]> Git — Sourcephile - gargantext.git/blob - src/Gargantext/Core/Methods/Distances/Conditional.hs
[Distances] Distributional WIP
[gargantext.git] / src / Gargantext / Core / Methods / Distances / Conditional.hs
1 {-|
2 Module : Gargantext.Core.Methods.Distances
3 Description :
4 Copyright : (c) CNRS, 2017-Present
5 License : AGPL + CECILL v3
6 Maintainer : team@gargantext.org
7 Stability : experimental
8 Portability : POSIX
9
10 Motivation and definition of the @Conditional@ distance.
11 -}
12
13 {-# LANGUAGE BangPatterns #-}
14 {-# LANGUAGE Strict #-}
15 module Gargantext.Core.Methods.Distances.Conditional
16 where
17
18 import Data.Matrix hiding (identity)
19
20 import Data.List (sortOn)
21
22 import Data.Map (Map)
23 import qualified Data.Map as M
24
25 import qualified Data.Set as S
26
27 import qualified Data.Vector as V
28
29 import Gargantext.Prelude
30 import Gargantext.Core.Viz.Graph.Utils
31
32 ------------------------------------------------------------------------
33 ------------------------------------------------------------------------
34 -- | Optimisation issue
35
36 toBeOptimized :: (Num a, Fractional a, Ord a) => Matrix a -> Matrix a
37 toBeOptimized m = proba Col m
38
39 ------------------------------------------------------------------------
40 -- | Main Functions
41 -- Compute the probability from axis
42 -- x' = x / (sum Col x)
43 proba :: (Num a, Fractional a) => Axis -> Matrix a -> Matrix a
44 proba a m = mapOn a (\c x -> x / V.sum (axis a c m)) m
45
46
47 mapOn :: Axis -> (AxisId -> a -> a) -> Matrix a -> Matrix a
48 mapOn a f m = V.foldl' f' m (V.enumFromTo 1 (nOf a m))
49 where
50 f' m' c = mapOnly a f c m'
51
52 mapOnly :: Axis -> (AxisId -> a -> a) -> AxisId -> Matrix a -> Matrix a
53 mapOnly Col = mapCol
54 mapOnly Row = mapRow
55
56 mapAll :: (a -> a) -> Matrix a -> Matrix a
57 mapAll f m = mapOn Col (\_ -> f) m
58
59
60 ---------------------------------------------------------------
61 -- | Compute a distance from axis
62 -- xs = (sum Col x') - x'
63 distFromSum :: (Num a, Fractional a)
64 => Axis -> Matrix a -> Matrix a
65 distFromSum a m = mapOn a (\c x -> V.sum (axis a c m) - x) m
66 ---------------------------------------------------------------
67 ---------------------------------------------------------------
68 -- | To compute included/excluded or specific/generic scores
69 opWith :: (Fractional a1, Num a1)
70 => (Matrix a2 -> t -> Matrix a1) -> Matrix a2 -> t -> Matrix a1
71 opWith op xs ys = mapAll (\x -> x / (2*n -1)) (xs `op` ys)
72 where
73 n = fromIntegral $ nOf Col xs
74 ---------------------------------------------------------------
75
76
77 -------------------------------------------------------
78 conditional :: (Num a, Fractional a, Ord a) => Matrix a -> Matrix a
79 conditional m = filterMat (threshold m') m'
80 where
81 ------------------------------------------------------------------------
82 -- | Main Operations
83 -- x' = x / (sum Col x)
84 x' = proba Col m
85
86 ------------------------------------------------------------------------
87 -- xs = (sum Col x') - x'
88 xs = distFromSum Col x'
89 -- ys = (sum Row x') - x'
90 ys = distFromSum Row x'
91
92 ------------------------------------------------------------------------
93 -- | Top included or excluded
94 ie = opWith (+) xs ys
95 -- ie = ( xs + ys) / (2 * (x.shape[0] - 1))
96
97 -- | Top specific or generic
98 sg = opWith (-) xs ys
99 -- sg = ( xs - ys) / (2 * (x.shape[0] - 1))
100
101 nodes_kept :: [Int]
102 nodes_kept = take k' $ S.toList
103 $ foldl' (\s (n1,n2) -> insert [n1,n2] s) S.empty
104 $ map fst
105 $ nodes_included k <> nodes_specific k
106
107 nodes_included n = take n $ sortOn snd $ toListsWithIndex ie
108 nodes_specific n = take n $ sortOn snd $ toListsWithIndex sg
109 insert as s = foldl' (\s' a -> S.insert a s') s as
110 k' = 2*k
111 k = 10
112
113 dico_nodes :: Map Int Int
114 dico_nodes = M.fromList $ zip ([1..] :: [Int]) nodes_kept
115 --dico_nodes_rev = M.fromList $ zip nodes_kept [1..]
116
117 m' = matrix (length nodes_kept)
118 (length nodes_kept)
119 (\(i,j) -> getElem ((M.!) dico_nodes i) ((M.!) dico_nodes j) x')
120
121 threshold m'' = V.minimum
122 $ V.map (\cId -> V.maximum $ getCol cId m'')
123 (V.enumFromTo 1 (nOf Col m'') )
124
125 filterMat t m'' = mapAll (\x -> filter' t x) m''
126 where
127 filter' t' x = case (x >= t') of
128 True -> x
129 False -> 0
130 ------------------------------------------------------------------------