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 TODO
  14 noApax :: Ord a => Map a Occ -> Map a Occ
  15 noApax m = M.filter (>1) m
  16
  17 -}
  18
  19 {-# LANGUAGE BangPatterns      #-}
  20 {-# LANGUAGE NoImplicitPrelude #-}
  21 {-# LANGUAGE OverloadedStrings #-}
  22
  23 module Gargantext.Text.Metrics
  24   where
  25
  26 import Data.Text (Text, pack)
  27 import Data.Map (Map)
  28 import qualified Data.List as L
  29 import qualified Data.Map  as M
  30 import qualified Data.Set  as S
  31 import qualified Data.Text as T
  32 import qualified Data.Vector as V
  33 import qualified Data.Vector.Unboxed as VU
  34 import Data.Tuple.Extra (both)
  35 --import GHC.Real (Ratio)
  36 --import qualified Data.Text.Metrics as DTM
  37 import Data.Array.Accelerate (toList)
  38 import Math.KMeans (kmeans, euclidSq, elements)
  39
  40
  41 import Gargantext.Prelude
  42
  43 import Gargantext.Text.Metrics.Count (occurrences, cooc)
  44 import Gargantext.Text.Terms (TermType(MonoMulti), terms)
  45 import Gargantext.Core (Lang(EN))
  46 import Gargantext.Core.Types (Terms(..))
  47 import Gargantext.Text.Context (splitBy, SplitContext(Sentences))
  48
  49 import Gargantext.Viz.Graph.Distances.Matrice
  50 import Gargantext.Viz.Graph.Index
  51
  52 import qualified Data.Array.Accelerate.Interpreter as DAA
  53 import qualified Data.Array.Accelerate as DAA
  54
  55 import GHC.Real (round)
  56
  57 filterCooc :: Ord t => Map (t, t) Int -> Map (t, t) Int
  58 filterCooc cc = filterCooc' ts cc
  59   where
  60     ts     = map _scored_terms $ takeSome 350 5 2 $ coocScored cc
  61
  62 filterCooc' :: Ord t => [t] -> Map (t, t) Int -> Map (t, t) Int
  63 filterCooc' ts m = foldl' (\m' k -> M.insert k (maybe errMessage identity $ M.lookup k m) m') M.empty selection
  64   where
  65     errMessage = panic "Filter cooc: no key"
  66     selection  = [(x,y) | x <- ts, y <- ts, x > y]
  67
  68
  69 type MapListSize = Int
  70 type SampleBins  = Double
  71 type Clusters    = Int
  72
  73 -- | Map list creation
  74 -- Kmeans split into (Clusters::Int) main clusters with Inclusion/Exclusion (relevance score)
  75 -- Sample the main cluster ordered by specificity/genericity in (SampleBins::Double) parts
  76 -- each parts is then ordered by Inclusion/Exclusion
  77 -- take n scored terms in each parts where n * SampleBins = MapListSize.
