Commit 912341c4 authored by Quentin Lobbé's avatar Quentin Lobbé

Add sub cooc matrix to each PhyloGroup of level 1 and more

parent 1179d082
Pipeline #253 failed with stage
...@@ -122,6 +122,7 @@ data PhyloGroup = ...@@ -122,6 +122,7 @@ data PhyloGroup =
, _phylo_groupLabel :: Text , _phylo_groupLabel :: Text
, _phylo_groupNgrams :: [Int] , _phylo_groupNgrams :: [Int]
, _phylo_groupQuality :: Map Text Double , _phylo_groupQuality :: Map Text Double
, _phylo_groupCooc :: Map (Int, Int) Double
, _phylo_groupPeriodParents :: [Pointer] , _phylo_groupPeriodParents :: [Pointer]
, _phylo_groupPeriodChilds :: [Pointer] , _phylo_groupPeriodChilds :: [Pointer]
......
...@@ -31,7 +31,7 @@ import Control.Lens hiding (makeLenses, both, Level) ...@@ -31,7 +31,7 @@ import Control.Lens hiding (makeLenses, both, Level)
import Data.Bool (Bool, not) import Data.Bool (Bool, not)
import Data.List (concat, union, intersect, tails, tail, head, last, null, zip, sort, length, any, (++), (!!), nub, sortOn, reverse, splitAt, take) import Data.List (concat, union, intersect, tails, tail, head, last, null, zip, sort, length, any, (++), (!!), nub, sortOn, reverse, splitAt, take)
import Data.Map (Map, elems, member, adjust, singleton, (!), keys, restrictKeys, mapWithKey) import Data.Map (Map, elems, member, adjust, singleton, empty, (!), keys, restrictKeys, mapWithKey, filterWithKey)
import Data.Semigroup (Semigroup) import Data.Semigroup (Semigroup)
import Data.Set (Set) import Data.Set (Set)
import Data.Text (Text, unwords, toLower, words) import Data.Text (Text, unwords, toLower, words)
...@@ -82,8 +82,8 @@ getProximity p group group' = case p of ...@@ -82,8 +82,8 @@ getProximity p group group' = case p of
-- | To get the next or previous PhyloPeriod based on a given PhyloPeriodId -- | To get the next or previous PhyloPeriod based on a given PhyloPeriodId
getNextPeriods :: PairTo -> PhyloPeriodId -> [PhyloPeriodId] -> [PhyloPeriodId] getNextPeriods :: PairTo -> PhyloPeriodId -> [PhyloPeriodId] -> [PhyloPeriodId]
getNextPeriods to id l = case to of getNextPeriods to id l = case to of
Childs -> (tail . snd) next Childs -> unNested id ((tail . snd) next)
Parents -> (reverse . fst) next Parents -> unNested id ((reverse . fst) next)
_ -> panic ("[ERR][Viz.Phylo.Example.getNextPeriods] PairTo type not defined") _ -> panic ("[ERR][Viz.Phylo.Example.getNextPeriods] PairTo type not defined")
where where
-------------------------------------- --------------------------------------
...@@ -95,6 +95,17 @@ getNextPeriods to id l = case to of ...@@ -95,6 +95,17 @@ getNextPeriods to id l = case to of
Nothing -> panic ("[ERR][Viz.Phylo.Example.getNextPeriods] PhyloPeriodId not defined") Nothing -> panic ("[ERR][Viz.Phylo.Example.getNextPeriods] PhyloPeriodId not defined")
Just i -> i Just i -> i
-------------------------------------- --------------------------------------
-- | To have an non-overlapping next period
unNested :: PhyloPeriodId -> [PhyloPeriodId] -> [PhyloPeriodId]
unNested x l
| null l = []
| nested (fst $ head l) x = unNested x (tail l)
| nested (snd $ head l) x = unNested x (tail l)
| otherwise = l
--------------------------------------
nested :: Date -> PhyloPeriodId -> Bool
nested d prd = d >= fst prd && d <= snd prd
--------------------------------------
-- | To find the best set (max = 2) of Childs/Parents candidates based on a given Proximity mesure until a maximum depth (max = Period + 5 units ) -- | To find the best set (max = 2) of Childs/Parents candidates based on a given Proximity mesure until a maximum depth (max = Period + 5 units )
...@@ -217,15 +228,24 @@ phyloCooc = fisToCooc phyloFisFiltered phyloLinked_0_1 ...@@ -217,15 +228,24 @@ phyloCooc = fisToCooc phyloFisFiltered phyloLinked_0_1
-- | To Cliques into Groups -- | To Cliques into Groups
cliqueToGroup :: PhyloPeriodId -> Int -> Int -> Ngrams -> (Clique,Support) -> Phylo -> PhyloGroup cliqueToGroup :: PhyloPeriodId -> Int -> Int -> Ngrams -> (Clique,Support) -> Map (Date, Date) Fis -> Phylo -> PhyloGroup
cliqueToGroup period lvl idx label fis p = PhyloGroup ((period, lvl), idx) cliqueToGroup period lvl idx label fis m p = PhyloGroup ((period, lvl), idx)
label label
(sort $ map (\x -> ngramsToIdx x p) ngrams
$ Set.toList
$ fst fis
)
(singleton "support" (fromIntegral $ snd fis)) (singleton "support" (fromIntegral $ snd fis))
cooc
[] [] [] [] [] [] [] []
where
--------------------------------------
ngrams :: [Int]
ngrams = sort $ map (\x -> ngramsToIdx x p)
$ Set.toList
$ fst fis
--------------------------------------
cooc :: Map (Int, Int) Double
cooc = filterWithKey (\k _ -> elem (fst k) ngrams && elem (snd k) ngrams)
$ fisToCooc (restrictKeys m $ Set.fromList [period]) p
--------------------------------------
-- | To transform Fis into PhyloLevels -- | To transform Fis into PhyloLevels
...@@ -236,7 +256,7 @@ fisToPhyloLevel m p = over (phylo_periods . traverse) ...@@ -236,7 +256,7 @@ fisToPhyloLevel m p = over (phylo_periods . traverse)
fisList = zip [1..] (Map.toList (m ! periodId)) fisList = zip [1..] (Map.toList (m ! periodId))
in over (phylo_periodLevels) in over (phylo_periodLevels)
(\levels -> (\levels ->
let groups = map (\fis -> cliqueToGroup periodId 1 (fst fis) "" (snd fis) p) fisList let groups = map (\fis -> cliqueToGroup periodId 1 (fst fis) "" (snd fis) m p) fisList
in levels ++ [PhyloLevel (periodId, 1) groups] in levels ++ [PhyloLevel (periodId, 1) groups]
) period ) p ) period ) p
......
...@@ -213,6 +213,7 @@ initGroup ngrams lbl idx lvl from to p = PhyloGroup ...@@ -213,6 +213,7 @@ initGroup ngrams lbl idx lvl from to p = PhyloGroup
lbl lbl
(sort $ map (\x -> ngramsToIdx x p) ngrams) (sort $ map (\x -> ngramsToIdx x p) ngrams)
(Map.empty) (Map.empty)
(Map.empty)
[] [] [] [] [] [] [] []
......
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment