src/Data/Gargantext/Prelude.hs

   1 {-# OPTIONS_GHC -fno-warn-name-shadowing #-}
   2 {-# OPTIONS_GHC -fno-warn-type-defaults  #-}
   3 {-# LANGUAGE     NoImplicitPrelude       #-}
   4
   5 {-
   6 TODO: import head impossible from Protolude: why ?
   7 -}
   8
   9 module Data.Gargantext.Prelude
  10   ( module Data.Gargantext.Prelude
  11   , module Protolude
  12   , headMay
  13   )
  14   where
  15
  16 import Protolude ( Bool(True, False), Int, Double, Integer
  17                  , Fractional, Num, Maybe, Floating, Char
  18                  , Ord, Integral, Foldable, RealFrac, Monad, filter
  19                  , reverse, map, zip, drop, take, zipWith
  20                  , sum, fromIntegral, length, fmap
  21                  , takeWhile, sqrt, undefined, identity
  22                  , abs, maximum, minimum, return, snd, truncate
  23                  , (+), (*), (/), (-), (.), (>=), ($), (**), (^)
  24                  )
  25
  26 -- TODO import functions optimized in Utils.Count
  27 -- import Protolude hiding (head, last, all, any, sum, product, length)
  28 -- import Data.Gargantext.Utils.Count
  29
  30 import qualified Data.List     as L hiding (head, sum)
  31 import qualified Control.Monad as M
  32 import qualified Data.Map as Map
  33 import qualified Data.Vector as V
  34 import Safe (headMay)
  35
  36
  37 pf :: (a -> Bool) -> [a] -> [a]
  38 pf = filter
  39
  40 pr :: [a] -> [a]
  41 pr = reverse
  42
  43 pm :: (a -> b) -> [a] -> [b]
  44 pm = map
  45
  46 pm2 :: (t -> b) -> [[t]] -> [[b]]
  47 pm2 fun = pm (pm fun)
  48
  49 pz :: [a] -> [b] -> [(a, b)]
  50 pz  = zip
  51
  52 pd :: Int -> [a] -> [a]
  53 pd  = drop
  54
  55 ptk :: Int -> [a] -> [a]
  56 ptk = take
  57
  58 pzw :: (a -> b -> c) -> [a] -> [b] -> [c]
  59 pzw = zipWith
  60
  61 -- Exponential Average
  62 eavg :: [Double] -> Double
  63 eavg (x:xs) = a*x + (1-a)*(eavg xs)
  64   where a = 0.70
  65 eavg [] = 0
  66
  67 -- Simple Average
  68 mean :: Fractional a => [a] -> a
  69 mean xs = if L.null xs then 0.0
  70                        else sum xs / fromIntegral (length xs)
  71
  72 sumMaybe :: Num a => [Maybe a] -> Maybe a
  73 sumMaybe = fmap sum . M.sequence
  74
  75 variance :: Floating a => [a] -> a
  76 variance xs = mean $ pm (\x -> (x - m) ** 2) xs where
  77     m = mean xs
  78
  79 deviation :: [Double] -> Double
  80 deviation = sqrt . variance
  81
  82 movingAverage :: Fractional b => Int -> [b] -> [b]
  83 movingAverage steps xs = pm mean $ chunkAlong steps 1 xs
  84
  85 ma :: [Double] -> [Double]
  86 ma = movingAverage 3
  87
  88
  89 -- | Function to split a range into chunks
  90 chunkAlong :: Int -> Int -> [a] -> [[a]]
  91 chunkAlong a b l = only (while  dropAlong)
  92     where
  93         only      = pm (take a)
  94         while     = takeWhile  (\x -> length x >= a)
  95         dropAlong = L.scanl (\x _y -> drop b x) l ([1..] :: [Integer])
  96
  97 -- | Optimized version (Vector)
  98 chunkAlong' :: Int -> Int -> V.Vector a -> V.Vector (V.Vector a)
  99 chunkAlong' a b l = only (while  dropAlong)
 100     where
 101         only      = V.map (V.take a)
 102         while     = V.takeWhile  (\x -> V.length x >= a)
 103         dropAlong = V.scanl (\x _y -> V.drop b x) l (V.fromList [1..])
