src/Gargantext/Core/Text/List/Learn.hs

   1 {-|
   2 Module      : Gargantext.Core.Text.List.Learn
   3 Description : Learn to make lists
   4 Copyright   : (c) CNRS, 2018-Present
   5 License     : AGPL + CECILL v3
   6 Maintainer  : team@gargantext.org
   7 Stability   : experimental
   8 Portability : POSIX
   9
  10 CSV parser for Gargantext corpus files.
  11
  12 -}
  13
  14 {-# OPTIONS_GHC -fno-warn-orphans #-}
  15
  16
  17 module Gargantext.Core.Text.List.Learn
  18   where
  19
  20 import qualified Data.IntMap as IntMap
  21 import qualified Data.List   as List
  22 import Data.Map (Map)
  23 import qualified Data.Map    as Map
  24 import qualified Data.SVM    as SVM
  25 import qualified Data.Vector as Vec
  26
  27 import Gargantext.Core
  28 import Gargantext.Core.Text.Metrics.Count (occurrencesWith)
  29 import Gargantext.Core.Types.Main (ListType(..))
  30 import Gargantext.Prelude
  31 import Gargantext.Database.GargDB
  32
  33 ------------------------------------------------------------------------
  34 train :: Double -> Double -> SVM.Problem -> IO SVM.Model
  35 train x y = (SVM.train (SVM.CSvc x) (SVM.RBF y))
  36
  37 predict :: SVM.Model -> [Vec.Vector Double] -> IO [Double]
  38 predict m vs = mapM (predict' m) vs
  39   where
  40     predict' m' vs' = SVM.predict m' (IntMap.fromList $ (zip [1..]) $ Vec.toList vs')
  41
  42 ------------------------------------------------------------------------
  43 trainList :: Double -> Double -> Map ListType [Vec.Vector Double] -> IO SVM.Model
  44 trainList x y = (train x y) . trainList'
  45   where
  46     trainList' :: Map ListType [Vec.Vector Double] -> SVM.Problem
  47     trainList' = mapVec2problem . (Map.mapKeys (fromIntegral . toDBid))
  48
  49     mapVec2problem :: Map Double [Vec.Vector Double] -> SVM.Problem
  50     mapVec2problem = List.concat . (map (\(a,as) -> zip (repeat a) as)) . Map.toList . (Map.map vecs2maps)
  51
  52     vecs2maps :: [Vec.Vector Double] -> [IntMap.IntMap Double]
  53     vecs2maps = map (IntMap.fromList . (zip [1..]) . Vec.toList)
  54
  55
  56 predictList :: Model -> [Vec.Vector Double] -> IO [Maybe ListType]
  57 predictList (ModelSVM m _ _) vs = map (Just . fromDBid . round) <$> predict m vs
  58
  59 ------------------------------------------------------------------------
  60 data Model = ModelSVM { modelSVM :: SVM.Model
  61                       , param1 :: Maybe Double
  62                       , param2 :: Maybe Double
  63                       }
  64 --{-
  65 instance SaveFile Model
  66   where
  67     saveFile' fp (ModelSVM m _ _) = SVM.saveModel m fp
  68
  69 instance ReadFile Model
  70   where
  71     readFile' fp = do
  72       m <- SVM.loadModel fp
  73       pure $ ModelSVM m Nothing Nothing
  74 --}
  75 ------------------------------------------------------------------------
  76 -- | TODO
  77 -- shuffle list
  78 -- split list : train / test
  79 -- grid parameters on best result on test
  80
  81 type Train = Map ListType [Vec.Vector Double]
  82 type Tests = Map ListType [Vec.Vector Double]
  83 type Score = Double
  84 type Param = Double
  85
  86 grid :: (MonadBase IO m)
  87      => Param -> Param -> Train -> [Tests] -> m (Maybe Model)
  88 grid _ _ _ []  = panic "Gargantext.Core.Text.List.Learn.grid : empty test data"
  89 grid s e tr te = do
  90   let
  91     grid' :: (MonadBase IO m)
  92           => Double -> Double
  93           -> Train
  94           -> [Tests]
  95           -> m (Score, Model)
  96     grid' x y tr' te' = do
  97       model'' <- liftBase $ trainList x y tr'
  98
  99       let
 100         model' = ModelSVM model'' (Just x) (Just y)
 101
 102         score' :: [(ListType, Maybe ListType)] -> Map (Maybe Bool) Int
 103         score' = occurrencesWith (\(a,b) -> (==) <$> Just a <*> b)
 104
 105         score'' :: Map (Maybe Bool) Int -> Double
 106         score'' m'' = maybe 0 (\t -> (fromIntegral t)/total) (Map.lookup (Just True) m'')
 107           where
 108             total = fromIntegral $ foldl (+) 0 $ Map.elems m''
 109
 110         getScore m t = do
 111           let (res, toGuess) = List.unzip
 112                              $ List.concat
 113                              $ map (\(k,vs) -> zip (repeat k) vs)
 114                              $ Map.toList t
 115
 116           res' <- liftBase $ predictList m toGuess
 117           pure $ score'' $ score' $ List.zip res res'
 118
 119       score <- mapM (getScore model') te'
 120       pure (mean score, model')
 121
 122   r <- head . List.reverse
 123             . (List.sortOn fst)
 124            <$> mapM (\(x,y) -> grid' x y tr te)
 125                          [(x,y) | x <- [s..e], y <- [s..e]]
 126
 127   printDebug "GRID SEARCH" (map fst r)
 128   --printDebug "file" fp
 129   --fp <- saveFile (ModelSVM model')
 130   --save best result
 131   pure $ snd <$> r