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

   1 {-|
   2 Module      : Gargantext.Core.Text.Corpus.Parsers.WOS
   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 Here is a longer description of this module, containing some
  11 commentary with @some markup@.
  12 -}
  13
  14
  15 module Gargantext.Core.Text.Corpus.Parsers.WOS (parser, keys) where
  16
  17 import Control.Applicative
  18 import Data.Attoparsec.ByteString (Parser, string, takeTill, take, manyTill, many1)
  19 import Data.Attoparsec.ByteString.Char8 (anyChar, isEndOfLine)
  20 import Data.ByteString (ByteString)
  21 import Data.ByteString.Char8 (pack)
  22 import Gargantext.Core.Text.Corpus.Parsers.RIS (fieldWith)
  23 import Prelude hiding (takeWhile, take, concat, readFile, lines, concat)
  24
  25 -------------------------------------------------------------
  26 -- | wosParser parses ISI format from
  27 -- Web Of Science Database
  28 parser :: Parser [[(ByteString, ByteString)]]
  29 parser = do
  30     -- TODO Warning if version /= 1.0
  31     -- FIXME anyChar (string ..) /= exact string "\nVR 1.0" ?
  32     _  <- manyTill anyChar (string $ pack "\nVR 1.0")
  33     ns <- many1 notice <*  (string $ pack "\nEF"    )
  34     pure ns
  35
  36 notice :: Parser [(ByteString, ByteString)]
  37 notice = start *> many (fieldWith field) <* end
  38     where
  39       field :: Parser ByteString
  40       field = "\n" *> take 2 <* " "
  41
  42       start :: Parser ByteString
  43       start = "\nPT " *> takeTill isEndOfLine
  44
  45       end :: Parser [Char]
  46       end = manyTill anyChar (string $ pack "\nER\n")
  47
  48
  49 keys :: ByteString -> ByteString
  50 keys champs
  51             | champs == "AF" = "authors"
  52             | champs == "TI" = "title"
  53             | champs == "SO" = "source"
  54             | champs == "DI" = "doi"
  55             | champs == "PD" = "publication_date"
  56             | champs == "AB" = "abstract"
  57             | otherwise  = champs