src/Gargantext/Core/Text/Metrics/Hetero.purs

   1 {-|
   2 Module      : Gargantext.
   3 Description :
   4 Copyright   : (c) CNRS, 2017-Present
   5 License     : AGPL + CECILL v3
   6 Maintainer  : team@gargantext.org
   7 Stability   : experimental
   8 Portability : POSIX
   9
  10 Here is a longer description of this module, containing some
  11 commentary with @some markup@.
  12 -}
  13
  14 module Gargantext.Text.Hetero where
  15
  16 import Data.List.Split as S
  17 import Data.Map as M
  18 import Data.Set as S
  19 import Database.PostgreSQL.Simple as PGS
  20 import GHC.Real as R
  21 import Gargantext.Database.Admin.Gargandb
  22 import Gargantext.Database.Admin.Private
  23 import Gargantext.Database.Simple
  24 import Gargantext.Text.Count (occurrences)
  25 import Gargantext.Text.Words (cleanText)
  26 import Opaleye.Internal.Column (Column)
  27 import Opaleye.PGTypes (PGInt4)
  28
  29 --main = do
  30 --    t <- getTextquery
  31 --    print (Prelude.map (heterogeinity . concat) $ S.chunksOf 3 t)
  32
  33 -- heterogeinity sur concat texts
  34 heterogeinity' :: Int -> Int -> Int -> IO [Integer]
  35 heterogeinity' corpus_id limit x = do
  36     t <- getAbstract corpus_id limit
  37     Prelude.mapM (dicoStruct . occurrences) $ (S.chunksOf x) .  cleanText $ concat t
  38
  39 heterogeinity'' :: Int -> Int -> Int -> IO [Integer]
  40 heterogeinity'' corpus_id limit size = do
  41     t <- getAbstract corpus_id limit
  42     Prelude.mapM (dicoStruct . occurrences) $ (S.chunksOf size) .  cleanText $ concat t
  43
  44
  45 dicoStruct :: (Integral r, Monad m) => M.Map t r -> m r
  46 dicoStruct dict_occ = do
  47     let keys_size = toInteger $ length $ M.keys dict_occ
  48     let total_occ = sum $ Prelude.map (\(x, y) -> y) $ M.toList dict_occ
  49     return $ div total_occ (fromIntegral keys_size)
  50
  51 -- heterogeinity sur UCT (Unité de Context Textuel)
  52 heterogeinity :: [Char] -> IO Integer
  53 heterogeinity string = do
  54     let dict_occ = occurrences $ cleanText string
  55
  56     let keys_size = toInteger $ length $ M.keys dict_occ
  57     let total_occ = sum $ Prelude.map (\(x, y) -> y) $ M.toList dict_occ
  58
  59     return $ div total_occ (fromIntegral keys_size)
  60
  61
  62 --computeHeterogeinity
  63 --  :: Fractional t =>
  64 --       Opaleye.Internal.Column.Column Opaleye.PGTypes.PGInt4
  65 --            -> IO (t, Integer, Integer)
  66 computeHeterogeinity corpus_id = do
  67     c <- PGS.connect infoGargandb
  68     t <- getText c (nodeHyperdataText corpus_id)
  69     heterogeinity $ Prelude.concat t
  70
  71 main2 = do
  72     let corpus_ids = [
  73                 ("ALL", 272927) -- 73
  74                ,("Histoire", 1387736) -- 28
  75                ,("Sciences Po", 1296892) -- 37
  76                ,("Phylosophie", 1170004) -- 20
  77                ,("Psychologie", 1345852) -- 37
  78                ,("Sociologie", 1246452)  -- 42
  79                ]
  80
  81     r <- Prelude.map computeHeterogeinity $ Prelude.map (\(t,id) -> id) corpus_ids
  82     return r
  83
  84