentId corpusId) con
  68                 _             -> undefined -- TODO Query not supported
  69
  70   textFlow' termType contexts
  71
  72
  73 textFlow' :: TermType Lang -> [T.Text] -> IO Graph
  74 textFlow' termType contexts = do
  75   -- Context :: Text -> [Text]
  76   -- Contexts = Paragraphs n | Sentences n | Chars n
  77
  78   myterms <- extractTerms termType contexts
  79   -- TermsType = Mono | Multi | MonoMulti
  80   -- myterms # filter (\t -> not . elem t stopList)
  81   --         # groupBy (Stem|GroupList|Ontology)
  82   --printDebug "terms" myterms
  83   --printDebug "myterms" (sum $ map length myterms)
  84
  85   -- Bulding the map list
  86   -- compute copresences of terms, i.e. cooccurrences of terms in same context of text
  87   -- Cooc = Map (Term, Term) Int
  88   let myCooc1 = coocOn (_terms_label) myterms
  89   --printDebug "myCooc1 size" (M.size myCooc1)
  90
  91   -- Remove Apax: appears one time only => lighting the matrix
  92   let myCooc2 = Map.filter (>0) myCooc1
  93   --printDebug "myCooc2 size" (M.size myCooc2)
  94   --printDebug "myCooc2" myCooc2
  95   g <- cooc2graph myCooc2
  96   pure g
  97 -}