src/Gargantext/Core/Text/Corpus/Parsers/Book.hs

   1 {-|
   2 Module      : Gargantext.Core.Text.Corpus.Parsers.Book
   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 Get Book into GarganText
  11
  12 -}
  13
  14 module Gargantext.Core.Text.Corpus.Parsers.Book
  15   where
  16
  17 import Data.Maybe
  18 import Data.Text (Text)
  19 import GHC.IO (FilePath)
  20 import Gargantext.Core (Lang(..))
  21 import Gargantext.Core.Text.Corpus.Parsers.CSV (hyperdataDocument2csv)
  22 import Gargantext.Core.Text.Corpus.Parsers.FrameWrite (text2titleParagraphs)
  23 import Gargantext.Database.Admin.Types.Hyperdata (HyperdataDocument(..))
  24 import Gargantext.Prelude
  25 import System.Directory -- (getDirectoryContents)
  26 import qualified Data.ByteString.Lazy as DBL
  27 import qualified Data.List as List
  28 import qualified Data.Text as DT
  29
  30 ------------------------------------------------------------------------
  31 -- Main Export Function
  32
  33 type FileOut = FilePath
  34
  35 book2csv :: Int -> FileDir -> FileOut -> IO ()
  36 book2csv n f_in f_out = do
  37   files <- filesOf f_in
  38   texts <- readPublis f_in files
  39   let publis = List.concat $ map (file2publi n) texts
  40   let docs = map (\(y,p) -> publiToHyperdata y p) $ List.zip [1..] publis
  41   DBL.writeFile f_out (hyperdataDocument2csv docs)
  42
  43 filesOf :: FileDir -> IO [FilePath]
  44 filesOf fd = List.sort                                        -- sort by filename
  45         <$>  List.filter (\f -> DT.length (cs f) > 2)
  46         <$>  getDirectoryContents fd
  47
  48 readPublis :: FileDir -> [FilePath] -> IO [(FilePath, Text)]
  49 readPublis fd fps = mapM (\fp -> DBL.readFile (fd <> fp) >>= \txt -> pure (fp, cs txt)) fps
  50
  51 ------------------------------------------------------------------------
  52 -- Main Types
  53 data Publi = Publi { publi_authors :: [Text]
  54                    , publi_source  :: Text
  55                    , publi_title   :: Text
  56                    , publi_text    :: Text
  57                    }
  58   deriving (Show)
  59
  60 data FileInfo = FileInfo { fi_authors :: [Text]
  61                          , fi_source  :: Text
  62                          }
  63   deriving (Show)
  64
  65 type FileDir = FilePath
  66 ---------------------------------------------------------------------
  67
  68 file2publi :: Int -> (FilePath, Text) -> [Publi]
  69 file2publi n (fp,theText) = map (\(t,txt) -> Publi authors source t txt) theTexts
  70   where
  71     theTexts = text2titleParagraphs n theText
  72     FileInfo authors source = fileNameInfo fp
  73
  74 fileNameInfo :: FilePath -> FileInfo
  75 fileNameInfo fp = toFileInfo xs
  76   where
  77     xs = DT.splitOn "_" $ DT.pack fp
  78     toFileInfo (a:b:_) = FileInfo (DT.splitOn "-and-" a) (cs b)
  79     toFileInfo _       = panic "error"
  80
  81 ---------------------------------------------------------------------
  82 publiToHyperdata :: Int -> Publi -> HyperdataDocument
  83 publiToHyperdata y (Publi a s t txt) =
  84        HyperdataDocument { _hd_bdd = Just "Book File"
  85                         , _hd_doi = Nothing
  86                         , _hd_url = Nothing
  87                         , _hd_uniqId = Nothing
  88                         , _hd_uniqIdBdd = Nothing
  89                         , _hd_page = Nothing
  90                         , _hd_title = Just t
  91                         , _hd_authors = Just (DT.concat a)
  92                         , _hd_institutes = Nothing
  93                         , _hd_source = Just s
  94                         , _hd_abstract = Just txt
  95                         , _hd_publication_date = Nothing
  96                         , _hd_publication_year = Just y
  97                         , _hd_publication_month = Just 1
  98                         , _hd_publication_day = Just 1
  99                         , _hd_publication_hour = Nothing
 100                         , _hd_publication_minute = Nothing
 101                         , _hd_publication_second = Nothing
 102                         , _hd_language_iso2 = Just $ DT.pack $ show FR
 103                         }
 104
 105 -------------------------------------------------------------
 106 -- MISC tool to remove urls for instance
 107 clean :: Text -> Text
 108 clean = DT.unwords . List.filter (\w -> DT.length w < 20) . DT.words