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