src/Gargantext/Text/Metrics.hs

   1 {-|
   2 Module      : Gargantext.Text.Metrics
   3 Description : All parsers of Gargantext in one file.
   4 Copyright   : (c) CNRS, 2017 - present
   5 License     : AGPL + CECILL v3
   6 Maintainer  : team@gargantext.org
   7 Stability   : experimental
   8 Portability : POSIX
   9
  10 Mainly reexport functions in @Data.Text.Metrics@
  11 -}
  12
  13 {-# LANGUAGE NoImplicitPrelude #-}
  14
  15 module Gargantext.Text.Metrics where
  16
  17 import Data.Text (Text, pack)
  18 import qualified Data.Text as T
  19 import Data.List (concat)
  20
  21 --import GHC.Real (Ratio)
  22 --import qualified Data.Text.Metrics as DTM
  23
  24 import Gargantext.Prelude
  25
  26 import Gargantext.Text.Metrics.Count (occurrences, cooc)
  27 import Gargantext.Text.Terms (TermType(Multi), terms)
  28 import Gargantext.Core (Lang(EN))
  29 import Gargantext.Core.Types (Terms(..))
  30 import Gargantext.Text.Context (splitBy, SplitContext(Sentences))
  31
  32 --noApax :: Ord a => Map a Occ -> Map a Occ
  33 --noApax m = M.filter (>1) m
  34
  35
  36 metrics_text :: Text
  37 metrics_text = T.concat ["A table is an object."
  38                         ,"A glas is an object too."
  39                         ,"Using a glas to dring is a function."
  40                         ,"Using a spoon to eat is a function."
  41                         ,"The spoon is an object to eat."
  42                         ]
  43
  44 metrics_sentences' :: [Text]
  45 metrics_sentences' = splitBy (Sentences 0) metrics_text
  46
  47 -- | Sentences
  48 metrics_sentences :: [Text]
  49 metrics_sentences = ["A table is an object."
  50                     ,"A glas is an object too."
  51                     ,"The glas and the spoon are on the table."
  52                     ,"The spoon is an object to eat."
  53                     ,"The spoon is on the table and the plate and the glas."]
  54
  55
  56 metrics_sentences_Test = metrics_sentences == metrics_sentences'
  57
  58 -- | Terms reordered to visually check occurrences
  59 metrics_terms :: [[Text]]
  60 metrics_terms = undefined
  61
  62 metrics_terms' :: IO [[Terms]]
  63 metrics_terms' = mapM (terms Multi EN) $ splitBy (Sentences 0) metrics_text
  64
  65 --metrics_terms_Test =  metrics_terms == ((map _terms_label) <$> metrics_terms')
  66
  67 -- | Occurrences
  68 {-
  69 fromList [ (fromList ["table"] ,fromList [(["table"] , 3 )])]
  70          , (fromList ["object"],fromList [(["object"], 3 )])
  71          , (fromList ["glas"]  ,fromList [(["glas"]  , 2 )])
  72          , (fromList ["spoon"] ,fromList [(["spoon"] , 2 )])
  73 -}
  74 metrics_occ = occurrences <$> concat <$> (mapM (terms Multi EN) $ splitBy (Sentences 0) metrics_text)
  75
  76 {-
  77 -- fromList [((["glas"],["object"]),6)
  78             ,((["glas"],["spoon"]),4)
  79             ,((["glas"],["table"]),6),((["object"],["spoon"]),6),((["object"],["table"]),9),((["spoon"],["table"]),6)]
  80
  81 -}
  82 metrics_cooc = cooc <$> (mapM (terms Multi EN) $ splitBy (Sentences 0) metrics_text)
  83
  84 metrics_cooc' = (mapM (terms Multi EN) $ splitBy (Sentences 0) "The table object. The table object.")
  85
  86
  87