 104
 105 -- | TODO Inverse of chunk ? unchunkAlong ?
 106 unchunkAlong :: Int -> Int -> [[a]] -> [a]
 107 unchunkAlong = undefined
 108
 109
 110 -- splitAlong [2,3,4] ("helloworld" :: [Char]) == ["he", "llo", "worl", "d"]
 111 splitAlong :: [Int] -> [Char] -> [[Char]]
 112 splitAlong _ [] = [] -- No list? done
 113 splitAlong [] xs = [xs] -- No place to split at? Return the remainder
 114 splitAlong (x:xs) ys = take x ys : splitAlong xs (drop x ys) -- take until our split spot, recurse with next split spot and list remainder
 115
 116 takeWhileM :: (Monad m) => (a -> Bool) -> [m a] -> m [a]
 117 takeWhileM _ [] = return []
 118 takeWhileM p (a:as) = do
 119     v <- a
 120     if p v
 121         then do
 122             vs <- takeWhileM p as
 123             return (v:vs)
 124         else return []
 125
 126 -- SUMS
 127 -- To select the right algorithme according to the type:
 128 -- https://github.com/mikeizbicki/ifcxt
 129
 130 sumSimple :: Num a => [a] -> a
 131 sumSimple = L.foldl' (+) 0
 132
 133 -- | https://en.wikipedia.org/wiki/Kahan_summation_algorithm
 134 sumKahan :: Num a => [a] -> a
 135 sumKahan = snd . L.foldl' go (0,0)
 136     where
 137         go (c,t) i = ((t'-t)-y,t')
 138             where
 139                 y  = i-c
 140                 t' = t+y
 141
 142 -- | compute part of the dict
 143 count2map :: (Ord k, Foldable t) => t k -> Map.Map k Double
 144 count2map xs = Map.map (/ (fromIntegral (length xs))) (count2map' xs)
 145
 146 -- | insert in a dict
 147 count2map' :: (Ord k, Foldable t) => t k -> Map.Map k Double
 148 count2map' xs = L.foldl' (\x y -> Map.insertWith' (+) y 1 x) Map.empty xs
 149
 150
 151 trunc :: (RealFrac a, Integral c, Integral b) => b -> a -> c
 152 trunc n = truncate . (* 10^n)
 153
 154 trunc' :: Int -> Double -> Double
 155 trunc' n x = fromIntegral $ truncate $ (x * 10^n)
 156
 157
 158 bool2int :: Num a => Bool -> a
 159 bool2int b = case b of
 160                   True  -> 1
 161                   False -> 0
 162
 163 bool2double :: Bool -> Double
 164 bool2double bool = case bool of
 165                   True  -> 1.0
 166                   False -> 0.0
 167
 168
 169
 170 -- Normalizing && scaling data
 171 scale :: [Double] -> [Double]
 172 scale = scaleMinMax
 173
 174 scaleMinMax :: [Double] -> [Double]
 175 scaleMinMax xs = pm (\x -> (x - mi / (ma - mi + 1) )) xs'
 176     where
 177         ma  = maximum xs'
 178         mi  = minimum xs'
 179         xs' = pm abs xs
 180
 181 scaleNormalize :: [Double] -> [Double]
 182 scaleNormalize xs = pm (\x -> (x - v / (m + 1))) xs'
 183     where
 184         v = variance xs'
 185         m = mean     xs'
 186         xs' = pm abs xs
 187
 188
 189
 190 normalize :: [Double] -> [Double]
 191 normalize as = normalizeWith identity as
 192
 193 normalizeWith :: Fractional b => (a -> b) -> [a] -> [b]
 194 normalizeWith extract bs = pm (\x -> x/(sum bs')) bs'
 195     where
 196         bs' = pm extract bs
 197
 198 -- Zip functions to add
 199 zipFst :: ([b] -> [a]) -> [b] -> [(a, b)]
 200 zipFst  f xs = zip (f xs) xs
 201
 202 zipSnd :: ([a] -> [b]) -> [a] -> [(a, b)]
 203 zipSnd f xs = zip xs (f xs)