src/Data/Gargantext/Parsers.hs

   1 {-|
   2 Module      : Data.Gargantext.Parsers
   3 Description : All parsers of Gargantext in one file.
   4 Copyright   : (c) CNRS, 2017
   5 License     : AGPL + CECILL v3
   6 Maintainer  : alexandre.delanoe@iscpif.fr
   7 Stability   : experimental
   8 Portability : POSIX
   9
  10 Gargantext enables analyzing semi-structured text that should be parsed
  11 in order to be analyzed.
  12
  13 The parsers suppose, we know the format of the Text (TextFormat data
  14 type) according which the right parser is chosen among the list of
  15 available parsers.
  16
  17 This module mainly describe how to add a new parser to Gargantext,
  18 please follow the types.
  19 -}
  20
  21
  22 module Data.Gargantext.Parsers ( module Data.Gargantext.Parsers.WOS
  23                                --, module Data.Gargantext.Parsers.XML
  24                                --, module Data.Gargantext.Parsers.DOC
  25                                --, module Data.Gargantext.Parsers.ODS
  26                                )
  27     where
  28
  29
  30 import Data.Attoparsec.ByteString
  31 import Data.ByteString (ByteString)
  32 import Data.Map                    as DM
  33 import Data.Either.Extra(Either(..))
  34
  35 import Control.Monad (join)
  36 import Codec.Archive.Zip
  37 import Path.IO (resolveFile')
  38 -- import qualified Data.ByteString.Lazy as B
  39 import Control.Applicative ( (<$>) )
  40
  41
  42
  43 import Control.Concurrent.Async as CCA (mapConcurrently)
  44
  45
  46 import Data.Gargantext.Parsers.WOS (wosParser)
  47 -- import Data.Gargantext.Parsers.XML (xmlParser)
  48 -- import Data.Gargantext.Parsers.DOC (docParser)
  49 -- import Data.Gargantext.Parsers.ODS (odsParser)
  50
  51 import Data.Gargantext.Prelude
  52 import Data.Gargantext.Types.Main (ErrorMessage(), GargParser(), Corpus)
  53
  54
  55 -- | According to the format of Input file,
  56 -- different parser are available.
  57 data FileFormat = WOS        -- Implemented (ISI Format)
  58                 | XML        -- Not Implemented / see :
  59                              -- > http://chrisdone.com/posts/fast-haskell-c-parsing-xml
  60                 | DOC        -- Not Implemented / import Pandoc
  61                 | ODS        -- Not Implemented / import Pandoc
  62                 | PDF        -- Not Implemented / pdftotext and import Pandoc ?
  63
  64
  65
  66 -- | withParser:
  67 -- According the format of the text, choosing the right parser.
  68 withParser :: FileFormat -> GargParser
  69 withParser WOS = wosParser
  70 --withParser XML = xmlParser
  71 --withParser DOC = docParser
  72 --withParser ODS = odsParser
  73 withParser _   = error "[ERROR] Parser not implemented yet"
  74
  75
  76 runParser :: FileFormat -> ByteString -> Either ErrorMessage (IO (Maybe Corpus))
  77 runParser format text = parseOnly (withParser format) text
  78
  79
  80 parseZip :: FilePath -> ByteString -> IO Corpus
  81 parseZip = undefined
  82
  83 parseFile :: FileFormat -> ByteString -> IO Corpus
  84 parseFile p x = case runParser p x of
  85         Left  _ -> pure 0
  86         Right r -> pure $ length r
  87
  88
  89 openZipFiles :: FilePath -> IO [ByteString]
  90 openZipFiles fp = do
  91     path    <- resolveFile' fp
  92     entries <- withArchive path (DM.keys <$> getEntries)
  93     bs      <- mapConcurrently (\s -> withArchive path (getEntry s)) entries
  94     pure bs
  95
  96
  97 wosParserTest :: FilePath -> IO [Int]
  98 wosParserTest fp = join $ mapConcurrently (parseFile WOS) <$> openZipFiles fp
  99
 100