src/Gargantext/Database/Action/Flow/Pairing.hs

   1 {-|
   2 Module      : Gargantext.Database.Flow
   3 Description : Database Flow
   4 Copyright   : (c) CNRS, 2017-Present
   5 License     : AGPL + CECILL v3
   6 Maintainer  : team@gargantext.org
   7 Stability   : experimental
   8 Portability : POSIX
   9
  10 -}
  11
  12 {-# LANGUAGE FlexibleContexts  #-}
  13 {-# LANGUAGE QuasiQuotes       #-}
  14 {-# LANGUAGE NoImplicitPrelude #-}
  15 {-# LANGUAGE OverloadedStrings #-}
  16 {-# LANGUAGE RankNTypes        #-}
  17 -- {-# LANGUAGE Arrows #-}
  18
  19 module Gargantext.Database.Action.Flow.Pairing
  20   (pairing)
  21     where
  22
  23 import Control.Lens (_Just, (^.))
  24 import Data.Map (Map, fromList)
  25 import Data.Maybe (catMaybes)
  26 import Data.Text (Text, toLower)
  27 import Database.PostgreSQL.Simple.SqlQQ (sql)
  28 import Gargantext.Core.Types (TableResult(..))
  29 import Gargantext.Database.Action.Flow.Utils
  30 import Gargantext.Database.Admin.Types.Node (AnnuaireId, CorpusId, ListId{-, DocId, ContactId-})
  31 import Gargantext.Database.Admin.Utils (Cmd, runPGSQuery)
  32 import Gargantext.Database.Node.Children (getAllContacts)
  33 import Gargantext.Database.Node.Contact -- (HyperdataContact(..))
  34 import Gargantext.Database.Schema.Ngrams -- (NgramsType(..))
  35 import Gargantext.Prelude hiding (sum)
  36 import Safe (lastMay)
  37 import qualified Data.Map as DM
  38 import qualified Data.Text as DT
  39
  40 -- TODO mv this type in Types Main
  41 type Terms = Text
  42
  43 {-
  44 pairing'' :: (CorpusId, CorpusId) -> (DocId -> DocId)
  45 pairing'' = undefined
  46
  47 pairing' :: (CorpusId, AnnuaireId) -> (DocId -> ContactId)
  48 pairing' = undefined
  49 -}
  50
  51 -- | TODO : add paring policy as parameter
  52 pairing :: CorpusId   -- (CorpusId,   ListId) -- Pair (Either CorpusId AnnuaireId) ListId
  53         -> AnnuaireId -- (AnnuaireId, ListId) -- Pair (Either CorpusId AnnuaireId) ListId
  54         -> ListId
  55         -> Cmd err Int
  56 pairing cId aId lId = do
  57   contacts' <- getAllContacts aId
  58   let contactsMap = pairingPolicyToMap toLower
  59                   $ toMaps extractNgramsT (tr_docs contacts')
  60
  61   ngramsMap' <- getNgramsTindexed cId Authors
  62   let ngramsMap = pairingPolicyToMap lastName ngramsMap'
  63
  64   let indexedNgrams = pairMaps contactsMap ngramsMap
  65
  66   insertDocNgrams lId indexedNgrams
  67
  68 lastName :: Terms -> Terms
  69 lastName texte = DT.toLower
  70                $ maybe texte (\x -> if DT.length x > 3 then x else texte) (lastName' texte)
  71   where
  72     lastName' = lastMay . DT.splitOn " "
  73
  74 -- TODO: this method is dangerous (maybe equalities of the result are not taken into account
  75 -- emergency demo plan...)
  76 pairingPolicyToMap :: (Terms -> Terms)
  77                    -> Map (NgramsT Ngrams) a
  78                    -> Map (NgramsT Ngrams) a
  79 pairingPolicyToMap f = DM.mapKeys (pairingPolicy f)
  80
  81 pairingPolicy :: (Terms -> Terms)
  82               -> NgramsT Ngrams
  83               -> NgramsT Ngrams
  84 pairingPolicy f (NgramsT nt (Ngrams ng _)) = (NgramsT nt (Ngrams (f ng) 1))
  85
  86 -- | TODO : use Occurrences in place of Int
  87 extractNgramsT :: HyperdataContact
  88                -> Map (NgramsT Ngrams) Int
  89 extractNgramsT contact = fromList [(NgramsT Authors    a' , 1)| a' <- authors    ]
  90   where
  91     authors    = map text2ngrams
  92                $ catMaybes [ contact^.(hc_who . _Just . cw_lastName) ]
  93
  94
  95 pairMaps :: Map (NgramsT Ngrams) a
  96          -> Map (NgramsT Ngrams) NgramsId
  97          -> Map NgramsIndexed (Map NgramsType a)
  98 pairMaps m1 m2 =
  99   DM.fromList
 100     [ (NgramsIndexed ng nId, DM.singleton nt n2i)
 101     | (k@(NgramsT nt ng),n2i) <- DM.toList m1
 102     , Just nId <- [DM.lookup k m2]
 103     ]
 104
 105 -----------------------------------------------------------------------
 106 getNgramsTindexed :: CorpusId
 107                   -> NgramsType
 108                   -> Cmd err (Map (NgramsT Ngrams) NgramsId)
 109 getNgramsTindexed corpusId ngramsType' = fromList
 110     <$> map (\(ngramsId',t,n) -> (NgramsT ngramsType' (Ngrams t n),ngramsId'))
 111     <$> selectNgramsTindexed corpusId ngramsType'
 112   where
 113     selectNgramsTindexed :: CorpusId
 114                          -> NgramsType
 115                          -> Cmd err [(NgramsId, Terms, Int)]
 116     selectNgramsTindexed corpusId' ngramsType'' = runPGSQuery selectQuery (corpusId', ngramsTypeId ngramsType'')
 117       where
 118         selectQuery = [sql| SELECT n.id,n.terms,n.n from ngrams n
 119                       JOIN node_node_ngrams occ ON occ.ngrams_id = n.id
 120                       -- JOIN node_node_ngrams2 occ ON occ.ngrams_id = n.id
 121                       JOIN nodes_nodes      nn  ON nn.node2_id   = occ.node2_id
 122
 123                       WHERE nn.node1_id     = ?
 124                         AND occ.ngrams_type = ?
 125                         AND occ.node2_id = nn.node2_id
 126                       GROUP BY n.id;
 127                      |]
 128
 129 {- | TODO more typed SQL queries
 130 selectNgramsTindexed :: CorpusId -> NgramsType -> Query NgramsRead
 131 selectNgramsTindexed corpusId ngramsType = proc () -> do
 132     nodeNode   <- queryNodeNodeTable     -< ()
 133     nodeNgrams <- queryNodesNgramsTable  -< ()
 134     ngrams     <- queryNgramsTable       -< ()
 135
 136     restrict -< node1_id nodeNode .== pgInt4 corpusId
 137     restrict -< node2_id nodeNode .== node_id nodeNgrams
 138     restrict -< ngrams_id ngrams  .== node_ngrams nodeNgrams
 139
 140     result <- aggregate groupBy (ngrams_id ngrams)
 141     returnA -< result
 142 --}