src/Gargantext/Ngrams/FrequentItemSet.hs

   1 {-|
   2 Module      : Gargantext.Ngrams.FrequentItemSet
   3 Description : Ngrams tools
   4 Copyright   : (c) CNRS, 2018
   5 License     : AGPL + CECILL v3
   6 Maintainer  : team@gargantext.org
   7 Stability   : experimental
   8 Portability : POSIX
   9
  10 Domain Specific Language to manage Frequent Item Set (FIS)
  11
  12 -}
  13
  14 module Gargantext.Ngrams.FrequentItemSet
  15   ( Fis, Size
  16   , occ, cooc
  17   , all, between
  18   , module HLCM
  19   )
  20   where
  21
  22 import Data.List (tail, filter)
  23 import Data.Either
  24
  25 import HLCM
  26
  27 import Gargantext.Prelude
  28
  29 type Size = Either Int (Int, Int)
  30
  31
  32 ------------------------------------------------------------------------
  33 -- | Occurrence is Frequent Item Set of size 1
  34 occ  :: Frequency -> [[Item]] -> [Fis]
  35 occ f is  = fisWithSize (Left 1) f is
  36
  37 -- | Cooccurrence is Frequent Item Set of size 2
  38 cooc :: Frequency -> [[Item]] -> [Fis]
  39 cooc f is = fisWithSize (Left 2) f is
  40
  41 all :: Frequency -> [[Item]] -> [Fis]
  42 all f is  = fisWith Nothing f is
  43
  44 ------------------------------------------------------------------------
  45 between :: (Int, Int) -> Frequency -> [[Item]] -> [Fis]
  46 between (x,y) f is = fisWithSize (Right (x,y)) f is
  47
  48 --maximum :: Int -> Frequency -> [[Item]] -> [Fis]
  49 --maximum m f is = between (0,m) f is
  50
  51
  52 ------------------------------------------------------------------------
  53 ------------------------------------------------------------------------
  54 -- | Data type to type the Frequent Item Set
  55 -- TODO replace List with Set in fisItemSet
  56 -- be careful : risks to erase HLCM behavior
  57 type Fis = Fis' Item
  58 data Fis' a = Fis' { _fisCount   :: Int
  59                    , _fisItemSet :: [a]
  60                    } deriving (Show)
  61
  62 -- | Sugar from items to FIS
  63 items2fis :: [Item] -> Maybe Fis
  64 items2fis is = case head is of
  65                  Nothing -> Nothing
  66                  Just h  -> Just (Fis' h (tail is))
  67
  68 ------------------------------------------------------------------------
  69 ------------------------------------------------------------------------
  70
  71 fisWithSize :: Size -> Frequency -> [[Item]] -> [Fis]
  72 fisWithSize n f is = case n of
  73   Left   n'   -> fisWith (Just (\x -> length x == (n'+1)    )) f is
  74   Right (a,b) -> fisWith (Just (\x -> cond1 a x && cond2 b x)) f is
  75                     where
  76                       cond1 a' x = length x >= a'
  77                       cond2 b' x = length x <= b'
  78
  79
  80 fisWith :: Maybe ([Item] -> Bool) -> Frequency -> [[Item]] -> [Fis]
  81 fisWith s f is = unMaybe $ map items2fis $ filter' $ runLCMmatrix is f
  82   where
  83     filter' = case s of
  84                 Nothing  -> identity
  85                 Just fun -> filter fun
  86
  87 ------------------------------------------------------------------------
  88 ------------------------------------------------------------------------