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 --data Size = Point | Segment
  32
  33 ------------------------------------------------------------------------
  34 -- | Occurrence is Frequent Item Set of size 1
  35 occ  :: Frequency -> [[Item]] -> [Fis]
  36 occ f is  = fisWithSize (Left 1) f is
  37
  38 -- | Cooccurrence is Frequent Item Set of size 2
  39 cooc :: Frequency -> [[Item]] -> [Fis]
  40 cooc f is = fisWithSize (Left 2) f is
  41
  42 all :: Frequency -> [[Item]] -> [Fis]
  43 all f is  = fisWith Nothing f is
  44
  45 ------------------------------------------------------------------------
  46 between :: (Int, Int) -> Frequency -> [[Item]] -> [Fis]
  47 between (x,y) f is = fisWithSize (Right (x,y)) f is
  48
  49 --maximum :: Int -> Frequency -> [[Item]] -> [Fis]
  50 --maximum m f is = between (0,m) f is
  51
  52
  53 ------------------------------------------------------------------------
  54 ------------------------------------------------------------------------
  55 -- | Data type to type the Frequent Item Set
  56 -- TODO replace List with Set in fisItemSet
  57 -- be careful : risks to erase HLCM behavior
  58 type Fis = Fis' Item
  59 data Fis' a = Fis' { _fisCount   :: Int
  60                    , _fisItemSet :: [a]
  61                    } deriving (Show)
  62
  63 -- | Sugar from items to FIS
  64 items2fis :: [Item] -> Maybe Fis
  65 items2fis is = case head is of
  66                  Nothing -> Nothing
  67                  Just h  -> Just (Fis' h (tail is))
  68
  69 ------------------------------------------------------------------------
  70 ------------------------------------------------------------------------
  71
  72 fisWithSize :: Size -> Frequency -> [[Item]] -> [Fis]
  73 fisWithSize n f is = case n of
  74   Left   n'   -> fisWith (Just (\x -> length x == (n'+1)    )) f is
  75   Right (a,b) -> fisWith (Just (\x -> cond1 a x && cond2 b x)) f is
  76                     where
  77                       cond1 a' x = length x >= a'
  78                       cond2 b' x = length x <= b'
  79
  80
  81 fisWith :: Maybe ([Item] -> Bool) -> Frequency -> [[Item]] -> [Fis]
  82 fisWith s f is = unMaybe $ map items2fis $ filter' $ runLCMmatrix is f
  83   where
  84     filter' = case s of
  85                 Nothing  -> identity
  86                 Just fun -> filter fun
  87
  88 ------------------------------------------------------------------------
  89 ------------------------------------------------------------------------