src/Gargantext/Text/Parsers/RIS.hs

   1 {-|
   2 Module      : Gargantext.Text.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 {-# LANGUAGE NoImplicitPrelude #-}
  20 {-# LANGUAGE OverloadedStrings #-}
  21
  22 module Gargantext.Text.Parsers.RIS (risParser, risDate, toDate, presseParser) where
  23
  24 import Data.Either (either)
  25 import Data.List (lookup)
  26 import Data.Tuple.Extra (first)
  27 import Control.Applicative
  28 import Data.Attoparsec.ByteString (Parser, try, string, takeTill, take, manyTill, many1, endOfInput, parseOnly)
  29 import Data.Attoparsec.ByteString.Char8 (anyChar, isEndOfLine)
  30 import Data.ByteString (ByteString, concat, length)
  31 import Data.ByteString.Char8 (pack)
  32 import Data.Monoid ((<>))
  33 import Gargantext.Prelude hiding (takeWhile, take, concat, readFile, lines, concat)
  34 import qualified Data.List as DL
  35 -------------------------------------------------------------
  36
  37 risParser :: Parser [[(ByteString, ByteString)]]
  38 risParser = do
  39     n  <- notice "TY  -"
  40     ns <- many1 (notice "\nTY  -")
  41     pure $ [n] <> ns
  42
  43 notice :: Parser ByteString -> Parser [(ByteString, ByteString)]
  44 notice s = start *> many field <* end
  45     where
  46       start :: Parser ByteString
  47       start = s *> takeTill isEndOfLine
  48
  49       end :: Parser ByteString
  50       end =  "\nER  -" *> takeTill isEndOfLine
  51
  52 field :: Parser (ByteString, ByteString)
  53 field = do
  54     name  <- "\n" *> take 2 <* "  - "
  55     txt   <- takeTill isEndOfLine
  56     txts  <- try lines
  57     let txts' = case DL.length txts > 0 of
  58             True  -> txts
  59             False -> []
  60     pure (translate name, concat ([txt] <> txts'))
  61
  62 lines :: Parser [ByteString]
  63 lines = many line
  64     where
  65         line :: Parser ByteString
  66         line = "\n\n" *> takeTill isEndOfLine
  67
  68 translate :: ByteString -> ByteString
  69 translate champs
  70             | champs == "AU" = "authors"
  71             | champs == "TI" = "title"
  72             | champs == "JF" = "source"
  73             | champs == "LA" = "language"
  74             | champs == "DI" = "doi"
  75             | champs == "UR" = "url"
  76             | champs == "N2" = "abstract"
  77             | otherwise  = champs
  78 -------------------------------------------------------------
  79
  80 presseParser :: [(ByteString, ByteString)] -> [(ByteString, ByteString)]
  81 presseParser = (toDate "DA" (\x -> either (const []) identity $ parseOnly risDate x))
  82              . (toDate "LA" presseLang)
  83
  84 risDate :: Parser [(ByteString, ByteString)]
  85 risDate = do
  86   day <- take 2 <* "/"
  87   mon <- take 2 <* "/"
  88   yea <- take 4
  89   pure $ map (first (\x -> "publication_" <> x))
  90        [ ("day",day)
  91        , ("month", mon)
  92        , ("year", yea)
  93        , ("date", yea <> "-" <> mon <> "-" <> day <> "T0:0:0")
  94        ]
  95
  96 toDate :: ByteString -> (ByteString -> [(ByteString, ByteString)])
  97        -> [(ByteString, ByteString)] -> [(ByteString, ByteString)]
  98 toDate k f m = m <> ( maybe [] f (lookup k m) )
  99
 100 presseLang :: ByteString -> [(ByteString, ByteString)]
 101 presseLang "Français" = [("language", "FR")]
 102 presseLang "English"  = [("langauge", "EN")]
 103 presseLang _ = undefined
 104
 105 {-
 106 fixTitle :: [(ByteString, ByteString)] -> [(ByteString, ByteString)]
 107 fixTitle ns = ns <> [ti, ab]
 108   where
 109     ti = case
 110 -}