src/Gargantext/API/Node/Corpus/Export.hs

   1 {-|
   2 Module      : Gargantext.API.Node.Corpus.Export
   3 Description : Get Metrics from Storage (Database like)
   4 Copyright   : (c) CNRS, 2017-Present
   5 License     : AGPL + CECILL v3
   6 Maintainer  : team@gargantext.org
   7 Stability   : experimental
   8 Portability : POSIX
   9
  10 Main exports of Gargantext:
  11 - corpus
  12 - document and ngrams
  13 - lists
  14 -}
  15
  16 module Gargantext.API.Node.Corpus.Export
  17   where
  18
  19
  20 import Data.Map (Map)
  21 import Data.Maybe (fromMaybe)
  22 import Data.Set (Set)
  23 import Data.Text (Text)
  24 import Gargantext.API.Node.Corpus.Export.Types
  25 import Gargantext.API.Ngrams.Types
  26 import Gargantext.API.Ngrams.Tools (filterListWithRoot, mapTermListRoot, getRepo)
  27 import Gargantext.API.Prelude (GargNoServer)
  28 import Gargantext.Prelude.Crypto.Hash (hash)
  29 import Gargantext.Core.Types
  30 import Gargantext.Database.Action.Metrics.NgramsByNode (getNgramsByNodeOnlyUser)
  31 import Gargantext.Database.Admin.Config (userMaster)
  32 import Gargantext.Database.Admin.Types.Hyperdata (HyperdataDocument(..))
  33 import Gargantext.Database.Prelude (Cmd)
  34 import Gargantext.Database.Query.Table.Node
  35 import Gargantext.Database.Query.Table.Node.Error (HasNodeError)
  36 import Gargantext.Database.Query.Table.Node.Select (selectNodesWithUsername)
  37 import Gargantext.Database.Query.Table.NodeNode (selectDocNodes)
  38 import Gargantext.Database.Schema.Ngrams (NgramsType(..))
  39 import Gargantext.Database.Schema.Node (_node_id, _node_hyperdata)
  40 import Gargantext.Prelude
  41 import qualified Data.List as List
  42 import qualified Data.Map as Map
  43 import qualified Data.Set as Set
  44
  45 --------------------------------------------------
  46 -- | Hashes are ordered by Set
  47 getCorpus :: CorpusId
  48           -> Maybe ListId
  49           -> Maybe NgramsType
  50           -> GargNoServer Corpus
  51 getCorpus cId lId nt' = do
  52
  53   let
  54     nt = case nt' of
  55       Nothing -> NgramsTerms
  56       Just  t -> t
  57
  58   ns   <- Map.fromList
  59        <$> map (\n -> (_node_id n, n))
  60        <$> selectDocNodes cId
  61   repo <- getRepo
  62   ngs  <- getNodeNgrams cId lId nt repo
  63   let  -- uniqId is hash computed already for each document imported in database
  64     r = Map.intersectionWith (\a b -> Document a (Ngrams (Set.toList b) (hash b)) (d_hash a b)
  65                              ) ns ngs
  66           where
  67             d_hash  a b = hash [ fromMaybe "" (_hd_uniqId $ _node_hyperdata a)
  68                                        , hash b
  69                                        ]
  70   pure $ Corpus (Map.elems r) (hash $ List.map _d_hash
  71                                             $ Map.elems r
  72                               )
  73
  74 getNodeNgrams :: HasNodeError err
  75         => CorpusId
  76         -> Maybe ListId
  77         -> NgramsType
  78         -> NgramsRepo
  79         -> Cmd err (Map NodeId (Set Text))
  80 getNodeNgrams cId lId' nt repo = do
  81   lId <- case lId' of
  82     Nothing -> defaultList cId
  83     Just  l -> pure l
  84
  85   lIds <- selectNodesWithUsername NodeList userMaster
  86   let ngs = filterListWithRoot MapTerm $ mapTermListRoot [lId] nt repo
  87   r <- getNgramsByNodeOnlyUser cId (lIds <> [lId]) nt (Map.keys ngs)
  88   pure r
  89
  90 -- TODO
  91 -- Exports List
  92 -- Version number of the list