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

   1 {-|
   2 Module      : Gargantext.Core.Text.Corpus.Parsers.RIS
   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
  11 RIS is a standardized tag format developed by Research Information
  12 Systems, Incorporated (the format name refers to the company) to enable
  13 citation programs to exchange data.
  14
  15 [More](https://en.wikipedia.org/wiki/RIS_(file_format))
  16
  17 -}
  18
  19
  20 module Gargantext.Core.Text.Corpus.Parsers.RIS (parser, onField, fieldWith, lines) where
  21
  22 import Data.List (lookup)
  23 import Control.Applicative
  24 import Data.Attoparsec.ByteString (Parser, try, takeTill, take, many1)
  25 import Data.Attoparsec.ByteString.Char8 (isEndOfLine)
  26 import Data.ByteString (ByteString, concat)
  27 import Data.Monoid ((<>))
  28 import Gargantext.Prelude hiding (takeWhile, take)
  29 import qualified Data.List as DL
  30 -------------------------------------------------------------
  31
  32 parser :: Parser [[(ByteString, ByteString)]]
  33 parser = do
  34     n  <- notice "TY  -"
  35     ns <- many1 (notice "\nTY  -")
  36     pure $ [n] <> ns
  37
  38 notice :: Parser ByteString -> Parser [(ByteString, ByteString)]
  39 notice s = start *> many (fieldWith field)  <* end
  40     where
  41       field :: Parser ByteString
  42       field = "\n" *> take 2 <* "  - "
  43
  44       start :: Parser ByteString
  45       start = s *> takeTill isEndOfLine
  46
  47       end :: Parser ByteString
  48       end =  "\nER  -" *> takeTill isEndOfLine
  49
  50
  51 fieldWith :: Parser ByteString -> Parser (ByteString, ByteString)
  52 fieldWith n = do
  53     name  <- n
  54     txt   <- takeTill isEndOfLine
  55     txts  <- try lines
  56     let txts' = case DL.length txts > 0 of
  57             True  -> txts
  58             False -> []
  59     pure (name, concat ([txt] <> txts'))
  60
  61
  62 lines :: Parser [ByteString]
  63 lines = many line
  64     where
  65         line :: Parser ByteString
  66         line = "\n " *> takeTill isEndOfLine
  67
  68 -------------------------------------------------------------
  69 -- Field for First elem of a Tuple, Key for corresponding Map
  70 onField :: ByteString -> (ByteString -> [(ByteString, ByteString)])
  71        -> [(ByteString, ByteString)] -> [(ByteString, ByteString)]
  72 onField k f m = m <> ( maybe [] f (lookup k m) )
  73