src/Gargantext/Parsers.hs

   1 {-|
   2 Module      : Gargantext.Parsers
   3 Description : All parsers of Gargantext in one file.
   4 Copyright   : (c) CNRS, 2017
   5 License     : AGPL + CECILL v3
   6 Maintainer  : team@gargantext.org
   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 to 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 module Gargantext.Parsers -- (parse, FileFormat(..))
  22     where
  23
  24 import Gargantext.Prelude
  25
  26 import System.FilePath (takeExtension, FilePath())
  27 import Data.Attoparsec.ByteString (parseOnly, Parser)
  28 import qualified Data.ByteString as DB
  29 import qualified Data.Map        as DM
  30 import Data.Either.Extra (partitionEithers)
  31 import Data.Ord()
  32 import Data.Foldable (concat)
  33 import Data.String()
  34 import Data.Either.Extra(Either())
  35
  36 import Data.Text (Text)
  37 import Data.Text.Encoding (decodeUtf8)
  38 ----
  39 --import Control.Monad (join)
  40 import Codec.Archive.Zip (withArchive, getEntry, getEntries)
  41 import Path.IO (resolveFile')
  42 ------ import qualified Data.ByteString.Lazy as B
  43 --import Control.Applicative ( (<$>) )
  44 import Control.Concurrent.Async as CCA (mapConcurrently)
  45
  46 import Data.String (String())
  47 import Gargantext.Parsers.WOS (wosParser)
  48 ---- import Gargantext.Parsers.XML (xmlParser)
  49 ---- import Gargantext.Parsers.DOC (docParser)
  50 ---- import Gargantext.Parsers.ODT (odtParser)
  51
  52 --import Gargantext.Prelude (pm)
  53 --import Gargantext.Types.Main (ErrorMessage(), Corpus)
  54
  55
  56 -- | According to the format of Input file,
  57 -- different parser are available.
  58 data FileFormat = WOS        -- Implemented (ISI Format)
  59 --                | DOC        -- Not Implemented / import Pandoc
  60 --                | ODT        -- Not Implemented / import Pandoc
  61 --                | PDF        -- Not Implemented / pdftotext and import Pandoc ?
  62 --                | XML        -- Not Implemented / see :
  63 --                             -- > http://chrisdone.com/posts/fast-haskell-c-parsing-xml
  64
  65 -- TODO: to debug maybe add the filepath in error message
  66 type ParseError = String
  67
  68
  69 parse :: FileFormat -> FilePath -> IO ([ParseError], [[(Text, Text)]])
  70 parse format path = do
  71     files <- case takeExtension path of
  72               ".zip" -> openZip              path
  73               _      -> pure <$> DB.readFile path
  74     (as, bs) <- partitionEithers <$> mapConcurrently (runParser format) files
  75     pure (as, map toText $ concat bs)
  76       where
  77         -- TODO : decode with bayesian inference on encodings
  78         toText = map (\(a,b) -> (decodeUtf8 a, decodeUtf8 b))
  79
  80
  81 -- | withParser:
  82 -- According the format of the text, choosing the right parser.
  83 -- TODO  withParser :: FileFormat -> Parser [Document]
  84 withParser :: FileFormat -> Parser [[(DB.ByteString, DB.ByteString)]]
  85 withParser WOS = wosParser
  86 --withParser DOC = docParser
  87 --withParser ODT = odtParser
  88 --withParser XML = xmlParser
  89 --withParser _   = error "[ERROR] Parser not implemented yet"
  90
  91 runParser :: FileFormat -> DB.ByteString
  92           -> IO (Either String [[(DB.ByteString, DB.ByteString)]])
  93 runParser format text = pure $ parseOnly (withParser format) text
  94
  95 openZip :: FilePath -> IO [DB.ByteString]
  96 openZip fp = do
  97     path    <- resolveFile' fp
  98     entries <- withArchive path (DM.keys <$> getEntries)
  99     bs      <- mapConcurrently (\s -> withArchive path (getEntry s)) entries
 100     pure bs
 101
 102