  78 takeSome :: Ord t => MapListSize -> SampleBins -> Clusters -> [Scored t] -> [Scored t]
  79 takeSome l s k scores = L.take l
  80                     $ takeSample n m
  81                     $ splitKmeans k scores
  82   where
  83     -- TODO: benchmark with accelerate-example kmeans version
  84     splitKmeans x xs = elements
  85                      $ V.head
  86                      $ kmeans (\i -> VU.fromList ([(_scored_incExc i :: Double)]))
  87                               euclidSq x xs
  88     n = round ((fromIntegral l)/s)
  89     m = round $ (fromIntegral $ length scores) / (s)
  90     takeSample n m xs = L.concat $ map (L.take n)
  91                                  $ L.reverse $ map (L.sortOn _scored_incExc)
  92                                  -- TODO use kmeans s instead of splitEvery
  93                                  -- in order to split in s heteregenous parts
  94                                  -- without homogeneous order hypothesis
  95                                  $ splitEvery m
  96                                  $ L.reverse $ L.sortOn _scored_speGen xs
  97
  98
  99 data Scored t = Scored { _scored_terms  :: !t
 100                        , _scored_incExc :: !InclusionExclusion
 101                        , _scored_speGen :: !SpecificityGenericity
 102                      } deriving (Show)
 103
 104 coocScored :: Ord t => Map (t,t) Int -> [Scored t]
 105 coocScored m = zipWith (\(i,t) (inc,spe) -> Scored t inc spe) (M.toList fi) scores
 106   where
 107     (ti,fi) = createIndices m
 108     (is, ss) = incExcSpeGen $ cooc2mat ti m
 109     scores = DAA.toList $ DAA.run $ DAA.zip (DAA.use is) (DAA.use ss)
 110
 111
 112
 113
 114
 115
 116
 117
 118
 119
 120
 121
 122
 123
 124
 125
 126
 127
 128
 129
 130 incExcSpeGen_sorted :: Ord t => Map (t,t) Int -> ([(t,Double)],[(t,Double)])
 131 incExcSpeGen_sorted m = both ordonne (incExcSpeGen $ cooc2mat ti m)
 132   where
 133     (ti,fi) = createIndices m
 134     ordonne x = L.reverse $ L.sortOn snd $ zip (map snd $ M.toList fi) (toList x)
 135
 136
 137
 138 metrics_text :: Text
 139 metrics_text = T.intercalate " " metrics_sentences
 140
 141 metrics_sentences' :: [Text]
 142 metrics_sentences' = splitBy (Sentences 0) metrics_text
 143
 144 -- | Sentences
 145 metrics_sentences :: [Text]
 146 metrics_sentences = [ "There is a table with a glass of wine and a spoon."
 147                     , "I can see the glass on the table."
 148                     , "There was only a spoon on that table."
 149                     , "The glass just fall from the table, pouring wine everywhere."
 150                     , "I wish the glass did not contain wine."
 151                     ]
 152
 153 metrics_sentences_Test = metrics_sentences == metrics_sentences'
 154
 155 -- | Terms reordered to visually check occurrences
 156 -- >>>
 157 {- [ [["table"],["glass"],["wine"],["spoon"]]
 158    , [["glass"],["table"]]
 159    , [["spoon"],["table"]]
 160    , [["glass"],["table"],["wine"]]
 161    , [["glass"],["wine"]]
 162    ]
 163 -}
 164
 165 metrics_terms :: IO [[Terms]]
 166 metrics_terms = mapM (terms MonoMulti EN) $ splitBy (Sentences 0) metrics_text
 167
 168 -- | Occurrences
 169 {-
 170 fromList [ (fromList ["table"] ,fromList [(["table"] , 3 )])]
 171          , (fromList ["object"],fromList [(["object"], 3 )])
 172          , (fromList ["glas"]  ,fromList [(["glas"]  , 2 )])
 173          , (fromList ["spoon"] ,fromList [(["spoon"] , 2 )])
 174 -}
 175 metrics_occ = occurrences <$> L.concat <$> metrics_terms
 176
 177 {-
 178 -- fromList [((["glas"],["object"]),6)
 179             ,((["glas"],["spoon"]),4)
 180             ,((["glas"],["table"]),6),((["object"],["spoon"]),6),((["object"],["table"]),9),((["spoon"],["table"]),6)]
 181
 182 -}
 183 metrics_cooc = cooc <$> metrics_terms
 184
 185 metrics_cooc_mat = do
 186   m <- metrics_cooc
 187   let (ti,_) = createIndices m
 188   let mat_cooc = cooc2mat ti m
 189   pure ( ti
 190        , mat_cooc
 191        , incExcSpeGen_proba  mat_cooc
 192        , incExcSpeGen        mat_cooc
 193        )
 194
 195 metrics_incExcSpeGen = incExcSpeGen_sorted <$> metrics_cooc
 196