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

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