src/Gargantext/Database/Flow.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 add :: Corpus -> [Documents] -> IO Int
  13 if new id -> extractNgrams + extract Authors + extract Sources
  14 Map (Ngrams, NodeId)
  15 insert Ngrams -> NgramsId
  16 Map (NgramsId, NodeId) -> insert
  17
  18 data NgramsType = Sources | Authors | Terms
  19
  20 nodes_ngrams : column type, column list
  21
  22
  23 documents
  24 sources
  25 authors
  26
  27 -}
  28
  29 {-# LANGUAGE NoImplicitPrelude #-}
  30 {-# LANGUAGE OverloadedStrings #-}
  31
  32 module Gargantext.Database.Flow
  33   (flow)
  34     where
  35 import System.FilePath (FilePath)
  36 import Data.Maybe (Maybe(..))
  37 import Gargantext.Core.Types (NodePoly(..))
  38 import Gargantext.Prelude
  39 import Gargantext.Database.Bashql (runCmd', del)
  40 import Gargantext.Database.Node (getRoot, mkRoot, mkCorpus)
  41 import Gargantext.Database.User (getUser, UserLight(..))
  42 import Gargantext.Database.Node.Document.Import (insertDocuments)
  43 import Gargantext.Text.Parsers (parseDocs, FileFormat(WOS))
  44
  45 flow :: FilePath -> IO ()
  46 flow fp = do
  47   masterUser <- runCmd' (getUser "gargantua")
  48
  49   let masterUserId = case masterUser of
  50         Nothing   -> panic "Error: User does not exist (yet)" -- mk NodeUser gargantua_id "Node Gargantua"
  51         Just user -> userLight_id user
  52
  53   rootId' <- map _node_id <$> runCmd' (getRoot masterUserId)
  54
  55   rootId'' <- case rootId' of
  56         []  -> runCmd' (mkRoot masterUserId)
  57         un  -> case length un >= 2 of
  58                  True  -> panic "Error: more than 1 userNode / user"
  59                  False -> pure rootId'
  60   let rootId = maybe (panic "error rootId") identity (head rootId'')
  61   printDebug "Root ID : " rootId
  62
  63   corpusId' <- runCmd' $ mkCorpus (Just "Corpus WOS") Nothing rootId masterUserId
  64   let corpusId = maybe (panic "error corpusId") identity (head corpusId')
  65   printDebug "Corpus ID : " corpusId
  66
  67   docs <- parseDocs WOS fp
  68   ids  <- runCmd' $ insertDocuments masterUserId corpusId docs
  69   printDebug "Docs IDs : " ids
  70
  71   idsRepeat  <- runCmd' $ insertDocuments masterUserId corpusId docs
  72   printDebug "Docs IDs : " idsRepeat
  73
  74   _ <- runCmd' (del [corpusId])
  75   pure ()
  76
  77 {-
  78   --folderId <- mk Folder parentId (Name "Data") (Descr "All corpora DATA here")
  79   folderId <- mk Folder rootId "Data"
  80   corpusId <- mk Corpus folderId (Name "WOS")  (Descr "WOS database description")
  81 -}
  82
  83
  84
  85 {-
  86   docs <- parseDocuments WOS "doc/.."
  87   ids  <- add (Documents corpusId) docs
  88
  89   user_id <- runCmd' (get RootUser "alexandre")
  90   rootUser_id <- runCmd' (getRootUser $ userLight_id user_id
  91   corpusId <- mk Corpus
  92 -}
  93
  94